I spend a lot of time inside revenue systems, which means I see a lot of the same decisions show up at different companies under different names. Lately I’ve been thinking about one change that happened so gradually we rarely talk about it.

Companies have taken over more and more of the work required to create sales opportunities.

A traditional sales rep had to make more of their own luck. They found the accounts, started the conversations, and followed up. They always had one ear to the ground, monitoring the market.

That model was okay. It produced sales but it wasn’t the most efficient and it was difficult to predict. Results often depended on individual heroics and buyers could easily be missed if a rep never happened to find them.

Inbound improved the process.

Marketing could create demand at scale. Forms could capture interest. Routing software could get the right lead to the right person. Scheduling tools could put a meeting directly on a calendar. Instead of asking every rep to build an entire market around themselves, a company could build a system that created opportunity for the whole team.

That was amazing progress. I don’t want to undo that – but think about it – inbound changed the sales job more than we usually admit. The role moved from finding and closing, to responding and closing, and sometimes all the way to receiving and closing.

The company began doing more of the work to create opportunities. The rep still owned the result, but they no longer owned the whole path that produced it.

Every leader feels this one in the gut, so let’s just say it out loud: A lot of reps act like they own the leads assigned to them. They forget the marketers and the site builders and the consultants who obsessed over how to get a stranger onto a page and move them to fill out a form. They forget the ad buyers and the copywriters and the designers who spent years A/B testing their way to a single sign-up. They forget the teammate who did the grinding to get that name onto the screen.

With inbound, reps have less of a stake in the leads they work because they’re often assigned rather than acquired. The resulting attitude can look like entitlement and it can feel annoying but it really just points to a design flaw in the sales org. What we need to do is realign incentive with effort.

Commission is often described as a reward for closing. I think it’s more useful to look at it as an alignment mechanism. The company can’t sit beside a rep for every prospecting hour, or every small judgment that moves a buyer forward. It pays on the outcome because the outcome is easier to observe.

That logic works when the rep creates the opportunity and then wins it but inbound makes the equation more complicated. A closed deal now reflects the brand, the campaign, the website, the routing rules, the timing, the territory, and the rep’s work. Some of the value comes from effort. Some comes from allocation.

This is where the gap begins.

Ownership arrives before effort

I didn’t notice the misalignment all at once. I saw it gradually.

I first started to see it in pipeline reviews, when a rep insisted an account was active even though nobody had spoken with the buyer in months. Then I saw it in routing conversations, when moving an untouched lead felt like taking something away.

I started hearing it in how people talked about “their” leads, even when the company had done most of the work to produce them.

Here’s how it goes:

• A lead is assigned to a rep’s name.

• It appears in their CRM.

• They add notes, enroll it in a sequence, and see it on a dashboard labeled “my pipeline.”

Very quickly, the lead begins to feel like theirs.

That feeling of ownership isn’t inherently bad. This psychological stance can create care, responsibility, and better work. We want reps to develop a real sense of stewardship over the people and opportunities they’re serving.

The problem is the order of operations.

In many inbound systems, ownership arrives before investment. The rep gets the attachment that comes with possession before they’ve done the work that normally gives ownership meaning. Then a subtle shift happens. The lead can become something to keep rather than something to advance.

A rep I was coaching once sent me an email – she had woken up to find three meetings (count ‘em, THREE!) booked on her calendar for that day. She wasn’t happy about it (huh?!). She wanted to restrict her calendar to create a full-day buffer between when a form was filled out and when a meeting could be booked. She didn’t want to have to change her plans the day of. Thing is – she’s a sales rep! Her job is to take meetings with interested customers. From my point of view, unplanned meetings should be part of the plan and they should feel like gold. These types of meetings are painstakingly won. They shouldn’t be received as a nuisance.

I understand that preparation matters but sales reps should be so acquainted with their product and their customer that they don’t need much prep time for a standard intro demo. So many pieces of this experience rubbed me the wrong way.

I started to realize it was just another example of the trend I’d been observing for years – that trend annoyed me more than the email itself. The rep hadn’t seen all the work that had gone into creating the demand. She hadn’t found those people, or earned their attention, or convinced them to take a meeting. The opportunity appeared at the end of a long system she didn’t own or really even influence.

By the time the leads reached her they just felt like extra tasks on her calendar.

It’s easy to call that entitlement but I don’t think labels help. The more useful question is why a system designed to create opportunity can make that opportunity feel so distant from the work of selling.

When opportunity becomes inventory

The same distance appears in pipeline management.

At one company, we found reps holding as many as 175 open deals. In a single stage, 953 deals had very little meaningful movement, and roughly 80 percent of the pipeline was no longer active in any practical sense.

No rep can meaningfully work 175 active deals. At that volume, pipeline becomes storage.

We saw the same thing with a large book of academic leads. The CRM showed emails, restarted sequences, and updated records, but almost nobody was responding. When the book was reassigned, another rep found real pipeline inside it.

This wasn’t a distinction between a good salesperson and a bad one. It was the difference between activity and progress. The system let the first rep preserve ownership with enough motion to keep the records in place, even when the buyers weren’t moving.

Possession and stewardship aren’t the same thing.

Make ownership conditional on progress

The answer isn’t to remove ownership. People do better work when they feel responsible for an outcome, and buyers benefit from continuity.

A more useful approach is to make ownership conditional on progress. An opportunity stays with a rep while the buyer is responding, the next step is clear, and the rep is doing the work. If it sits without meaningful movement, it returns to a shared pool or goes to someone with more capacity.

The rules have to measure the right thing. One automated email or token call shouldn’t preserve ownership. At the same time, not every deal moves on the same schedule. A complex enterprise deal may be quiet for good reasons, while a smaller inbound opportunity may require tighter follow-up.

The point isn’t to create one universal timer. It’s to define responsible ownership for each sales motion, make the rules visible, and apply them consistently. Then reassignment isn’t punishment. It’s how the company keeps opportunity moving.

Let contribution change access

Conditional ownership connects possession to ongoing work. Companies can also connect access to demonstrated contribution.

That doesn’t mean rebuilding the old model or making every rep create their entire pipeline. It means better access can follow demonstrated readiness. A rep might earn more high-intent inbound through consistent follow-through, self-sourced pipeline, revived opportunities, completed training, or strong call quality.

The exact currency will vary, but the relationship should be visible. Instead of waiting for routing to determine their month, reps can improve their skills, create pipeline, and demonstrate that they’re ready for more.

Inbound removed friction from lead creation. Good incentive design can preserve that efficiency while restoring a clear connection between effort and possibility.

Fairness isn’t the same as equality

Lead distribution naturally becomes a fairness conversation because it affects a rep’s income. But fairness can’t mean that every rep receives the same number and quality of leads regardless of capacity, follow-through, or performance.

A better measure is legitimacy.

A legitimate system has clear rules. It explains what is rewarded, accounts for differences in segment and sales cycle, and gives people a reasonable path to improve their access. Reps don’t need every outcome to be equal, but they do need to understand why it happened.

