Comparison · Engagement model

Monthly retainer or fixed price for AI

Engaging an AI partner with a monthly retainer and dedicated team, or with a fixed-price one-off project. What works for an SME.

Daniel Levis· Published 8 October 2026 ·Updated 31 October 2026

In brief

If the scope is clear, stable and limited, a fixed-price project gives budget certainty and is the simplest choice. If the process is complex, evolves as you build it, or you plan to continue over time, a monthly retainer with a dedicated team handles change better and reduces renegotiations. The right question is how stable your scope is, not which model sounds cheaper.

Option A

Monthly retainer with a dedicated team

You pay a monthly fee for three dedicated people (an Account Manager, a Project Manager and an AI Engineer) and one release in production every month, and client-approved milestones drive the work.

Pros

  • +Absorbs scope changes without renegotiating the contract every time
  • +With Soraia you sign the Sprint Goal at the start of the month and get a release in production at the end of it
  • +Evolution continues in a later Sprint, maintenance moves to Hypercare
  • +A stable team that builds knowledge of your process month after month

Cons

  • −Total cost is less predictable up front if you don't set a clear goal
  • −Requires an internal point person to set priorities and approve milestones
  • −If the pace of work drops, you risk paying for hours you don't fully use

Best for

  • SMEs with a complex process, or one that changes as you build it
  • Those expecting the agent to evolve over time, not just a single release
  • Companies that want a team that stays and knows the context

Option B

Fixed-price AI project (one-off)

You agree scope, deliverables and price in advance, and pay a set amount for that defined perimeter.

Pros

  • +Certain, predictable budget: you know exactly what you spend
  • +Easy to get approved by a CFO or board because it's a single figure
  • +Clear vendor accountability on the agreed perimeter

Cons

  • −Every change to scope becomes a renegotiation, often slow
  • −Post-release maintenance and evolution are usually excluded
  • −If the perimeter is mis-estimated, the vendor cuts quality or you go over budget
  • −Poorly suited to processes you only truly understand while building them

Best for

  • Clear, stable, bounded scope from the start
  • A single well-defined agent or automation
  • Those who need a firm number for an already-approved budget
Criterion Monthly retainer with a dedicated team Fixed-price AI project (one-off)
Budget predictability Medium (monthly fee, depends on pace) High (fixed figure)
Flexibility on scope changes High (you decide the Sprint priorities) Low (renegotiation for each change)
Evolution and maintenance A later Sprint, or Hypercare Usually excluded from the fixed price
Fit for changing processes Yes No, needs a stable scope
Code ownership Client's from day one Depends on the contract
Typical starting point From 5,000 euro per month One-off figure agreed on scope

The verdict

This isn't an ideological choice, it depends on how stable your scope is. If you know exactly what you want to build and it won't change, fixed price gives budget certainty and is easier to approve. If the process is complex, only becomes clear as you build it, or you plan to evolve and maintain it, a monthly retainer with a dedicated team handles change without constant renegotiations. Soraia works on a retainer precisely because AI agents almost always evolve after the first release, but still sets milestones and an agreed target: so you get the flexibility of a retainer plus a clear obligation of result.

FAQ

What people usually ask us.

With a monthly retainer, do I still pay if I do little work one month?
With Soraia you don't buy hours, you buy a Sprint: we sign the Sprint Goal at the start of the month, the demo is mid-month and the release goes to production at the end of it. The retainer starts from 5,000 euro per month, even for just one month, with three dedicated people. What matters is the release, which the client approves.
Isn't fixed price always safer for my budget?
On the single number, yes, but only if scope doesn't change. In practice, on AI projects the perimeter evolves as you build, and each change becomes a renegotiation. Fixed price is safe when scope is clear, stable and bounded; otherwise it hides extras and delays.
Does a retainer lock me into the vendor?
Not if the code is yours. With Soraia the code is the client's from day one, documented, with no perpetual subscriptions and no lock-in: you can stop or bring it in-house whenever you want. See how we build AI agents.
How do you handle maintenance and evolution after the first release?
After the first release in 4 weeks, evolution continues in a later Sprint and maintenance moves to Hypercare, from 500 euro per month: monitoring, bug fixes and security updates. In a fixed-price project this part is usually agreed separately, often with a new contract.
How do I choose between the two models for my case?
Start from how stable your scope is and how likely the agent is to evolve. If you have a closed perimeter, fixed price is simple; if the process is complex or ongoing, a retainer holds up better. If you're unsure, let's talk for 20 minutes and assess it together, or look at AI adoption to grow your internal team in parallel.

Not sure which one fits your case?

20 minutes with the CEO to work out the right choice for your processes. No pitch, no obligation.

20 min with Daniel