Use case

Our AI product development process

We use AI to write code and prepare tests. The team decides what to build and reviews each delivery.

We use tools such as Claude and Cursor to implement components, review code and prepare tests. We handle design, scope and technical decisions. Delivery time depends on the project, its integrations and the reviews it needs.

How the workflow changes with AI

We design in Figma and use the agreed components and styles to build the app. AI helps implement screens and prepare tests. We review the code and behaviour before approving a delivery.

We also use it to reorganise code, look for missing test cases and spot possible problems in database changes. The team checks the suggestions and decides which to apply.

How the team works

Design and development share tools and workflows. That helps us review the whole product and keep the context when work passes from one person to another.

Scope, team, schedule and service terms are agreed in the proposal and contract before work starts.

What we review ourselves

The team handles architecture, scope and the client relationship. Deciding what to build requires understanding the business and what users need. We also review code changes before they go to production.

We manually review database migrations, permissions and authentication before merging changes, using OWASP Top 10 as a reference.

Real cases from Productea

The Power Plugin case study shows what we built for them. The portfolio also includes earlier projects such as Capboard, Inmovilla and Swipcar, with different scopes and ways of working.

Frequently asked questions

Does AI write the code that ships to production?+

AI writes drafts. A senior reviews, adapts, refactors, and approves before commit. Every PR goes through human review and automated tests.

What about IP if you use AI?+

Code ownership is documented in the contract. Data handling depends on the tools, configuration and terms agreed for the project. Privacy requirements are reviewed before sensitive data is used.

How do you review security?+

We manually review migrations, permissions and authentication using OWASP Top 10 as a reference. Changes go through human review before they are merged.

Which AI tools do you use exactly?+

Claude (Anthropic) as main code and docs engine. Cursor as AI-assisted IDE. OpenAI for UX research and asset generation. Plus proprietary workflows built on top of these APIs.