When the rules feel arbitrary, people create workarounds and the CRM stops reflecting reality. Transparent rules reduce that friction. They tell everyone: this opportunity is yours while you’re moving it; this is how you earn access to more; and this is what happens when you no longer have capacity.

That’s a stronger form of alignment than round-robin alone can create.

The goal is energy

I’m still the person who sees a new lead and immediately wants to know more.

Who are they? Why did they raise their hand? What are they trying to solve? Is there a real opportunity here? How can we move it forward?

That part of selling is exciting. I want reps to feel that.

You don’t create that energy by lecturing people about how lucky they are. It grows when reps can see that their judgment matters, their effort changes what becomes available to them, and good work leads somewhere.

Inbound gave companies a better way to create and distribute demand, but it also separated effort, ownership, and reward. The next step is to realign those elements.

We can preserve the efficiency of inbound while designing ownership that’s active, earned, and useful. The company gets more from the opportunities it creates. The rep gets a sense that their work is integral but also part of a larger system.

Share this with someone who could use a fresh take on aligning incentive in the age of inbound.

Share

We were on a call last week talking about this very scenario. There was a lead that came in, hit the scoring model, missed the threshold, and got routed to a nurture sequence. There was no discovery call and so there was no discovery.

How many opportunities are being sacrificed to aggregated efficiency?

Here are the numbers that should make everyone uncomfortable.

Only 5 to 10 percent of marketing-qualified leads ever become a discovery call. Of the deals that make it to a qualified pipeline, 25 to 30 percent close. Run that math on your funnel and look at what’s sitting in the gap. That’s not a qualification problem. That’s a conversation problem.

RevOps leaders have spent the last decade responding to a mandate: make the revenue machine more efficient, get the right leads to the right reps at the right time, reduce friction, and improve conversion rates. It’s a reasonable mandate, but the execution went sideways.

The tool that broke things was lead scoring

On paper it made sense: assign point values to behavioral signals and demographic fit, set a threshold, and let only the qualified leads through. The problem is what got optimized in the process. RevOps teams started measuring success by how clean the pipeline looked, not by how much the team was learning. Efficiency became the goal and discovery was the cost.

Sales leaders compounded the damage by not pushing back. Somewhere along the way, salespeople were handed the power to define what constituted a waste of their time. And they used it!

“Pre-qualify everything, filter harder, only show me the leads I know I can close.”

Leadership nodded along because it sounded like focus. Nobody asked the obvious question: who decided salespeople get to make that call?

The filter didn’t deliver certainty. It delivered fewer data points and a sales team that has lost the habit of doing real discovery.

The promise of lead scoring was that it would surface the buyers and screen out the noise. What it actually did was train salespeople to expect pre-validation. Now they want to know before they pick up the phone that the person on the other end is ready to buy, fits the ICP, has budget, has authority, and has already decided they need what you’re selling. At that point you’re not doing discovery. You’re taking an order.

That’s not the job.

The job is to take the call

Salespeople are paid to have their time wasted. Not because waste is good, but because you can’t know what you’re walking into until you’re in the conversation. The call is where the information lives. You find out if the form data was accurate (often it isn’t), if the lead understands their own problem (often they don’t), and if there’s something real worth pursuing.

Discovery is the highest form of leverage in your sales pipeline. Every other stage is downstream of it. Qualification, proposal, negotiation, close: all of it depends on the quality of what you learned in that first conversation. When you filter discovery out in the name of efficiency, you’re not saving time. You’re destroying the input that everything else runs on.

A great discovery call does four things:

• Do Discovery - Ask good questions, be genuinely curious, listen without an agenda.

• Cultivate Good Energy - bring a smile, enthusiastic about being connected, grateful for the time

• Cast a Vision - Create a scenario they can envision for themselves and help them see what solving this problem could look like.

• Determine Fit - Do we want to work with them? Do they want to work with us?

Most sales teams treat the fourth job as the only job. They jump on the call looking for disqualifiers and the conversation becomes a checklist. And because they’re skipping or not focused on the first three, they’re not even good at the fourth. You can’t determine fit without doing discovery first. The sequence matters.

Here’s what the efficiency model doesn’t measure

Every lead that gets bounced by a scoring system is a person who had an interaction with your brand and then didn’t hear from a human. Some of them were genuinely not ready; that’s fine. But some of them were a year away from being ready, and now they’re going to remember how your system treated them. Some of them work at the wrong company today and the right company in eighteen months. Some of them are going to talk to someone who is your ideal customer.

The goodwill that comes from taking a call seriously, even when the lead isn’t qualified, is one of the most undervalued assets in a sales organization. People remember how you treated them when you had nothing to gain: whether you were dismissive and clearly running a qualification script, or curious, generous, and actually helpful even when you couldn’t sell them anything that day. This is the long-cycle pipeline that never shows up in your attribution model.

There’s also a rep development argument that almost never gets made. Salespeople who talk to more people at more stages of market awareness get better faster. They learn how the market talks about problems, they learn the objections before they become objections, and they develop the ability to read a conversation and move it somewhere useful. That’s the actual skill the job requires. You can’t develop that on a diet of pre-validated, high-intent leads.

The fix requires sales leadership to make a cultural call

Stop letting salespeople define what constitutes a waste of their time, and stop optimizing your lead scoring model to protect the pipeline from unqualified conversations. Start measuring whether your team is doing real discovery, casting real vision, and building real relationships, even in the calls that don’t convert.

RevOps has a role here too, and it isn’t building a better filter. It’s building infrastructure that makes discovery better: pre-call research tooling, clear frameworks for what a great discovery call actually looks like, and feedback loops from conversations back into the qualification model. The technology should make the conversation smarter, not replace it.

Discovery is the first job of sales and the most important one. Everything that follows is just a consequence of how well your team does it.

For now, give them permission to take the call.

I’ve been working on a framework for revenue operations observability for the better part of a year now, and this one idea is the anchor.

Observation is never neutral.

You can’t measure a business before you’ve decided where you’re standing and what you’re looking at. The place you stand and the direction you look are what produces the observation. If you skip that step, your data is ungrounded. Numbers without a model behind them won’t tell you anything relevant about the business.

This is where I see most reporting in most companies fail. Someone asks for a report and someone else obliges but nobody pauses to ask which lever of the business the report will measure, what time horizon it should speak to, or which level of the organization needs to see it. The report gets built but then nobody quite knows what to do with it. The numbers are accurate but the observation behind them is unclear.

Observability doesn’t start with measurement. It starts with defining the model you’re going to use to observe the business.

The rest of this piece is about that model. The framework is a 3x3x3 matrix that combines three revenue levers, three time horizons, and three levels of the organization. That gives you 27 combinations. Every meaningful report you build will be in one of those 27 cells.

The three levers

Every business has three principal levers that produce revenue. The levers are volume, conversion, and value. Another way to describe them is inputs, throughputs, and outputs.

