Summer, 1888: George Eastman launches the Kodak camera with a deceptively simple business model, “You press the button, we do the rest.” The camera was almost beside the point. The money was in the film, a “consumption-based” roll that made the whole thing run. In the process, Eastman built a usage loop and monetized what happened inside it.

That’s the AI economy in a nutshell: The software is “the button,” and the tokens, inference calls, and workflow runs are “the rest.” Yet, most sales organizations are still building compensation plans like they’re selling cameras.

AI-native companies make up roughly 5% of QuotaPath’s customer base — and after analyzing hundreds of their comp plans, we found that the best are rewriting the rules and, in fact, modifying them at 3.5x the rate of the average SaaS company over the last 12 months. The most forward-thinking among them aren’t simply adjusting old plans but sometimes starting from scratch.

Here are some of their lessons — and what the rest of the market is about to learn the hard way.

In SaaS, paying 10% of ACV made sense when you kept 80 cents on the dollar. In AI, those economics are gone. Many AI companies operate today at 50–60% gross margin, where every token has a real cost and every inference call eats into it. 

Paying 10% on revenue when you’re keeping 55 cents is a very different business decision than paying 10% when you’re keeping 80. It’s not that commissions are wrong, but rather, the math underneath them has changed, and most plans haven’t caught up.

SaaS had a clean unit: seats. AI does not. Every other part of the chain has evolved on this, from CRM to billing, except for comp. 

And while 58% of AI companies still include a subscription component, both consumption-based and outcome-based pricing have grown meaningfully in the last six months (at 35% and 18% respectively). What’s more, 37% plan to change their pricing model within the next year.

When revenue shifts from signed seat count to usage, the focus of compensation necessarily shifts from a payroll decision to a strategic design problem.

Companies with <$1M ARR are 86% acquisition-driven. By the time they hit $20–50M, expansion is generating 38% of net-new ARR. AI-native companies treat expansion as a core driver and not just a “nice-to-have.” And though PLG as a term went out the window when OpenView went belly-up, almost every AI-native company is running a PLG play today, in which product-influenced dollars can account for ~90% of total revenue. 

This all makes sense, as there’s no other way to go from $0–40 million in revenue in six months without a viral tweet, as Eric Simons described in a Topline episode last June.

Most AI products require substantial human effort before customers see value. Companies are effectively paying to get customers on the product, even at the cost of near-term margin. This is especially true in deployment, where forward-deployed engineers, hands-on onboarding, and service-heavy implementations can really add up the hours. One AI vendor at QuotaPath had an onboarding roadmap with milestones totaling 6+ months!

This changes both what you can afford to pay your sellers and what you should ask them to optimize for.

Cassie Young (Partner at Primary Ventures) put it plainly when describing one of her portcos, 1Mind, on a forthcoming episode of Topline: They track customer count, not revenue dollars. Their CEO, Amanda Kahlow, learned from previous mistakes as founder of 6sense: In a land-and-expand model, the logo is the asset, and revenue follows.

The comp plans reflect it. Multiple AI-native companies in our base are paying explicit new-logo bonuses, a flat cash reward for each new account, regardless of deal size. One notable example — a larger AI customer — pays approximately $5,000 per new logo (!). In the SaaS era that number would’ve seemed eccentric, but in an expansion-first world, it’s just honest incentive design.

This surprised us most: AI-native companies are not only paying bonuses for closing deals quickly, but some are paying higher commission rates for deals that skip the proof of concept entirely.

Think about what that signals — POC used to be table stakes! 

A fast close on a smaller committed deal, followed by real usage, beats a slow close on a larger deal that takes six months to get into production. Effectively, it’s “get the logo in. Prove value. Expand. Repeat.”

This also explains why multi-year incentives have nearly disappeared from AI-native comp. Whereas traditionally 30–40% of SaaS plans offered a premium for two- or three-year commitments, in our AI-native cohort, we found only one company doing so. The reason is simple: Pricing is still in flux, and locking a customer into last year’s terms is a liability for both sides. 

My Spidey Sense thinks this will change rapidly next year as these AI-native companies raise their post-Seed or Series B rounds, and LTV comes to the forefront.

One company eliminated the higher commission rate for AE-sourced deals — a common incentive to reward hustle. Now, every deal pays the same rate, regardless of source. Another went from maintaining differentiated rates across their entire product portfolio (which was a nightmare to administer and easy to game) to a single, flat rate across everything they sell.

The underlying logic: Complexity pulls cognitive focus away from selling. If a rep is doing mental math about product mix and sourcing credit, they’re not thinking about the customer. In an expansion-driven motion, that’s a real cost.

AI-native companies aren’t ignoring expansion revenue — they’re just treating it differently than new logo ARR.

The most common structure: pay reps on expansion ARR within a defined window after close, but at a reduced rate or with a quota multiplier. One company applies a 0.8x multiplier, meaning $100K in expansion counts as $80K toward the number. Others use a lower commission rate outright.

The message to the rep is clear: Expansion matters, and it’s something we’ll pay for, but as a supplement to new logos, not a substitute.

If revenue grows through usage, comp has to follow. The real question is: Where in the process is your seller actually changing the outcome? 

Snowflake figured this out early: Reps carried consumption quotas, and comp extended beyond the deal to the usage that followed. 

That same principle is now showing up across AI-native companies, where they’re anchoring comp to the one or two moments that drive adoption, activation, or expansion, and not to every step in the funnel. The emerging structure: one component covering new business tied to committed spend, and a second covering what happens after — usage growth, adoption milestones, and expansion events.

Most sales organizations are still pricing cameras, while the AI economy runs on film.

Paying an AE like a traditional SaaS closer often means paying for the wrong thing: sometimes overpaying, sometimes underpaying, but mostly just rewarding behaviors that look like winning, while the real growth levers go unnoticed.

AI-native companies that are getting this right share one orientation: They’ve stopped asking “How do we structure commissions?” in favor of “What behaviors actually produce compounding value for our customers?” The comp plan is just the answer to that question, written in dollars.

That means paying for logos over deal size. Speed over contract length. Simplicity over complexity. And tying the payout to when a customer realizes value — not when they sign.

Because the camera was never the business. The film was. And your compensation plan should reflect that difference.

AJ Bruno is the co-founder and CEO of QuotaPath, a sales compensation platform that simplifies commission tracking for revenue teams. As a second-time founder, AJ has built and led GTM strategy for sales, lead generation, and account management teams, scaling from $0 to $25M in ARR and overseeing organizations with over 200 employees. AJ is also an active startup advisor and investor, and a co-host of the Topline podcast.

Become a Topline insider by joining our Slack channel.

We want to hear your feedback! Let us know your thoughts.