Outcome-As-A-Service is crushing the SaaS based GTM model

The subscription model isn't dying. But it is, quietly and unevenly, beginning to lose on a new dimension.

The companies that recognize outcomes as a competitive variable before their competitors do won't just win more deals but also define what the entire category charges for and that's a far more durable advantage than any feature on a roadmap.

The Subscription Is Not Dying. But It's Losing.

Here's a scenario worth sitting with. Two companies sell AI-powered customer support automation. Company A charges you a platform license - $180,000 a year, fixed, regardless of outcomes. Company B says: "We'll reduce your support costs by 40%. If we don't hit that number, you don't pay."

Now ask yourself - not as a technologist, but as a CFO who just got grilled about AI ROI in the last board meeting - which vendor do you call back?  
‍

This isn't hypothetical but actual competitive terrain forming underneath the software industry right now, and most companies haven't fully reckoned with what it means for how they price, how they sell, and ultimately, how they survive.

The Shift Nobody Is Talking About Loudly Enough

Most of the AI conversation in 2025 is about capability. Which model is better. Whether agents can replace knowledge workers. How fast things are moving. These are interesting debates - but they're not the debate that will determine which software companies are still standing five years from now.

The debate that matters is a pricing debate.

For three decades, SaaS built an extraordinary industry on a simple idea: charge per seat, expand through adoption, grow net revenue retention. It worked because software was a tool, and tools were priced by how many people used them. The unit of value was the user.

Agentic AI quietly broke that logic. When a single AI agent can do the work of five support reps, ten data analysts, or an entire onboarding team, the per-seat model stops being a pricing strategy and starts looking like a legacy assumption. As the a16z framing of this shift puts it plainly: software is becoming labor. And labor has never been priced by the seat. Labor is priced by output.  

‍

This is what "Outcome as a Service" actually means - not a rebrand, not a marketing pivot, but a fundamental repricing of where value is captured in the software stack. And to be precise: this isn't about SaaS dying.  

The subscription model isn't going anywhere. What's changing is that outcomes are becoming a competitive variable - a new dimension on which vendors will increasingly win or lose - layered on top of the existing commercial landscape.

Why Is SaaS Pricing Being Disrupted by Outcome-Based Models in AI Companies?

The per-seat pricing model was built for a world where software was a tool used by people. Agentic AI broke the unit of value: when a single agent can do the work of ten support reps, per-seat pricing stops being a commercial strategy and starts looking like a legacy assumption. Companies moving to performance-based pricing in AI are not making a marketing decision — they are responding to buyers who have stopped tolerating the gap between what the vendor promises and what the contract guarantees. Outcome-based pricing collapses that gap by design. It forces the vendor to price on what the customer actually cares about: results, not access.

What Zendesk and Salesforce Are Actually Signaling

The companies worth watching aren't announcing this shift loudly. They're quietly pricing around it.
‍

Zendesk was first among major CX platforms to introduce outcome-based pricing tied to automated resolutions - cases fully closed by AI, zero human intervention. Not deflection. Resolution. They paired it with a real-time automation dashboard, because if you're charging for outcomes, customers need to verify them. That dashboard is what makes the pricing model credible.

Salesforce followed a parallel path with Agentforce's Flex Credit - $0.10 per agent action - structured so enterprises can actively swap human licenses for digital labor credits. The language matters: digital labor, not software, not automation. They're signaling that what they sell is no longer platform access. It's capacity for work to get done.

‍

KEY INSIGHT: 40% of buyers already cite seat reduction as their primary lever to decrease software spending, according to BCG. Vendors who don't get ahead of this are handing their customers a reason to churn.

Both of these are early-stage experiments. The attribution challenges are real, the measurement infrastructure is still being figured out, and the pricing mechanics will evolve. But the direction is unmistakable.

Why This Matters More Than Technical Superiority

This is the part that should make any founder or product leader genuinely uncomfortable.

You could build the best agentic AI system in the market - better recall rates, lower latency, more reliable reasoning than anything a competitor ships - and still lose. Not because your product is worse, but because your pricing model makes it structurally harder for a risk-averse CFO to say yes.

The data tells you exactly why this dynamic is so powerful right now:
‍

That's why outcome pricing is increasingly functioning as a trust signal. The willingness to tie your revenue to a result communicates something about your confidence in your own product that a fixed annual license simply cannot.