Volume is how many revenue opportunities are entering the system. Without sufficient volume, you don’t produce growth. Quantity alone is misleading though, so volume has to be observed alongside source, quality, timing, and fit. A thousand bad leads in the door is a completely different business from a hundred good ones. The metrics that live under volume include total lead count, marketing-qualified and sales-qualified leads, source attribution, channel mix, event-driven inflows, and the cost to acquire each one.

Conversion is how efficiently the system transforms an opportunity into revenue. The thing most people miss about conversion is that it includes velocity. Speed matters as much as the rate, sometimes more. A 30% close rate at 60 days is a completely different business from a 30% close rate at 6 months. Both numbers look fine on a slide, but they describe very different operations. Conversion metrics include stage conversion rates, overall win rate, sales cycle length, time in stage, and pipeline velocity.

Value is how much revenue a single opportunity produces. Value can show up as ACV, MRR, LTV, contract size, or expansion revenue, depending on how you account for it. Value must always be bound to a time window, whether that window is one year, two years, or the full customer lifetime. Value also gets measured very differently depending on the business model, which means the first question to ask about any business is how it thinks about value. A SaaS subscription is a completely different beast from a consumption contract, and consumption is going to keep growing as a model because of AI. Conversations about value have to start with the business model.

Each metric that matters comes back to one of these three levers, or to the interaction between them. If you can’t ground a number in volume, conversion, or value, you won’t know what that number relates to.

The three horizons

The next coordinate is time. There are three horizons of observation, and each one answers a different question.

Horizon 1 is operational. It operates in real time, minute by minute. The question this horizon answers is whether the system is running right now. Operational observability is like the temperature gauge on your car. The gauge has one job, which is to tell you when the engine is about to seize. Operational observability does the same thing for a revenue system. It tells you whether the lead form is firing, whether the routing logic is putting leads where they should go, and whether the integrations between your tools are syncing. When the temperature spikes, you want to know fast. Operational reports look like alerts, error logs, sync status pages, and form-submission counters that should be ticking up but suddenly aren’t.

Horizon 2 is directional. It tracks leading indicators on a time frame of hours to weeks. The question this horizon answers is where the vectors are pointing right now. At Innovation Refunds we ran the business in 4-hour windows. By noon every day we could predict how many deals would close before close of business. We could do that because we had built directional observability into the rhythm of the day. We watched whether volume was steady or climbing, whether qualified pipeline was moving, whether stage conversion was holding, and whether average deal size was sitting where we expected it. If the leading indicators were pointing in the wrong direction, we knew before the day was over. Directional reports look like daily pacing dashboards, week-over-week conversion trends, and pipeline-by-stage views with movement velocity built in.

Horizon 3 is strategic. It covers months, quarters, and years. The question this horizon answers is whether the system is producing the outcomes it was designed to produce. This is the horizon that executives default to, and it is the easiest to measure but the slowest to give you signal. By the time horizon 3 says something is broken, horizon 1 and horizon 2 said so weeks or months earlier. A lot of companies build horizon 3 reports because the board asks for them and then they wonder why every quarter end is a fire drill. Strategic reports look like annual revenue trajectories, retention cohorts, LTV-to-CAC ratios, and progress against multi-quarter goals.

The three levels

The last coordinate is who is looking. The three levels are executive, team, and individual or process.

Executive reporting maps to the strategic horizon. The question is what a CEO or board needs to see to know the business is on track. Team reporting breaks the business into functional units. Marketing, sales, and customer success each need a view that shows their part of the system performing or not. Individual and process reporting goes one level lower. The question becomes how a specific rep’s activities and meetings and win rate are feeding into the system, or how a specific campaign is performing, or how an automated routing flow is doing its job when no human is touching it.

That last level is the one most companies forget. We automate things now so the system runs without a person involved in a lot of the steps, but you still have to observe what the system is doing, or you’ll be the last to know when it stops doing it.

The matrix

The matrix combines the three levers, three horizons, and three levels. That gives you 27 combinations. Every report you build should answer three questions at the outset. The first question asks which revenue lever the report is measuring. The second question asks over what time horizon. The third question asks at what level of the organization the answer needs to land.

If you can answer all three, the report will be meaningful.

Take a concrete example: a founder needs to raise $4.6M in 60 days, with a $200K minimum check size. The math says 23 investors, which works out to one investor every 3 days. That number is useful, but it doesn’t tell the founder how to pace the work. Most fundraises are back-loaded, with more closes in the second half of the window. The realistic expectation might be 4 commits in the first 3 weeks and 19 in the last 5. Once that description is on paper, the next question becomes how many investor conversations need to be in flight each week to produce that pace, which is the directional horizon at the individual level. Then the daily question becomes whether the right meetings are landing on the calendar at all, which is the operational horizon at the individual level. The strategic horizon was the $4.6M target.

The whole exercise describes the same business pursuing the same goal. It produces three completely different reports. Each one answers a different question, and each one matters at a different cadence.

From observation to control

The next piece of work after observability is controllability. The term observability isn’t mine, by the way. It comes from control theory, a branch of applied math that Rudolf Kalman developed in the late 1950s and early 1960s as computer systems were getting complex enough that engineers couldn’t see what was happening inside them anymore. Kalman’s question was whether, if you put a known input into a system, you could predict the output. If the answer was yes, the system was observable. Once a system was observable, it would be controllable.

A revenue system works the same way. Once you can observe the system across three levers, three horizons, and three levels, you can start designing controls. A control is an action with a predicted outcome. Most actions inside a sales or marketing org are taken without a predicted outcome attached. That is the gap controllability closes.

Controllability is the next article. For now, the test is just this: before you build any report, ask which lever, which horizon, and which level. If you can name all three, the report will be useful. If you can’t name them, you have a different job to do first. That job is defining what is being observed and why.

That is where revenue operations actually starts.

A few weeks ago I was at Stripe Sessions, listening to various versions of the future from the stage and the lunch tables. Most of my highlights took place in the hallways, but there was one stage moment that changed my perspective about what’s happening in the economy right now.

Patrick Collison put up a chart during his opening keynote: new business creation on Stripe, consistent and linear through most of 2025, then vertical in early 2026. He called it parabolic.

The capabilities of artificial intelligence models are moving at frontier pace. The absorption of those capabilities, and the financial gains that will accompany them, is slooow.

The companies holding the microphone at Sessions are not the broader economy. They are the leading edge. The Federal Reserve’s April synthesis puts US firm-level AI adoption at 18% by year-end 2025. MIT’s NANDA initiative reviewed 150 executive interviews, 350 employee surveys, and 300 public deployments, and found that 95% of enterprise GenAI pilots produced no measurable P&L impact. Somewhere between $30 and $40 billion has been spent and almost none of it has produced operating leverage. Wild!

The gap between capability and absorption is massive. Who’s going to usher in actual implementation? Whoever they are, they stand to make a lot of money – and by make, I mean create.

We saw this pattern with electrification

Pearl Street Station went live on September 4, 1882. Six 100-kilowatt dynamos in Lower Manhattan lit 400 lamps for 85 customers across a quarter-square-mile of the financial district. The promise of electrification had come to fruition but the ripples didn’t spread out at the speed of light (ba-dum-tss!). Forty years later, factories still hadn’t gotten much productivity out of it.

