Innovation and improvement projects in the public sector
We help municipalities clarify their needs, test ideas with the people who will use the solution, and take what works forward into secure operation. Fast, controlled and with room to learn along the way.
You know the challenges. We build the solution together with you.
AI-supported development makes it possible to test an idea early. But technology alone is not improvement work. We work closely with professionals and users, define the scope of the task and let what we learn from testing decide the next step.
The municipality can stop, change direction or move on after each phase. No major investment has to be approved before the benefit has been explored.
A new starting point:
Experimentation and testing with new technology
With AI, we can explore and experiment more, and find entirely new ways of solving tasks.
It applies to AI-assisted solution development, where ideas become testable prototypes in weeks instead of months. And it applies to large language models used for advanced analysis and knowledge support, where the municipality's own documents, routines and data can be made searchable and useful for its professionals. Tasks that were previously set aside as too big, too expensive or too complicated can now be tackled as well-defined projects with low risk and fast learning.
It's actually quite simple
Three phases, three decisions
01 Exploration:
We map the need, who the solution should help, existing workflows and requirements for data and security. If needed, we test whether the idea is technically feasible.
Deliverable: Needs assessment and a short decision memo
The municipality decides: Proceed with a prototype – or stop?
02 Prototype and test
We build a working prototype, let key users try it in relevant work situations and improve the solution based on concrete feedback.
Deliverable: Tested prototype and documented learning
The municipality decides: Adopt, develop further – or stop?
03 Deployment
Once the solution has proven its value, we clarify requirements for security, privacy, operations and quality. Some simple tools can be put to use quickly; others need more preparation.
Deliverable: Extended pilot, system description, test reports and a separate quote for delivery and operations
The municipality decides: Put into production?
PREDICTABLE SCOPE
Choose a size for each phase
The size of the assignment depends on factors such as the number of user groups, functional scope, what information is processed and the risk of errors. A small exploration can be followed by a larger prototype – or the other way around.
Example:
| Phase | Small | Medium | Large |
|---|---|---|---|
| Exploration | 10–20 hours | 20–30 hours | 30 hours or more |
| Prototype and test | 30–50 hours | 70–90 hours | 90 hours or more |
| Deployment | 5–30 hours | 30–80 hours | 80 hours or more |
Indicative hours per phase and package size
The figures indicate package scope, not hourly rates. Exploration and prototype are agreed at a fixed price per phase before work starts. Innocode carries the risk if the work takes longer within the agreed package. Deployment is indicative and priced in a separate quote once the prototype has been tested. Any changes to price or scope are agreed in writing before additional work.
Safe from the start
Speed must not come at the expense of control
We clarify what data the solution will process before choosing technology and ways of working. As a rule, development and testing are done with synthetic or anonymised data. Access is limited to what is needed; testing, quality assurance and a clear decision come before production. It is still the professionals who assess, quality-assure and make the decisions.
- Privacy and risk are assessed early
- Key users test before further investment
- People quality-assure the solution and make the decisions
From concrete needs to something you can try
The examples show different types of assignments. Status and further rollout are agreed in each individual project.
Sandefjord Municipality: Sparringspartner
An AI-based reflection tool for kindergarten and school staff who are concerned about children and young people. It helps them structure the situation and find relevant routines without replacing professional judgement.
Production rollout planned for autumn 2026.
SANDEFJORD MUNICIPALITY: Frivillig i Sandefjord
A platform that connects voluntary organisations with residents. Input from the organisations has guided priorities and made the solution more user-friendly.
Further development is being planned together with several municipalities.
FROGNER DISTRICT, OSLO: Ung Jobb
A mobile-friendly service developed with young people. The AI guide helps them highlight experience from everyday life and create a CV; jobs are collected from several sources.
Developed and tested together with the target group.
Nacka Municipality: Beslutstrålaren
In just a few months, Nacka Municipality and Innocode went from idea to a working solution that makes political decisions easy for residents to understand. The solution sends the decisions as push notifications in the Mitt Nacka app. It was developed as a rapid project together with the digitalisation team at Nacka Municipality, with the municipality’s professionals involved in every round.
Beslutstrålaren automatically retrieves new minutes from the municipal council and the municipal executive board. Using AI, each decision is summarised in plain language and tagged by topic and district, following the same structure as nacka.se. The original title is kept and the source is linked, so the AI summary is never mistaken for the municipality’s own wording.
Experienced people at work
Innocode's AI team is a permanent, cross-disciplinary senior team that has worked together for several years, including on Antiloop for the City of Helsingborg and AI solutions for Sandefjord Municipality, Nacka Municipality and Tromsø Municipality. The team covers the entire delivery itself: process management, service design, UX and accessibility, frontend, backend and integrations, quality assurance and DevOps. This means a smaller assignment can go from idea to tested solution without us having to bring in new people.
We use AI tools in development, but architecture, code review and testing are done by experienced team members. Experienced developers steer the architecture, and QA tests all code before it is put to use – including code written by AI. Development is done on synthetic test data, with separate test and production environments and controlled rollout.
Get in touch :)
The services are delivered by Yulieta – our dedicated innovation and product team
Yulieta is the name of Innocode's digital development department. We combine strategy, design and technology to turn ideas into good, working solutions.
Citizen 2026: Agile innovation and the new technology landscape
How can municipalities move, improve, and innovate more efficiently while still staying in control with new technology?
IN etNO CODE
On August 29, 2026, we will host an event called IN etNO CODE at our office in Lviv, Ukraine. This will be the fourth event of its kind that we have organized.