Ten Elements That Make an AI Project Succeed, and Why We at CivAI Incorporate Them from Day One

Christian Gossain describes ten elements in his Digital Project Survival Guide that determine whether an AI project succeeds or stalls. For us at CivAI, this is not an abstract theory; it is precisely why we organise projects for municipalities, schools, and care organisations the way we do.

Start with the problem, not with the technology

The first element is the most important. Start with a clearly defined problem, not with an infatuation for a new technology. In practice, we still too often see organisations that want to “do something with AI” without it being clear what issue it solves and for whom. That is why, with GovGPT, EduGPT, OrgGPT, and CareGPT, we always start with the client’s work process, not with the model.

Compliance and risk belong at the start, not at the end

Gossain also cites respect for compliance and risk as a critical factor, with legal and security specialists involved from the outset. This is exactly why European sovereignty is not a marketing term for us. Data storage, the GDPR, and the requirements of the EU AI Act are built into our platform, not a checklist to be ticked off afterwards.

Designing for data, from day one

Another element that stands out is determining in advance which data is needed to actually demonstrate the promised value. Without that design, an organisation will not know after the pilot whether the investment has yielded anything. This is also one of the reasons why we steer our clients towards measurable outcomes, not just a feeling that things will probably be fine.

What is still missing

The ten elements explicitly lack leadership, AI-specific ethical risks, and a learning design perspective. These are not minor issues. An AI project that works for a classroom or a service desk requires someone to safeguard the pedagogical or social aspect, not just the technology. That is why we always have a subject-matter expert alongside the technical team, whether that is an education specialist or an adviser with knowledge of the public sector.

Resilience and a learning attitude, two of the ten elements, may be the least exciting but are perhaps the most underestimated. An AI project rarely proceeds in a straight line. Organisations that build in space for setbacks and adjustments from the start ultimately fare better than those that had counted on a rigid schedule.

Meeting room, Architekten Bernhardt + Partner via Wikimedia Commons, licence CC BY 4.0