The reason was architectural. Pre-electric factories ran on line-shaft drive, so one steam engine sat in the basement, overhead shafts ran the length of the building, and belts dropped down to each machine. When electricity arrived, they pulled out the steam engine, replaced it with a giant electric motor, and left everything else in place. The result was almost no productivity gain.

The gains only came when factories rebuilt around unit drive. Every machine got its own motor, the overhead shafts came out, and the floor plan got redesigned around the new substrate. That retrofit didn’t happen until the 1920s.

The buildings just weren’t built for the technology that arrived in them.

Retrofit or renovate?

I’m gonna argue for renovation. Most firms are treating AI as something added on to an existing structure but that mentality is what generated the NANDA failure rate. The model may perform as advertised inside a pilot but then loses its edge when it has to coordinate with the rest of the business, which was never designed to receive what the model is producing. See what I mean?

Replatforming is the work of renovation, and renovation is six jobs running in parallel: a new data architecture, a new ontology, a new process design, a new set of decision rights, a new org structure, and a new equation for what a single employee actually costs once compute lives on the headcount line. You map what’s existing, redesign it, do the change management, and only then do you plug in the agents.

The NANDA study also separated success rates by who did the work. External partners succeeded at roughly twice the rate of internal builds. Renovation sits so far outside the muscle memory of an operating firm that the ones doing it alone produce most of the failures.

Why the frontier can’t bridge its own gap

Sam Altman appeared at Sessions in conversation with John Collison and named OpenAI’s role plainly. He wants OpenAI to be infrastructure, a utility that other people build on top of.

He has tried to scale OpenAI into the absorption layer anyway. The Forward Deployed Engineering organization, modeled on Palantir’s FDE, went from 2 engineers to roughly 40 in under a year. The instinct is on but the math still doesn’t work. A frontier-firm payroll can’t cover the surface area of every 200-person business in the country that needs the work done.

The skill set required to walk into one of those companies and make AI live inside it is rare: deep AI fluency, operator experience, change-management instinct, and the credibility to get past the company’s immune system. The people who have all four sit at the partner level of operating-experienced firms. I’m one of them. None of us would take a W-2 job at a frontier salary band, and the frontier cannot stand up a partner model fast enough to cover the absorption surface area of the global economy.

The frontier needs a partner category that doesn’t have a name yet.

Whoever solves pace owns the era

Two timelines are running in parallel. The absorption work itself will take decades. Electrification took forty years. Cloud is twenty years in and still half-deployed. The strategic window for category leadership is much shorter. The names attached to the cloud era (Andreessen, Bezos, Benioff) got attached during a window that lasted maybe five years.

The supply side of pace is going vertical. The humans doing the renovation work have AI doing the time-consuming parts alongside them, and the work compresses every quarter.

The demand side does not compress. Boards still meet quarterly. Procurement cycles still take six months. Cultural change still takes years. Conway’s Law still applies: any AI integration that respects the existing org structure will reproduce that org’s existing limitations, which means real rewiring requires reshaping the organization before the technology can do anything inside it. Substantial lift, but a lasting and extremely profitable one.

The binding constraint is the pace at which operating businesses can absorb structural change without breaking themselves.

The Collisons’ chart has a vertical axis labeled capability. The chart nobody has drawn yet has a vertical axis labeled absorption. It’s flat right now but it’ll become a hockey stick of its own.

The translator firms that exist in five years are being built right now, mostly in private, by operators who already see the gap.

A little over a year ago, a writer in Salt Lake City named wrote about a small project she’d been part of. Seven writers and a designer put together a collection of essays called CITY STATE and released it through a platform called Metalabel, which let them sell it in printed and digital form. The collection earned $732.33. After costs, and split eleven ways, it wasn’t much of a payday but the profitability wasn’t what interested me. A group of people made something together and owned the result together, utilizing a platform designed to make that kind of endeavor simple instead of expensive.

I’ve been following the person behind that platform ever since. co-founded Kickstarter and started Metalabel, and for the past year I’ve read/listened to most of what he publishes. His position: creative people have always worked at a structural disadvantage. The legal and financial scaffolding that protects an ordinary business wasn’t built with creatives in mind but building it yourself costs more than most creative work will ever return, so almost no one does. He decided that was a quandary worth solving at the level of the law.

What Colorado just did

As of June 2nd, 2026 a new type of legal entity officially exists. Governor Jared Polis signed a bill creating a new kind of company in Colorado: the Artist Corporation, or A-Corp. It’s a variant of the limited liability company, so it keeps everything an LLC already has but adds a few protections built specifically for creative work. In this new structure, the artists hold at least 51 percent of the voting power, permanently, and no operating agreement can sign that away. The company names an artistic mission in its founding documents, and that mission carries legal weight. The creative work associated with the company can never be transferred to outside investors, and if the company closes, the work returns to the people who made it. Economic rights can be separated from control, so an investor can share in what a project earns without getting a vote in how it gets made.

The A-Corp breaks the linkage between capital contribution and control. The idea is:

• Investors can provide capital.

• Artists can retain governance and creative ownership.

• Financial ownership and control rights can be separated more deliberately than in a standard corporation.

So in an A-Corp, the distinction between money, capital, and ownership becomes especially important:

• Money = cash contributed.

• Capital = the resources invested into the venture.

• Ownership = the bundle of economic and governance rights defined by the A-Corp structure.

• Creative control can remain with artists even when outside capital is involved.

The first reaction most business people have is to shrug. “So what?” A competent lawyer can already write most of this into a custom LLC agreement. This is true. The catch is the price. Doing it properly could run ten to fifty thousand dollars in legal fees, which is an absurd thing to ask of a painter or a band or a group of writers whose project might clear a few hundred dollars. When close to 40 percent of working artists make under twenty thousand dollars a year from their craft, a five-figure legal bill is not a real option. The protections that an A-Corp established have technically always been available but they were priced out of reach of the exact people who needed them most.

What the A-Corp does is make those protections the default – you get them by filling out a form rather than retaining a firm. The same move worked once already. Everything a public benefit corporation does was possible before 2013 with the right articles of incorporation. What changed was that mission-first ownership became something you could choose off the shelf, and within a decade it spread to thirty-six states. Strickler says six other states are already drafting their own versions of Colorado’s new law.

Services firms are intellectual assets

Equity is only worth holding if there is something durable to own. A firm that holds nothing but contracts for billable work has nothing to share except a slice of the treadmill. The A-Corp forces a services firm to ask: what is our durable value, who actually built it and how can we maximize its earning potential?

For us the answer is the engine. The engine is our repeatable process for building revenue systems, the frameworks those processes are built on and the judgment that gets sharper with every engagement. The engine is the asset with lasting worth. It’s ownable, rentable and yes, contributable as capital.

Levver’s engine is a creative asset.

Levver contributes intellectual capital in exactly the same way a software founder contributes source code or a designer contributes a methodology.

