From idea to validated MVP with AKILI-SPECS

Your idea is worth something once you validate it, and worth more once you build it with method. This guide proposes six steps to go from an idea to a minimum viable product (MVP) already validated with real customers. The first five help you find out whether the idea is worth it; the sixth is building. At Akili we take that sixth step with AKILI-SPECS, our open and free methodology.

The problem: building without validating is expensive

It is common for a team or a founder to build a complete solution before knowing whether anyone needs it. By the time it reaches the customer, the most valuable resources are already spent. That cost has three faces:

• Time: months spent on features nobody asked for.
• Money: budget invested before knowing whether there is demand.
• Focus: a team that changes direction again and again because it has no clear plan.

Validating does not slow the project down. On the contrary: it avoids building what is not needed and keeps the energy for what matters.

Steps 1 and 2: the problem and the customer

Everything starts with the pain, not the idea. A good problem can be explained in one sentence, and someone suffers it today. For step 1, answer two questions:

• What problem do we solve? Describe it without talking about the solution yet.
• What does that problem cost today? In time, money or hassle.

Step 2 gives the customer a first and last name. “Everyone” is not a customer. You need to know who lives the problem and, for companies, who uses it, who chooses and who pays, because they are often not the same person.

What comes out of these steps: a clear one-sentence problem and a concrete customer profile.

Step 3: look at the market before you build

Knowing the ground keeps you from building something that already exists or that nobody is looking for. Four questions are enough to start:

• Alternatives: how does the customer solve the problem today? By hand, in a spreadsheet or with another tool.
• Competition: who else offers it? What they do well and where they fall short.
• Size: how many people or companies have the problem? A rough estimate is enough at first.
• Difference: why would they choose you? One clear reason is worth more than ten features.

Steps 4 and 5: test assumptions and validate demand

Before writing a line of code, look for cheap evidence that the idea solves something and that someone wants it.

In step 4 you write down your assumptions: everything you believe is true and have not checked yet. Then you test them as cheaply as possible: interviews, a simple page that explains the offer or a paper prototype.

In step 5 you ask for a real commitment: joining a waiting list, accepting a meeting or signing a letter of intent. Then you measure the signals and decide, with evidence, whether to continue, adjust or stop. Stopping in time is also a good outcome.

Step 6: build the MVP with AKILI-SPECS

What you learned in steps 1 to 5 is already a specification: the problem, the customer, the market and the tested assumptions. AKILI-SPECS takes that specification and turns it into a plan; then artificial intelligence builds on that plan.

The outcome we aim for is a small MVP that works and adds value from day one, built fast and with method, without improvising. Every change is checked against what was requested; if it does not comply, it does not move forward.

What AKILI-SPECS is and how it works

AKILI-SPECS is Akili’s own methodology. It is open source and free: the code is on GitHub and the package is published on npm as akili-specs. It works as spec-driven development: the plan comes first, and then AI builds on it.

Its cycle has six moments:

1. Propose: the change or new feature is put forward.
2. Specify: requirements, design and tasks are written down.
3. Execute: AI builds following the plan.
4. Test: automated tests run on what was built.
5. Validate: the result is compared with what was requested.
6. Archive: the cycle closes with a continuous-improvement retrospective that makes the next round better.

To install it in a project, the official command is:

npx akili-specs init

Because the method is public, anyone can review it before working with us. It is verifiable proof of how we work.

What we aim for you to gain

We do not sell hours or tools: we aim for results for your business, such as time saved, more inquiries and teams that work better with AI. With AKILI-SPECS we pursue four benefits. They are goals we measure with you, not guaranteed results:

• Speed: fewer round trips between the idea and the MVP, because the plan is already clear.
• Quality: every change is validated against what was requested before moving on.
• Predictability: you know what is being built, in what order and why.
• Continuous improvement: every cycle ends with a retrospective that improves the next one.

How we work with you

We mainly work with three groups: medium and large companies in Valle del Cauca, universities and schools, and founders or startups. AI training always goes hand in hand with AKILI-SPECS, so teams do not just use AI but build with method.

The path starts with a free 30-minute session on Google Meet, where we understand your idea or the process that takes up most of your team’s time. If it makes sense to continue, we study the problem in depth, plan the MVP and build it with AKILI-SPECS while we train your team to work with AI.

For universities and schools

We offer the workshop “Validate and build your MVP with AI”, designed for entrepreneurship programs. Students and teachers walk through the six steps with their own ideas and build with AKILI-SPECS.

Frequently asked questions

What is a validated MVP?

It is the smallest version of a product that solves the main problem, built after checking with real customers that the problem exists and that someone wants the solution. Validation comes before building, not after.

Does AKILI-SPECS cost anything?

No. AKILI-SPECS is open source (MIT license) and free. The code is on GitHub and the package installs from npm with npx akili-specs init.

What if validation says no?

That is a good outcome. Stopping in time, or adjusting the idea, avoids spending time, money and focus on something nobody needs.

The next step

Fewer ideas on the shelf, more validated MVPs. Book your free assessment · 30 min, no cost.