Skip to main content

Our process

We look at every project as an explosion of ideas: we move fast, we break things down, and the final product is something great that works right. A clear process means you know what will happen, who is responsible and when you will see results.

Process

We tackle big problems

From the first conversation to a product running smoothly, AI assistants work alongside our team at every step, while people set the goals, approve the results and stay accountable.

The five steps of a project, each carried out by AI assistants and approved by people

Five stages of every project

Each stage ends with something concrete for you to see and approve.

Step one: discovery

We sit down with key stakeholders and current customers to discuss goals, learn more about the business and talk about your current digital strategy, then dive into every metric we can measure. AI assistants gather documents and data; together with you we settle goals, scope, risks and an estimate.

Planning and design

Discovery gives us two things: personas of your ideal customer (their pains, how they use the internet, even a short story about them) and a conversion funnel. Then come wireframes of the user's journey, followed by static or interactive mock-ups.

Development

Development is a whirlwind. We move fast in two-week iterations, building features that mirror the customer's journey. AI assistants write code and tests, engineers review and approve every change, and stakeholders give feedback after each iteration.

Launch and measure

Nothing is perfect the first time, and we refine until we get it right. Once launched, we measure everything; there is normally a backlog of features to add after version 1.0.

A final product, for now

The process may end, but the project usually goes on. We keep you involved in what comes next without added stress, aiming for a memorable experience and a high-quality product.

A person at the centre of a loop of five AI assistants: analysis, design, coding, testing and operations

Humans in the loop

Who decides at each step

Every task we hand to an AI assistant has a human involvement level set in advance: in the loop, on the loop or out of the loop.

Human in the loop

For key decisions such as scope, architecture, customer data and releases, the AI proposes and only continues once a person approves.

Human on the loop

For repeated work such as test-driven coding and review, the AI works on its own while an engineer supervises and steps in when needed.

Human out of the loop

Only for safe, reversible work such as code formatting or test data, always with a log to audit later.

Frequently asked questions

  • How long does a project take?

    It depends on the scope, the platforms and the integrations. After discovery you receive a written scope and an estimate before we start building.

  • Which language do we work in?

    English or Vietnamese, whichever suits you. Project documents can be written in one or both.

  • Who runs the product after launch?

    We can keep operating and developing it, or hand over the full source code and documentation to your team.

  • Who owns the code?

    For client projects, the source code and the product belong to the client under the contract.

Let's get started

Tell us about the product you want to build. We will come back with the right questions and a clear way forward.