There are two kinds of making inside any business and they are traditionally valued unequally. There’s creative work: building software, designing a chair, defining a framework, and then there’s operations work: getting the creations into the hands of people who will benefit from them.

We have maintained since the day we started that this division is unnecessary. Designing and running a revenue system is a creative act in itself. It takes the same course as any other kind of making. You begin with a formless idea, you find a medium to express that idea, and you build a structure that did not exist before. You build an asset.

The A-Corp provides a legal structure that can recognize that asset as a basis for ownership.

The A-Corp structure goes beyond valuing and protecting the work of artists.

Underneath the specifics, the A-Corp treats creative contribution as a first-class basis for ownership and protects the creator’s authority over the thing they create.

Traditional business structures tend to concentrate ownership and control around financial capital. Creative contributors can receive ownership, but their authority often becomes negotiable once outside capital enters the picture. The A-Corp says the people who create the value can hold the value, and it builds the infrastructure to make that real and enforceable.

Founders and operators can become co-owners of a venture without reducing either side to a contractor.

This is the leading edge of creativity in business formation. From my view, professional services firms are sitting right on that front line. Think about what a firm like Levver actually is. We own no factory, no inventory, no equipment that matters. What we have is a way of thinking and the people who can do the thinking. Our entire offering is intellectual. In that sense we have far more in common with a band or a film crew than with a manufacturer.

If I want to value the contribution of creative work equally with the contribution of capital, in regards to ownership and control of a cash engine how can I align the interests of both contributors?

In the model I’m imagining, the founder contributes the creative work and retains control of it. Levver contributes the operating engine: go-to-market strategy, pricing, pipeline development, and the systems that turn an idea into a sustainable company. We receive ownership for that contribution, and the founder receives ownership for theirs.

The premise is simple: both forms of work are assets. Building the thing and building the revenue system that allows it to survive are equally valuable contributions.

AI makes this increasingly important

As AI increases the leverage of individual creators, the bottleneck shifts from creation to commercialization. That’s exactly where firms like Levver operate.

The number of people it takes to build a real business keeps shrinking. In May, Stripe published an analysis of the companies formed through its incorporation product. Solo founders now account for 63 percent of the C corps formed through Atlas, an all-time high. The reason is not mysterious. As AI tools let one person design a product, launch it, handle support, and keep improving it, the headcount required to stand up a serious company has zealously declined. The best of these founders are not earning rounding-error money, either. Four years ago the top tenth of solo founders made about 34 times the median in their first six months. Last year that figure was 61 times. The ceiling on what a very small team can build has moved up sharply.

The same data shows where going solo tops out. Past the two-year mark, founding teams pull ahead of even the strongest soloists, and the soloists who keep pace are usually the ones who extend themselves through partners, advisors, and networks rather than carrying everything alone. Put those two findings next to each other and it’s pretty obvious: AI has made it possible for one creative person to handle the product side of a business almost entirely on their own. What they still lack, and what still separates the ones who break out, is the other half of the work: the engine that finds customers, sets prices, and turns a good product into a thriving company.

That empty co-founder seat is just waiting for professional services firms who know operations. The larger the business one person can now build, the more value rides on getting that second seat filled, and the more it matters to have a clean and fair way to share ownership of what the partnership creates. Right on cue, enter the A-Corp.

The structure I keep coming back to

The model I keep thinking about is a holding company that incubates a portfolio of independent ventures rather than operating as a single firm. Each venture is built around a strong idea and the person capable of bringing it to life, so could be a software product, a book, a branding methodology, a framework distilled from years of experience, or something else entirely. The category matters less than the existence of a creative idea with enough clout to warrant building a business around.

The A-Corp makes this structure repeatable. Instead of negotiating custom arrangements every time, ownership can be issued for contribution rather than cash. Economic participation can be separated from control, allowing collaborators and the holding company to share in a venture’s success without taking creative authority away from its founder. If a venture ends, the underlying creative work returns to its creator.

What makes the model so interesting is that it was designed to protect artists, but the same protections apply cleanly to software founders, strategists, designers, and other creators. It allows a firm like ours to participate as a genuine partner rather than a vendor selling services.

Geography is not a meaningful obstacle. Like Delaware corporations, A-Corps can be formed in Colorado regardless of where the founders live, making the structure compatible with distributed teams and partnerships.

What it costs to mean it

The most significant constraint is also the feature that makes the model work: the creator must retain majority control. A holding company can own economics, but not command. Most holding companies would see that as a limitation. I see it as the alignment mechanism. The venture succeeds or fails under the leadership of the person whose vision created it, and the law makes that relationship durable.

There are still some unanswered questions. How should ownership be allocated when multiple creators jointly found a venture? How does a portfolio of A-Corps behave from a tax and entity-structure perspective? And because the law is new, there’s little precedent for applying it to consulting, strategy, or other forms of knowledge work. I believe it fits, but that assumption hasn’t been tested yet.

Those uncertainties don’t dissuade me though. A legal structure created for painters and musicians just so happens to describe a firm I have wanted to build for years.

Imagine a company that looks like the people inside it, and grows in proportion to what they bring.

It could show up as a single venture or a portfolio of them. It could grow into an ecosystem of creators, operators, and investors whose interests are aligned from the beginning instead of negotiated after the fact.

Maybe I’m in the minority – and I’m always okay with that – but I think the A-Corp pushes the horizon line toward a future where ownership can more accurately reflect contribution, where creators and operators each bring a real asset to the table, and share in the upside of what they build together. It’s a whole new way to create.

We bought our first Clay license on AppSumo on July 10, 2023. Hilariously, by the end of it’s AppSumo run, it had just 3.5 Tacos.

Back then Clay started as an “AI Relationship Manager” that could write emails to engaged contacts for you. Cool. There was no hype cycle or LinkedIn influencers posting waterfall tutorials, no “Clay” as shorthand for outbound enrichment. We saw what the platform could become before most people understood what it was: an orchestration layer. We saw a place where you could build the logic that sat between raw data and usable intelligence.

Since then we’ve built Clay orchestration for 35+ clients across every use case the platform supports: inbound enrichment, outbound list building, account scoring, CRM hygiene, and signal detection. We know this product so well we can hear when something’s off by how a table runs, the same way a mechanic knows something’s off by the sound of an engine.

So when Clay rolled out a new pricing model in March 2026 and announced it as a win for customers, we didn’t take the press release at face value. We did what we always do. We read the documentation (all of it) – the pricing page, the internal memo they published, the legacy plan docs, the Actions and Data Credits explainers, the AI pricing breakdowns, the community threads, and the old blog posts about previous pricing versions. Many hours of research across every public source we could find.

What we found wasn’t a simple price cut. It was a fundamental restructuring of how Clay captures revenue. And depending on how you use the platform, it might save you money or it might cost you more. The difference comes down to a question most users haven’t thought to ask.

The old system was elegant but exploitable

Once Clay turned into the data platform it is today, the original pricing model was beautiful in its simplicity. You bought a plan, it came with credits and the credits got consumed when you did things like enrichments, AI prompts, provider calls, or workflow steps. There was one meter for everything.