The Real Bottleneck: Attribution

Here's where the honest conversation gets harder.

Outcome-based pricing sounds compelling in a pitch deck until you try to actually implement it - at which point you run straight into the hardest problem in enterprise software: attribution. How do you prove that a specific business result was caused by your product, and not by a process change, a seasonal trend, or the five other tools running simultaneously in the same stack?

The attribution problem lives entirely in that 70%. It's not a technical challenge. It's an organizational one - and it requires baseline measurement before deployment, contractual clarity about what counts as an outcome, and a level of instrumentation that most software vendors have never had to build before.

Zendesk's automation dashboard exists for exactly this reason. Outcome pricing without measurement infrastructure isn't a pricing model. It's a promise with no mechanism for accountability. Buyers - especially the financially conservative CFOs who now control AI budgets - know the difference.

This is also why the transition will be gradual, not sudden. That's likely the template for how most of the market evolves: not an overnight switch, but a progressive shift where outcome components get layered on top of existing structures until they become the dominant value signal.  
‍

What Pricing Model Should AI Startups Use: SaaS Subscription or Outcome-Based Pricing?

The answer depends on how defensible the attribution is. Subscription pricing is lower risk and operationally simpler — it works well when the product is a platform or tool that augments human work rather than replacing a measurable business outcome. Outcome-based pricing is higher reward and higher commitment: it wins more deals with financially conservative buyers, builds trust through skin-in-the-game positioning, and can command premium pricing when the outcome is clearly defined. The practical template for most AI startups is a hybrid — a base subscription that reduces commercial risk, layered with an outcome component that grows as the attribution infrastructure matures. The companies that figure this out first will not just win more deals. They will set what the category charges for.

The GTM Implications Nobody Wants to Sit With

If outcomes become a competitive variable, the go-to-market motion has to change - and not in comfortable ways.  

Here's how the model actually shifts across three dimensions:
‍

The first shift is who you sell to. EY's research on agentic AI go-to-market strategy is explicit on this: buying decisions are migrating away from IT departments toward business function leaders - the CFO, the Chief Customer Officer, the Head of Operations. These buyers don't care about your API architecture or your model's benchmark scores. They care about a number on a spreadsheet and whether you can move it.

The second shift is what your sales team needs to know. The traditional enterprise software sales playbook - demo the features, negotiate the contract, expand by seat - starts to break down when the value you're selling is a result rather than a platform capability. EY describes this as moving from product-conversant selling to operational advising. That's not a training update. That's a different hire.

The third shift - and the one finance teams consistently miss - is revenue recognition. Under ASC 606, there's a meaningful legal distinction between selling access to a platform and selling a delivered outcome. Outcome-based contracts can change when and how revenue hits your income statement, your ability to forecast, and your revenue visibility quarter to quarter. Finance and legal need to be in the room from day one.

KEY INSIGHT: The buying criteria shift is already underway. EY's research confirms that enterprise software decisions are actively migrating from IT departments to business function leaders. If your sales motion still leads with a product demo, you are already behind.

The Companies That Figure This Out First Will Set the Category Standard
‍

McKinsey's data on AI high performers - the top 6% driving disproportionate returns - reveals one consistent pattern: they are three times more likely to fundamentally redesign their workflows around AI rather than layer it onto existing processes. BCG echoes this, finding that leading companies allocate more than 80% of their AI investment to reshaping business functions, not to incremental productivity gains.

The same logic applies to vendors. The companies that win in an outcome-competitive market won't necessarily have the best models. They'll be the ones willing to restructure their pricing, their customer success motion, their sales hiring, and their measurement infrastructure around a different definition of value delivered.

The urgency is real. Bessemer's Cloud 100 data shows the cohort crossed $1.1 trillion in total value for the first time - a 36% jump from the previous year. AI-native companies are reaching $100M ARR in 5.7 years on average versus 7.5 years for the broader list, with some breakouts hitting it in under four years. The window for incumbents to lead this shift rather than react to it is not indefinite.

KEY INSIGHT: The companies reaching scale fastest aren't winning on product alone. They're winning on how they frame, price, and prove value and that playbook is being rewritten right now.

‍