Buyer Guides · 6 MIN
Fixed-scope AI pilot vs time and materials: which should you choose?
When a fixed-scope AI pilot fits and when time and materials is the better contract. A plain comparison of risk, flexibility, and incentives for buyers.
Choose a fixed-scope pilot when the question is "does this work on our data" and the boundaries are clear. Choose time and materials when the work is exploratory, the requirements will change weekly, or you are extending an existing system. Many teams do both in sequence, which is why the strongest structure is often a fixed pilot followed by an ongoing monthly team.
- Fixed scope moves delivery risk to the supplier, and time and materials leaves it with you.
- Fixed scope works only when the goal, data, and pass bar are defined up front.
- Time and materials suits discovery, changing requirements, and ongoing product work.
- Neither model protects you from unclear goals. Both need a written definition of done.
- The common sequence is a fixed-scope pilot to reach a decision, then a recurring team to build.
- Nactore scopes each engagement to the team and the problem, and does not push one model for every buyer.
What is the real difference between the two models?
In a fixed-scope engagement, you agree on the outcome, the timeline, and the commercial terms before work starts. In time and materials, you pay for the effort spent, and the scope can move as you learn.
The difference is who holds the uncertainty. Under fixed scope, the supplier absorbs the risk that the work takes longer than expected. Under time and materials, you do. Each model also creates different incentives, which is worth understanding before you sign anything.
How do the two models compare for AI work?
| Factor | Fixed-scope pilot | Time and materials |
|---|---|---|
| Budget certainty | High, agreed up front | Lower, depends on effort |
| Flexibility to change direction | Limited, changes go through a scope update | High, adjust each sprint |
| Who carries overrun risk | The supplier | The buyer |
| Best for | A bounded question with a pass bar | Discovery, evolving products, ongoing work |
| Main failure mode | Scope written too loosely, then disputed | Drift with no decision point |
| Incentive to watch | Cutting corners to hit the date | Work expanding to fill the hours |
When does a fixed-scope pilot make sense?
Fixed scope fits when you can state the experiment in a paragraph. You know the workflow, you can provide the data, and you can define what a pass looks like. AI work has a particular reason to favor this early on, because the largest unknown is whether the approach works at all. A short, fixed experiment answers that cheaply.
It also helps internal approval. A defined scope, timeline, and decision gate is easier to take to a finance partner or a board than an open-ended engagement. If you want the scoping detail, read how to scope an AI pilot.
When is time and materials the better choice?
Time and materials wins when the work cannot be specified in advance. Examples include extending an AI feature across a mature product, working through a long tail of integrations, or building something where you expect to change your mind based on user feedback.
It is also the natural fit after a successful pilot. Once you know the approach works, the next phase is a stream of improvements, and a fixed scope would either be too rigid or constantly renegotiated.
What are the hidden risks of each model?
Both models fail in predictable ways. Knowing them lets you write protections into the agreement.
- Fixed scope, vague definition. If "done" is fuzzy, the fixed price invites disputes. Protect yourself with a written pass bar and a list of exclusions.
- Fixed scope, change friction. When you learn something mid-pilot, changing direction needs a scope update. Agree how changes are handled before you start.
- Time and materials, no decision gate. Without milestones, spending continues whether or not it works. Require a review at set intervals with explicit authority to stop.
- Time and materials, opaque effort. Ask for visibility into what the hours bought, tied to shipped work, not just timesheets.
Whichever model you choose, tie payment or review points to evidence, such as an eval score or a working release in your staging environment, not to activity.
How do you combine the two?
The sequence we see work best is a fixed-scope pilot first and a recurring team second. The pilot answers the go or no-go question with a hard end date. If the answer is yes, a monthly engineering pod continues the work with a steady team, a visible backlog, and a monthly review of what shipped.
This keeps the early risk bounded and the later work flexible. It also keeps continuity, because the people who learned your data in the pilot carry that knowledge into the build. For what the recurring phase includes, see what an AI engineering pod delivers monthly.
What should you ask any supplier before choosing?
- How do you define done? Look for a measurable bar, not a feature list.
- What happens if the pilot fails the bar? A trustworthy answer includes a written stop recommendation, not a pitch for more work.
- How are changes handled? Look for a clear, written change process.
- Who owns the code and eval sets? This should be you. Have your counsel review the terms.
- Who actually does the work? Ask for named engineers, not a bench.
Frequently asked questions
Is fixed scope always cheaper than time and materials?
Not necessarily. Fixed scope gives you certainty, not a lower number. A supplier may price in contingency for the risk they carry. Compare on the basis of what you get and how risk is shared.
Can a fixed-scope pilot change midway?
Yes, through an agreed change process. Small adjustments within the stated goal are normal. A change to the goal itself should update the scope document.
Which is better for a first AI project?
Usually a fixed-scope pilot, because the main question is whether the approach works. It bounds your downside and produces a decision.
What if we are not sure what we want built?
Then you may need a short discovery phase before a pilot. Do not force a fixed scope onto a problem you cannot yet describe.
Match the contract to the uncertainty
Use fixed scope to resolve the biggest unknown quickly, and time and materials, or a monthly team, once the path is clear.
Want this built for your team? Book a free 30-minute call.
Want to apply this to your business?
Book a free 30-minute call. We will tell you what we would do first.