That simplicity made Clay easy to adopt and easy to explain but it also created a specific kind of user behavior that Clay almost certainly did not intend to subsidize forever.

Power users figured out the game quickly. Test on small batches before running full tables, order your waterfall providers from cheapest to most expensive, filter companies before enriching contacts, use conditional logic so expensive steps only fire when prerequisites are met, run free or cheap steps before paid fallbacks, disable auto-runs until workflows are locked, set AI spend limits and watch the dashboard like a hawk.

These weren’t hacks. Clay’s own documentation taught some of these tactics. They were rational responses to a system that rewarded careful orchestration.

But here’s the thing that matters: the most sophisticated users took this further. They brought their own API keys. They routed their own data providers through Clay’s HTTP request actions. They used Clay as a pure orchestration shell, consuming credits for the workflow logic while paying their vendors directly for the data.

From the user’s perspective, this was brilliant. It meant maximum flexibility, with minimum spend.

From Clay’s perspective, this was a leak. Those users were getting the full value of Clay’s platform: the routing, the conditional logic, the table structure, the integrations, plus the AI layer, and Clay was barely capturing any of the economics. The platform work was being consumed without being properly metered.

Two meters where there used to be one

The new model splits that single credit system into two distinct meters.

Actions measure platform usage. Every time Clay does work on your behalf, whether that’s running an enrichment, calling a provider, executing an AI prompt, or processing a workflow step, it costs Actions. Actions reset every billing cycle.

Data Credits measure data purchases. When you buy enrichment data through Clay’s marketplace of 150+ providers, you spend Data Credits. These can roll over within plan-specific limits.

The critical distinction: if you bring your own API keys and route your own vendors through Clay, you might reduce your Data Credit spend but Clay still charges you Actions for the platform work. The orchestration layer is no longer free.

Clay says data prices dropped 50 to 90 percent on many enrichments under the new system. They also say 90 percent of customers won’t hit their Action limits. Both of these claims may be true. But observe the math for a second.

If data is dramatically cheaper and most people won’t exhaust their Actions, where does the new revenue come from?

It comes from the users who were previously getting a free ride on the orchestration layer. The API-heavy power users – agencies running Clay as a workflow shell and builders who used Clay’s infrastructure while paying Clay almost nothing for it.

Who Wins, Who Loses, and Why

The community reaction tells the story pretty clearly. In Clay’s own public announcement thread the sentiment splits along predictable lines.

Users who buy most of their data natively through Clay’s marketplace are likely seeing real savings. They’ve got cheaper enrichments, access to advanced features at a lower price point than the old Pro tier (the new Growth plan runs $495 a month), and a more transparent breakdown of where their money goes. If you’re a revenue team that runs standard enrichment workflows and buys data through Clay’s built-in providers, the new model probably is cheaper. Clay isn’t lying about that.

But the power users noticed something else. One community member pointed out that HTTP function calls actually cost more under the new system than they did before. Another noted that if a team uses a heavy volume of API and HTTP calls, staying on the legacy plan is the better deal. A small-business user said Clay’s new pricing was pushing people like them out of the platform entirely.

These aren’t contradictions. They’re the natural result of a pricing model that now charges for two different kinds of value instead of one. If your usage skews toward data purchases, you benefit. If your usage skews toward orchestration, you pay more for what used to be bundled in.

The Growth tier at $495 a month does open up advanced features that previously required the more expensive Pro plan. That’s a legitimate improvement for mid-market teams who want access to better tooling without the old price tag. But the Actions meter means that high-volume builders, the users who were squeezing the most value out of Clay’s flexibility, now face a new line item they didn’t have before.

What this is really about

Strip away the pricing details and a clearer picture emerges. Clay is maturing from a flexible, credit-based enrichment tool into a GTM operating layer with infrastructure pricing to match.

The old model was growth-stage pricing. Simple, generous, designed to get users in and let them explore. That kind of pricing builds love and loyalty and market share. It’s also the kind of pricing that leaks money as the most sophisticated users figure out how to extract maximum value at minimum cost.

The new model is scale-stage pricing. It’s more complex, more precise, designed to capture revenue proportional to the value Clay’s platform actually delivers. When a customer uses Clay’s orchestration layer, Clay now gets paid for it regardless of where the data comes from. That’s a structurally better business.

If you’ve watched how SaaS platforms evolve, this pattern is familiar. Usage-based models almost always start simple: one meter, one credit, one unit of consumption. As the company scales and the user base diversifies, the single meter stops reflecting reality. Some users consume a lot of one thing and almost nothing of another. The company realizes certain high-value behaviors aren’t being monetized. So they add meters. They separate data from compute, or storage from queries, or seats from usage.

It’s the same physics. Clay just made the transition more visible than most.

The Broader Lesson

We still think Clay is one of the most important tools in the modern GTM stack. The platform’s ability to orchestrate data waterfalls across dozens of providers, apply AI-driven research and scoring, and sync clean data into systems is genuinely hard to replicate. We will keep building in Clay and we will keep recommending it to clients where it fits.

This pricing shift is worth understanding because it reveals something about how revenue systems mature, and that’s a topic we think about constantly.

Every platform has a revenue architecture. That architecture creates incentives. Users respond to those incentives rationally. Over time, some of those rational responses become optimization patterns that work against the platform’s economics. The platform notices. The platform restructures.

Clay’s old model incentivized users to minimize data spend and maximize orchestration usage. The new model re-prices those behaviors. The optimization playbook that worked for three years just changed.

If you’re a Clay user, the practical move is straightforward. Audit your current usage. Figure out how much of your spend is data purchases versus orchestration work. Model what your bill looks like under the new system before you migrate off a legacy plan. If you’re data-heavy, you’ll probably save money. If you’re orchestration-heavy, run the numbers carefully.

If you’re a revenue leader thinking about GTM tooling more broadly, the lesson is bigger. The platform you build on will eventually restructure its economics around the value you extract. That’s just how systems work. The best time to understand the incentives baked into your tools is before they change.

We spent many hours reading the fine print so you could spend 10 minutes reading this. If it saves you from a surprise on your next invoice, it was worth it.

David Ogilvy built a career on a single conviction: advertising exists to sell something, and you should be able to prove that it did. He came up writing mail-order copy, and he never stopped saying it was the best training he ever got. Direct response was the one corner of the business where results were not a matter of taste. You mailed the piece, counted the coupons that came back and the number was the number. No amount of creative self-regard could argue with it. He spent the back half of his career frustrated that the rest of the industry treated this as beneath them, when it was the only part of advertising that told the truth.

What Ogilvy wanted was an honest line between the thing you did and the result it produced. In his era you could draw that line, because the system was small enough to see end to end. There was one catalogue, one offer and one traceable action. The feedback was slow, and the data was thin, but it was real. When the orders came in, you knew what had produced them.

Digital advertising was supposed to be that dream realized at a scale Ogilvy never could have imagined. Not just the coupon, but the whole journey. Every ad seen, every link clicked, every path a stranger took from a first impression to a purchase, captured and counted and assembled into a complete picture of cause and effect. The promise was precision. After a century of guessing which half of the budget was wasted, we would finally know.

