AI agents and workflows for companies in the EU
Put AI to work And bring your existing AI use under control
Your first working AI workflow by the end of a two-week sprint. We agree the task, set up data access and define where human review is needed
First conversation · 20 minutes · free
From scattered trials
to a working system
We identify the tools, access and checks your first workflow needs. From there, we help you develop agents, train your team and manage AI use across the company
01AI inventory and rulesKnow which tools your team uses,
with what data, and for what purpose
We map the tools already in use, the work they support and the data involved. We flag cases that need a closer review
- A clear inventory of tools and owners
- Practical rules for handling data
- Questions for further review
02Company-managed accessGive your team approved AI tools
with access managed by the company
We choose a platform that fits your existing systems and review access settings and data-processing terms
- User and role management
- A focused set of approved tools
- A transition plan for the team
03Team training and playbooksTrain your team on everyday tasks
and turn proven methods into AI instructions
We train people for their roles and work with process owners to create reusable AI instructions: what to do, which data to use and how to check the result
- Training on tasks relevant to each role
- Named owners and version history
- Clear quality checks
04AI agents and workflowsBuild workflows around your tasks
and test them in everyday work
We build AI workflows and agents to search documents, reconcile data or prepare drafts. We agree which outputs and actions need human approval
- Workflows, agents and connected data sources
- Team usage and output quality
- A comparison before and after implementation
What the inventory looks like
Sample inventory · illustrative, not client data
| Tool | Task | Owner | Review point |
|---|---|---|---|
| ChatGPT | Email drafts | Sales lead | Data in prompts |
| AI in Workspace | Meeting summaries | Operations lead | Access controls |
Second Friday means your team has its first working AI workflow by the end of a two-week sprint. We agree the task, access and acceptance criteria before the sprint starts
Discuss AI adoptionOne workflow
Two weeks
Choose one task
Choose a repeatable process with a clear owner and a result you can check
Set the scope
Agree the sprint scope, data access and actions that need human approval
Build and test
Build the workflow and test typical cases and exceptions with your team
Hand over to the team
By the end of week two, hand over a working workflow, team instructions and test results
Managing AI use
Clear rules
Clear accountability
Document the AI tools your team uses, what data they may handle and who owns each use case. Give clients, auditors and legal advisers a clear account of your setup
What we prepare
An inventory of tools and use cases, named process owners, practical AI rules and training by role
Materials for legal review
Use-case descriptions and questions to discuss with your legal adviser. A qualified specialist provides the legal assessment and advice
Our scope and the EU AI Act
We help set up the working environment, inventory, rules and review materials. We do not provide legal opinions. Specific obligations under the EU AI Act depend on your role, the system and the use case; a qualified specialist should assess them. Read the European Commission’s guidance ↗
Find the right
starting point
Answer three questions about how your team uses AI today. We’ll suggest a sensible first step
Choose an answer to continue
Let’s start with
your team’s work
We’ll discuss your task, current tools and constraints, then define the first AI workflow and the scope of a two-week sprint
Before we
talk
Do we need our own server?
Not necessarily. The choice depends on your data, requirements and existing systems. We’ll assess whether corporate cloud tools are sufficient
What will we have after two weeks?
One working AI workflow within the agreed scope, team instructions and test results from your examples. Before the sprint, we agree the task, required access and acceptance criteria. Further development is scoped separately
How much does implementation cost?
The first 20-minute conversation is free. We price the sprint once we agree the task, scope and intended result. Further development is discussed separately
Do you provide legal advice?
No. We prepare an inventory of tools, rules for using AI and process descriptions. A qualified legal adviser assesses any use case that needs legal review