Most AI initiatives stall before any code is written, and often the reason is the commercial model rather than the technology. One side wants a fixed price to control risk. The other knows the scope will change as soon as real data is involved. Both are right, which is why the choice between a fixed-scope project and a monthly retainer deserves a deliberate decision. Here is a practical way to make it.
The two models in plain terms
Fixed-scope project. You agree on a defined outcome, a timeline and a price. The partner is accountable for delivering that outcome. Changes go through a change request.
Monthly retainer. You invest a recurring amount for senior capacity, which is directed toward a prioritized backlog. With a Brazilian nearshore partner, senior engagements typically start around USD 8,000 per month. The deliverables change as priorities change, and you can adjust or stop with notice.
Neither is better in general. Each one fits a different kind of uncertainty.
When a fixed-scope project fits
- The outcome is specific and testable. For example: connect your CRM and billing system, and generate invoices automatically from approved quotes.
- The inputs are known. The data sources exist, you can describe their format and you have access to them.
- The budget is capped. Your finance team needs a number before approving.
- Your team has limited bandwidth. You want to hand over a problem and receive a solution, not manage a backlog.
The strength of a fixed scope is clarity. The weakness is rigidity: if you learn something in week four that changes the right answer, you either pay for a change or build something you no longer want. That is why fixed-scope projects need a good discovery phase first.
When a monthly retainer fits
- The problem is still being discovered. You know you want AI to help customer support, but you do not yet know which parts are worth automating.
- There are many small items. Integrations, reports, prompts, fixes and experiments that do not justify separate proposals.
- You want continuity. The same senior person remembers why decisions were made and keeps improving the system after launch.
- AI quality needs tuning. Models, prompts and data change, so the work after launch is part of the work.
The strength of a retainer is flexibility. The weakness is that without discipline it can drift. You need clear quarterly outcomes and a monthly review, or the budget gets absorbed by whatever is loudest.
A simple decision framework
Answer these five questions honestly.
- Can I describe the finished result in one paragraph that a stranger could verify? If yes, lean fixed-scope. If no, lean retainer or a discovery phase.
- Do I control the data and system access today? If access is uncertain, fixed pricing is risky for both sides.
- Will requirements change as users try it? For AI assistants and agents, usually yes. That favors an iterative model.
- Do I need work after launch? Monitoring, tuning and new use cases point to a retainer.
- Who will own priorities on my side? A retainer needs someone who can decide what comes next every week or two.
The hybrid that often works best
Many successful initiatives combine both models in sequence:
- Assessment (short, fixed). Review the workflows, data and risks, and produce a written plan with priorities.
- Pilot (fixed scope, 30 to 60 days). Build one narrow use case into production, with defined acceptance criteria and a measurable result.
- Retainer. Once the pilot proves value, move to monthly capacity to extend, maintain and improve what you built.
This sequence limits your risk early, when you know the least, and adds flexibility later, when you know the most.
Hypothetical example: a US property management company wants to automate lease document review. The assessment defines one document type and one extraction target. The pilot builds an extraction workflow that a staff member reviews in a staging environment before it goes live. Once it works, the company moves to a retainer to add other document types, build a dashboard and handle exceptions.
What good communication looks like in either model
The commercial model does not change the need for visibility. In both cases you should receive a preview of what will be built, a regular check-in, written progress reports by email or on a dashboard, and staging environments for your approval. In a fixed-scope project these are the milestones. In a retainer they are the rhythm of the month. A senior project lead should be accountable for all of it, and when the work needs a specific skill, vetted remote specialists join for that phase.
Contract points that matter more than the model
- IP assignment to your company, effective on payment or delivery.
- Acceptance criteria written before work starts, not after.
- Change process for fixed-scope work, with a simple way to estimate and approve changes.
- Notice period for retainers, short enough to give you freedom and long enough to protect a handover.
- Documentation obligations in both models.
- USD invoicing and clear payment schedules.
Common mistakes
- Forcing a fixed price on an unclear problem, then fighting over scope.
- Starting a retainer with no outcomes, then wondering where the money went.
- Skipping the pilot and committing to a large build based on a demo.
- Choosing by lowest number rather than by who will be accountable.
The right model is the one that matches how much you already know. When the answer is clear, fix the scope. When it is not, pay for learning in small, measurable steps, and then decide. Either way, insist on a senior owner, clear communication and a first deliverable you can test.
Want to put this to work? Convertty runs an AI assessment of your workflows and delivers an implementation plan with a pilot in production in 30–60 days. Book a call.