We did not finally know. We built something far more sophisticated than Ogilvy’s coupon and ended up further from the truth than he ever was.

Promises, promises

The discipline that was supposed to deliver Ogilvy’s dream is multi-touch attribution. The idea is straightforward enough to explain at a dinner party. A customer rarely buys after a single interaction. She sees a video, ignores it. Sees a display ad three days later, ignores that too. Searches your category, clicks a competitor, comes back a week later through a paid search ad, signs up for a newsletter, and converts a month after that off an email. Multi-touch attribution promises to reconstruct that whole sequence and assign credit fairly across every touch, so you know what actually moved her along the decision path.

It’s a beautiful idea but in practice it’s closer to astrology than accounting.

The first problem is that the data telling you this story is not yours. It’s computed and reported by the platforms that sold you the advertising in the first place. The same company that takes your money also grades its own performance and hands you the report card.

Meta tells you how many conversions Meta drove. Google tells you how many conversions Google drove. Each of them, asked about the same customer, will happily claim her. Run the numbers across all your channels and you routinely find you have paid for more advertising than your actual sales. Everyone takes credit. Nobody is lying, exactly. They are each answering a slightly different question, using a model they designed, measuring a window they chose, and they are under no obligation to reconcile their answer with anyone else’s.

This is not a story about bad vendors or sloppy implementation. The walled gardens don’t share data with each other because sharing it would dissolve the thing that makes each of them indispensable. Apple’s tracking changes in 2021 didn’t break attribution because Apple was careless; they broke it because the cross-platform visibility attribution depended on was borrowed, not owned, and it could be revoked the moment a more powerful player decided its own interests ran the other way. The precision was conditional and the values were not aligned.

The house computes the odds

Most complaints about attribution end with a wish that the platforms would behave better but that doesn’t dig deep enough. The platforms are not behaving badly. They are behaving rationally and in their own best interests.

Picture what it would mean for a platform to give you genuinely clean attribution. You would learn, with confidence, which of your spend produced results and which produced nothing. You would cut “the nothing”. Your budget would shrink, or move to wherever the truth pointed, which might be away from that platform entirely. Clean attribution is, from the platform’s perspective, a tool for helping you spend less with them. No rational business builds that tool and points it at its own revenue.

What a platform wants instead is for you to keep playing. Not to lose — a player who only loses eventually walks away – but to win just often enough, that you can never quite tell whether the wins came from your skill, the platform’s algorithm, or the simple fact that you kept feeding the machine. The optimal state for the house is a player who is convinced the game is winnable and can never prove whether they are winning. That player spends forever.

This is the attribution casino. The fog is the product. The entire industry of attribution software benefits from the confusion, selling maps to a floor that the house keeps rearranging. The platforms just want to keep you in the building, looking for the exit, certain that the right tool is one purchase away.

None of this requires a conspiracy. It only requires that everyone act in their own interest, which is the one thing you can always count on a market to do. That is what makes it durable. You can’t appeal to anyone’s conscience, because no one is doing anything wrong. The system is working exactly as designed. It’s just not designed for you.

AI is the casino’s final form

For all its opacity, the old casino at least let you see the floor. Google built an empire on paid search, but it ran alongside an organic search product that people had reasons to trust, and that tension forced a certain amount of visibility into the open. You could see the search results. You could track your keyword rankings, watch your position move, measure click-through, study the volume on a term and decide whether to fight for it. The instrumentation was imperfect and the game was rigged in the house’s favor, but there was a game board, and you could study it. You could become a better player.

Watch what happens to that game board when discovery moves into AI.

Start with what’s already here, because this is no longer a forecast. In February 2026, OpenAI began placing ads inside ChatGPT — sponsored units rendered beneath the AI’s answer, matched not to keywords but to the context of your conversation, sold at premiums that dwarf ordinary display because the intent is so concentrated. By spring, any advertiser in the United States could buy them through a self-serve dashboard with no minimum spend. Google moved the same direction faster: ads now appear inside a quarter of its AI-generated answers, up from roughly one in twenty a year earlier. The migration of paid advertising into the answer itself has already happened. We are simply early in it.

Now follow where it goes. Optimizing for AI answers — answer engine optimization — is becoming the new search marketing. But the equivalent of the old game board doesn’t exist. There is no ranking report for an AI answer. There is no position to track, no search volume to study, no result page to inspect and reverse-engineer. The answer simply appears, fully formed, and you have no way to audit why it recommended a competitor instead of you, or whether the recommendation was earned or bought. You can produce content and hope the model references it but you can’t see whether it worked, because the surface where it would have shown is a black box with no dial on the front.

Then there’s the part the structural logic makes inevitable, whatever any individual company says today. The economics of frontier AI are brutal — the leading labs are burning capital at a rate that demands a revenue engine, and advertising is the most profitable engine ever built. Forecasts already put AI search ad spend on a path from around a billion dollars to roughly twenty-six billion within a few years. Not every player will go there. Perplexity, tellingly, abandoned advertising entirely in early 2026 and is betting it can win as the trustworthy, ad-free alternative — a bet that is itself an admission that ads and trust pull against each other. But that is a bet by a challenger trying to differentiate, made against the full gravity of the market. The dominant economics point the other way.

What makes the AI casino even worse is that you now pay to enter. In the old model, your attention was the product and you paid nothing at the door. In the AI model, you pay for the answer — in subscriptions, tokens, the metered cost of every query — and your attention is still being sold. The casino now charges a cover and takes a cut of every hand. Wait! There’s one more turn of the screw. OpenAI has said it will launch its own multi-touch attribution product, a tool to show advertisers the customer journey across ChatGPT and beyond. So the same logic that made platform-reported attribution untrustworthy for fifteen years is being rebuilt, from scratch, inside the most opaque discovery system anyone has yet devised, but not to worry, it’ll be sold back to you as clarity.

What to do with this

I’d like to end with a playbook but I don’t have one, and I’m suspicious of anyone who does.

The honest position is that the game is moving faster than our ability to map it. The old attribution game took fifteen years to understand, and we never really got it, we just learned to live with the lies.

So this is not a piece with a prescription. It’s a piece with a question, and the question is the most valuable instrument you can carry into the next decade of marketing. Every time someone hands you an attribution number ask the only question that matters: who computed this, and what did they want when they did?

That question won’t necessarily lead to clarity but it replaces false precision with earned skepticism, and in a casino, the skeptical player is the only one who keeps their shirt. There isn’t a lack of data. You’re drowning in data. The danger is that you will mistake the house’s scorekeeping for the truth.

Ogilvy wanted to know, plainly and provably, what worked. We built a machine of staggering sophistication, pointed it at his exact question, then engineered it deliberately, rationally, profitably, to make sure the answer would always stay just out of reach.

