How to generate measurable impact with generative AI

In recent years, financial institutions have embraced artificial intelligence with great enthusiasm. For many, the early stages were characterised by an exploratory “play phase”, where innovation and experimentation flourished and AI-driven pilot projects emerged across the organisation.

Previously, leadership teams could support AI initiatives just out of curiosity. Now, concrete demands are emerging: show me the money. It has become necessary to document value, quantify impact and ensure that AI solutions actually deliver measurable results.

A study from MIT shows that as many as 95 per cent of implemented AI solutions fail to produce measurable value. This is often due to an excessive focus on the technologies themselves and the fascination with what the tools can do. The pattern is familiar: when new technologies enter the market, excitement grows, investment follows, but results are slow to appear – after which scepticism sets in. Only once the technology matures does the value begin to materialise. We saw the same with the PC, the internet and ERP systems – technologies that are indispensable today but started with more hype than impact.

The first phase of the AI journey is typically technology-driven. The next phase is about the concrete value the technology creates. Many organisations are now at a point where enthusiasm and pilot projects are no longer enough. A clear shift in focus is needed: moving from fascination to solutions that address real business problems and deliver documented value.

The path from pilot to production is, however, complex. Scaling up AI pilots across an organisation and making the technology an integrated part of daily operations is a far from trivial task. Experience shows that successful scaling up requires AI initiatives to be driven by a clear business ambition – not by technological possibilities alone. It also demands business insight, strong governance, investment in data and skills, and active change management right down to individual users.

We are currently working on concrete projects for two globally operating banks where generative AI has been scaled effectively and systematically, resulting in measurable impact. These banks have made significant progress in integrating generative AI into broad workflows through a structured, value-focused approach. The work has included setting clear objectives, prioritizing the most relevant use cases, realizing tangible results, and continuously scaling and embedding AI in daily operations. The outcomes are measurable and substantial: higher employee satisfaction from reduced routine work, strong efficiency gains, and improved work quality that benefits both internal teams and end customers.

AI only creates results when it becomes an integrated part of organisational working patterns. Leadership must therefore set the direction and drive the change towards a responsible and value-creating implementation of generative AI in core processes – including through targeted investment in learning and development. It also requires a culture that systematically identifies and integrates value-adding AI solutions into processes. In this way, AI becomes a genuine boost for the business – and not just another series of experiments which are not anchored in the organisation.

Read the article in Finans in Danish.