AI MVP LAUNCH
Your AI idea.
A product people
can use.
Go from a promising idea to a focused first release. We bring the product, AI and engineering together.
A REAL TASK
What needs attention in this agreement?
Three items to review
- Payment due within 30 days
- Renewal requires written notice
- Liability clause needs your review
References linked to the original document
Illustrative interface · Your workflow is scoped with you
The idea works.
The product still needs work.
A convincing AI demo is a starting point. A first release needs to work with your users, their data and the systems around it.
Define one valuable user journey, agree how you will test it, and build the product around that journey.
- 01The first version keeps growingEvery new feature delays the core journey that would tell you whether users find it useful.
- 02Good answers are inconsistentA polished demo input works. Messy documents, missing context and unfamiliar questions expose the gaps.
- 03The workflow stops at the answerUsers still copy results into other tools. Permissions, saved work and integrations remain unfinished.
- 04Usage costs are unclearModel calls, long inputs and retries can change the economics as more people start using the product.
- 05Data boundaries need decisionsYou need to decide who can access which information and when a person must approve an action.
- 06Launch has no clear finish lineWithout acceptance criteria, the build stays in revision instead of reaching a release you can put in front of users.
BEYOND THE CHAT BOX
An answer is a moment.
A product completes the task.
For example: turn a document into
findings a user can check and save.
- 01
Start with trusted inputs
A signed-in user uploads a supported document. The product checks file limits and applies the agreed access rules.
The right information, in the right account - 02
Make the result reviewable
The AI extracts the agreed information and links back to the source. Missing or uncertain results have a defined next step.
Findings the user can inspect - 03
Let the user finish the job
The user reviews, corrects and saves the result, or sends it to an agreed integration after approval.
A completed journey, not a stranded answer
Illustrative workflow. Your inputs, review steps and integrations are agreed during scoping.
FROM IDEA TO FIRST RELEASE
Small enough to launch.
Complete enough to use.
One customer problem. One core journey.
The foundations to take it further.
One useful AI workflow
Define the task, inputs, outputs and boundaries before choosing the model.
Checks against real examples
Agree representative test cases, quality criteria and what happens when an answer is unsuitable.
Visible running costs
Estimate hosting and model usage. Include agreed usage limits, monitoring and cost controls.
A complete first release
Build the agreed user journey, permissions and integrations. Deploy in your accounts with documentation.
HOW WE GET YOU TO LAUNCH
A visible path.
A decision at every stage.
You review working software.
We agree the release criteria together.
- 01 · Define
Agree version one
Choose the user, core task, sample inputs and the features that belong in the first release.
You get
Written scope, quote and delivery plan
Done when
Scope and acceptance criteria are approved.
- 02 · Evaluate
Test the AI workflow
Check representative examples, failure cases and the approach to review, fallback and usage costs.
You get
Evaluation results and agreed AI approach
Done when
The agreed evaluation criteria are met, or a scope decision is made.
- 03 · Build
Connect the product
Build the interface, permissions, saved work and integrations around the agreed journey.
You get
A staging release you can review
Done when
The agreed end-to-end checks pass.
- 04 · Launch
Release and hand over
Deploy in your accounts, verify the release and explain how to operate and maintain it.
You get
Deployed product, code and documentation
Done when
The agreed release is verified and handed over.
CHOOSE THE RIGHT STARTING POINT
Ready for a focused
first release?
You do not need a finished specification. Bring the problem, the intended user and examples of the work your product should help them do.
AI MVP Launch fits when
- AI is part of the core value your product delivers.
- You can provide representative inputs and someone to review results.
- You are ready to choose a focused scope and test it with real users.
A CLEAR WAY TO START
Scope it together.
Then make it real.
Start with a free scoping call. Get a written scope, delivery plan and fixed quote before the build begins.
Prefer to talk? Book a free callFor the scope we agree. No one-size-fits-all build price.
- Working demos and agreed acceptance checks
- Deployment, documentation and handover
- 30 days of defect fixes against the agreed scope
Scoping fees & running costs
Deeper discovery can start with a $500 Scope Sprint, credited in full against the build. Feasibility experiments are quoted separately. Hosting, model usage and subscriptions are estimated during scoping and billed to your accounts.
Good questions.
Clear answers.
Scope, delivery and what happens next. The details that help you decide.
MVP Launch
Scope a focused first release with agreed deliverables and a written quote. Existing prototypes are reviewed for what can be reused.
Explore MVP LaunchAsk about your projectWhat is AI MVP Launch?
A scoped first release of a product whose core user journey uses AI. We agree features, evaluation examples, integrations, acceptance criteria and handover before development. Using AI coding tools alone does not make a product an AI MVP.
Can you start from an idea or an existing prototype?
Yes. Bring the user problem, any designs or prototype, and sample inputs you are authorised to share. We assess what can be reused before quoting. If you only need a blocker repaired, System Rescue is the better starting point.
How do you check the AI before launch?
We agree a representative evaluation set and acceptance criteria for the task, test useful outputs and failure cases, and define fallback or human review behaviour. Tests reduce risk; they do not guarantee every future output will be correct.
How much does it cost, and how long does it take?
The free scoping call establishes the core journey and dependencies. Straightforward projects receive a written scope and fixed quote. Deeper discovery can start with a $500 Scope Sprint, credited in full against the build. Extensive experiments or feasibility prototypes are quoted separately. Your delivery schedule is agreed in the scope.
Who pays for AI usage and owns the product?
You own the delivered code, designs, documentation and accounts. Hosting, model usage and third-party subscriptions are billed to your accounts. We estimate these costs during scoping and agree usage controls. Third-party models, services and libraries retain their own licence terms.
What happens after launch?
The build includes 30 calendar days of fixes for delivered functionality that does not meet the agreed requirements. New features, model or provider changes, and ongoing optimisation are scoped separately. Launch means the agreed software is deployed and handed over; customer acquisition and funding are not included.
LET’S DEFINE VERSION ONE