Running the experiment
The pilot prioritised larger, established AI platforms over niche tools — the kind of products with the track record and reliability to deliver consistent productivity gains across a team. We evaluated them against specific design tasks: research synthesis, ideation, asset generation, iteration speed. Two findings stood out. First, AI expands a designer's subject-matter range quickly. A designer working on a technically unfamiliar domain can accelerate context-building significantly with AI assistance. Second, the outputs require critical scrutiny. AI generates results that are consistently satisfactory but rarely optimal. The skill of evaluating and improving AI output is as important as knowing which tools to use. The work also surfaced a structural shift: designers who work with AI effectively tend to operate with broader scope, combining traditional design competence with enough literacy to direct, critique, and refine AI-generated work. That's a different profile than the one design teams have typically hired for.
"In the project done together with Nordkapp, we extensively experienced the added value of generative AI tools in design work, in the form of very concrete use cases." — Jani Heinola, Information Technology Business Partner at Telia
What it changed
The pilot produced a clear picture of where AI fits — and where it doesn't yet. Data analysis and design iteration are the two areas with the most immediate, demonstrable returns. Tasks that previously required hours of manual synthesis can be compressed. Iteration cycles that depended on designer availability can run faster with AI assistance in the loop. The open questions are harder. Intellectual property ownership of AI-generated assets is still legally ambiguous. Privacy compliance in tools that process user research data is not yet straightforward. Bias in training datasets can reproduce patterns that human-centered design is specifically trying to avoid. These are not reasons to avoid AI tools — they are constraints that design teams need to understand and manage. Telia left the pilot with a clearer view of both sides. The workflow integrations identified during the work have since been adopted. The unresolved challenges are on the agenda, not ignored.
The pilot covered the full Double Diamond: discovery, definition, development, and delivery. Findings apply across Telia's seven-market operation, where any process change affects a design team of more than 100 people.
Next projects.
(2016-25©)
