Halfway through 2025, the Information Technology (IT)-Business Process Outsourcing (BPO) world finds itself at yet another crossroads, this time ushered in by the rise of agentic artificial intelligence (AI). As providers and enterprises tinker, test, and tease out what autonomous agents really mean for day-to-day operations, a deeper conversation is emerging around how we put a price tag on all this innovation.
The old yardsticks, tokens, Application Programming Interface (API) calls, seat licenses, even user counts, feel increasingly out of step when machines are the ones doing the heavy lifting. Instead, momentum is building toward execution-based pricing: models that charge for the tangible work an agent completes, the actions it takes, the tasks it closes, the workflows it runs, rather than the resources it consumes.
In other words, we’re moving from metering effort to metering outcomes, and that shift could redefine the economics of IT-BPO for the decade ahead.
Reach out to discuss this topic in depth.
Why execution-based pricing Is the next big thing
Picture this: instead of counting API calls, juggling seat licenses, or decoding token usage, you simply pay whenever a job is done, no more, no less. That’s the essence of execution-based pricing.
How it works
- Pay per completed task or workflow. A ticket gets resolved end-to-end, a transaction is processed, a data pipeline is migrated, each finished “execution” triggers a charge
- Straightforward budgeting. Because the price of each execution is agreed up front, finance teams can forecast spend by multiplying that rate by expected volumes. No surprises lurking in the fine print of usage metrics
- Full transparency. Tokens, API calls, or the number of autonomous agents involved can be hard to measure, so we stop measuring them. All that matters is the outcome you asked for
Getting it right starts with clear definitions
Before the first bot springs into action, map the workflow and its key steps. Whether it’s one agent scraping multiple systems or a string of micro-agents handing off tasks, everyone must agree what “done” looks like. Once the scope is crisp, each successful execution, regardless of how long it takes or how many tokens it consumes—becomes a billable event.
Where this model shines
Execution-based pricing thrives on workflows that are repeatable, well-defined, and carry stable Service Level Agreements (SLAs). Think password resets, invoice processing, or nightly data syncs. Providers are rewarded for cranking through more executions faster (as long as quality stays high), while buyers only pay for outcomes delivered. Efficiency gains on one side translate directly into cost predictability on the other, a genuine win-win.
As autonomous agents shoulder more of the heavy lifting, counting effort makes less and less sense. Counting outcomes, on the other hand, keeps everyone laser-focused on what matters most: work completed, and value delivered.
One-size-fits-all? Not anymore.
When autonomous agents step in, every “execution” tells a different story. Some close out in seconds, others demand deep-dive troubleshooting. That gap is pushing early adopters of agentic AI in IT-BPO to embrace tiered, complexity-based pricing, rewarding heavy-lift work properly instead of lumping everything into a flat fee. Here’s a quick tour of the three tiers we’re seeing in the wild:
- Simple executions: Think password resets, account unlocks, or any click-click-done routine. Steps are crystal-clear, run-books short, and decision-making is near-zero. Because they burn almost no resources, they sit at the lowest price point per execution
- Medium executions: Now we’re talking multi-step tasks with a bit of brainpower, say, resolving a software-install glitch or verifying a transaction against a couple of data sources. They take longer (roughly 15–30 minutes), need tier-2 know-how, and involve a few decision branches. Expect pricing around 1.5–2× the simple rate to reflect the added lift
- Complex executions: The heavyweight category: major incident triage, L3 support across systems, fraud-flag deep dives, anything unique, high-risk, or sprawling enough to pull senior experts into the mix. These jobs can stretch to hours or days and rarely follow a cookie-cutter script. Providers usually quote them at a premium per execution, or carve them out entirely for time-and-materials or project fees
By matching price to complexity, both sides win. Providers get fairly paid for the gnarly stuff, while buyers gain transparency and predictable budgets anchored to real business outcomes, not guesswork around tokens or API calls.
Why complexity tiers keep execution-based pricing honest
Not all “executions” are created equal. One might wrap up in five minutes; another could turn into a five-hour slog. Charging a flat fee for both? Bad math for providers, and a mismatch with client expectations. Complexity-based tiers solve that by letting simple jobs stay cheap while heavier lifts earn a premium.
Fair pay for fair effort
A typical spread might look like $5 for a simple task, $10 for a medium one, and $20 for the complex stuff, or whatever ratio suits your industry. The exact dollars vary, but the idea is constant: price mirrors effort.
Define the tiers up front
To avoid the “Was that really complex?” debate, both sides lock in what simple, medium, and complex mean before the contract ink dries. Think sample use-cases, expected handle times, or other clear benchmarks. Clarity today means fewer headaches tomorrow.
When the model bends
Every rule has its edge cases. Super-gnarly, unpredictable jobs that defy tidy definitions can break the per-execution mold. That’s where hybrids shine: charge per execution for the bulk of routine work, but switch to time-and-materials or fixed fees for true outliers.
The upshot? Most IT and BPO deals thrive on this tiered approach, keeping billing aligned with real effort while giving both provider and client a pricing model they can trust.
If you are considering adopting agentic AI solutions for your enterprise or your client, feel free to contact us for a Point of View (PoV) on the state and adoption of pricing models in this space.
If you found this blog interesting, check out our The Dichotomy Of Pricing In Cybersecurity Services | Blog – Everest Group, which delves deeper into another topic regarding pricing.
If you have any questions or want to discuss the evolution of pricing in more depth, please contact Rahul Rana ([email protected]) and Sahil Shah ([email protected]).