Growing up in Minnesota in the 1990s, in a sports family, meant a lot of car rides tuned to KFAN sports talk radio. The hosts were celebrities back then, giants of the local public sphere. My Grandpa Jack was a die-hard Vikings, Twins, and Timberwolves fan despite living in Western Wisconsin, and he loved his fandom so much that he fell asleep every night to KFAN coming through low and fuzzy on a red LED Casio clock radio with faux wood paneling. It stayed on until morning.

The king of the scene was Dan “The Common Man” Cole. Common, to everybody. We loved him for the ridiculous bits, the one-of-a-kind personality, and a genuine love of our local teams. Minnesota didn’t win much, so the material was plentiful. The bits wrote themselves.

One of the mainstays of his show was the monikers. Common never took a call from Bill in Woodbury. It was Billy the Bass Guy who owns a Ranger boat, frequents Cabella’s, and loves to chat weekend fishing on one of Minnesota’s 10,000 lakes. Everyone who called in was a guy or a gal, self-proclaimed, and when a familiar one came through the speakers your ears perked up, because Common was about to razz them into a good rant and you were in for a treat.

The best moniker of all time belonged to my uncle Jeff: the Anti-Anti-Hunting Guy.

Jeff called in whenever Common or a guest brought an anti-hunting take to the air, and he had the standing to do it. He grew up waking before school to trudge into the fields and swamps of Wisconsin, and then he never stopped. He has lived his whole adult life as an outdoorsman, hunting and fishing through every season, harvesting only what he was going to eat, deeply passionate about nature and the animals in it. The moniker was a bit, but it was a truthful bit with an agenda. He wanted people to hear what hunting and fishing really were from someone who lived them and loved them, instead of from the caricature.

I’m telling you this for two reasons. The first is that it’s fun to dig back through these nostalgic memories. The second is that the moniker market never closed. It went national.

The machine went national

Sports talk radio was the hot-take economy before anyone had the phrase. Takes were the currency. Razzing was the content. Conflict was the format. Common understood all of it decades before an algorithm did.

But two things were different then. Everyone was in on the bit and the stakes were the Vikings.

Then media got democratized. Podcasts handed every ordinary Joe and Jane a platform, and the distribution algorithms learned what talk radio always knew: division holds attention better than agreement does. The architecture survived. The wink did not. And the stakes climbed from whether the Twins could hit left-handed pitching to AI, data centers, employment, IP law, and whether we keep control of the most powerful technology humans have ever built.

Same machine, different soul. On the questions that matter most, nearly everyone with a microphone has indexed hard to Pro or Anti and made it their whole bit, because that’s what the platforms pay for. The people offering genuinely nuanced middle-ground positions get dismissed as biased before anyone listens. The conversation we need is the one the format ignores.

Anti-anti is not pro

Look closely at what my uncle’s moniker actually says, because there’s something hiding in it.

In a logic class, a double negative collapses into a positive. In real life it doesn’t. Anti-Anti-Hunting never meant pro-killing-animals. Jeff was against the caricature, the cartoon version of hunting that the anti-hunting position needs in order to exist. What the double negative defends is the real, conditional thing: conservation, self-sufficiency, eating what you harvest, the grit and compassion that come from a life spent in the field. The caricature flattens all of that. The anti-anti position restores it.

So call me the Anti-Anti-AI Guy.

That does not make me a booster. The accelerationists who treat every concern as a moral panic have a caricature problem of their own. The Anti-Anti-AI position holds three things at once: the technology is real, the risks are real, and the truth about both belongs to the people who practice it.

Priced out, not argued out

The AI conversation has sorted itself into two camps with names now. Boomers, or accelerationists, on one pole, doomers on the other. The writer Casey Mock described the two brands as “superficially opposed, structurally identical,” and he’s right. Both are selling confidence because confidence is what the attention economy buys. The catastrophe fits in a tweet. The utopia fits in a tweet. The conditional truth needs three paragraphs and your patience and the format doesn’t sell patience.

The squeeze on the middle is structural and it’s measurable. Researchers at NYU who studied more than half a million social media posts found that each additional moral-emotional word in a post increased its spread by roughly twenty percent. Philip Tetlock spent decades scoring expert predictions and found that the confident hedgehogs got the airtime while the self-critical foxes got the accuracy, with fame and accuracy running in opposite directions. The middle didn’t lose the argument. It lost the format.

Sometimes the pole is right

Honesty requires a concession here, because the middle isn’t automatically virtuous.

Sometimes the sharp pole is correct and the reasonable middle is the trap. The tobacco industry understood this better than anyone and weaponized it for decades. “Doubt is our product,” reads the infamous internal memo. Manufactured balance, both-sides coverage, endless calls for more research, all of it engineered to delay a verdict everyone already knew. The anti-tobacco hardliners were right. The reasonable moderates were being played.

So the middle has to be earned. There’s a difference between a lean and a hedge. A lean is a position built through contact with the thing itself, and you can say out loud what would change your mind. A hedge is a place to hide. “It depends” can be honest calibration or it can be cowardice wearing calibration’s clothes, and the test is whether there’s practice behind it. Jeff’s middle was never a hedge. He had sharp opinions and decades of mornings in the swamp backing them up.

Where the value lives

Here is what the discourse misses while people are busy at the poles. Attention and value trade in two different markets. The poles capture the attention. The middle captures the value. Right now the mispricing is enormous.

I see it every day. At Levver we build revenue systems across many businesses, in many markets, industries, and verticals, and when you do that work you see the inside of the game. AI is very good, and continually getting better, at specific mechanisms: prospect research, document ingestion and data extraction, mining call transcripts for what actually happened in a deal, and dozens of other tasks that used to eat human hours. Build those efficiencies into the system so the humans can do what humans do best, and you’re running the strategy that has won every time a new technology has arrived. Reduce the cost of moving a deal through the pipeline and closing it, and the unit economics improve. Same story in ops workflows. Same story in customer success process and engagement. Find the places where improvement is meaningful and AI can genuinely help, and that’s where the value is. Not everywhere – there.

Notice the pattern: it works here, it fails there, and here’s how we test it. That takeaway isn’t trendy or trending but it’s the only one worth money.

And notice who holds it. The people who use these tools every day trend toward cautious optimism, while the wildest takes cluster among people who barely use them at all. The poles are full of spectators. The middle is full of operators. Jeff knew the truth about hunting because he practiced it, and the same rule applies here. It’s why my lean is openly optimistic: used correctly, AI is a tool set and a capability that humans have never had before, and that’s amazing. That lean was earned deployment by deployment, including the ones that didn’t work.

The lines are open

When I first started writing this, I ended on a question. Who bridges the divide? Who’s the Anti-Anti-AI guy or gal who sees the promise of this technology and the vision for how it could make humanity better, but takes the downsides seriously and thinks hard about ushering it into society in a way that protects human beings and this beautiful, flawed experience of being here?

I’ve stopped waiting for that person to show up on a podcast. The divide won’t be bridged from a pole, and the Anti-Anti-AI guy was never going to be a pundit. He’s the operator who already lives in the middle, doing the conditional work the format can’t sell. There are a lot more of us out here than the algorithm lets you hear.

The lines are open. The moniker’s available.