# Discovery: from a broad AI ambition to a first built solution

Source: ArrowEye (https://arroweye.be) · Language: English · Canonical page: https://arroweye.be/en/cases/use-case-discovery-healthcare

An organisation that knew it had to do something with AI, but not where: a workshop per department mapped and scored the opportunities; the strongest was worked out and built.

## Facts

- **Category:** Transform (discovery, agents and automations)
- **Sector:** Healthcare software company

## The question

The organisation knew it had to do something with AI, but had no idea where or how. Not which processes lend themselves to it, not which department holds the greatest value, not where to start. The question: help us make concrete where AI can be used in our organisation, department by department, and in which order.

## Our approach

A workshop with participants from sales, implementation, support, QA and development, so every department put its own work on the table. Per team we went through the tasks and bottlenecks and translated them into concrete applications, then scored them on business value, feasibility, data availability and time to impact. The strongest was selected, worked out with scope, approach and lead time, and built. A deliberate choice: first a solution that makes knowledge searchable, to learn and build trust; the next ones are prioritised and ready.

## Impact delivered

Results, before → now:

- Opportunities mapped: no overview → 59, prioritised

From "we have to do something with AI" to a validated backlog: the organisation now knows per department where AI can be used, what is feasible and where to start, and the teams made that choice themselves. The first solution has been worked out and built; the next ones are prioritised and ready.

## The route

1. work and bottlenecks per department
2. scored on value, feasibility, data and time to impact
3. the strongest worked out
4. built

At the same client: [From Copilot inspiration to a team that builds its own agents](/cases/copilot-inspiratie-enablement) and [Governed documentation agent across seven knowledge sources](/cases/governed-documentatie-agent).