Start with AI at work

Start with one specific situation: a tool colleagues already use, a repetitive task or a supplier proposal. You do not need to know every law or have a finished policy before you begin.

1. Define what should improve

Describe the current process, the problem and the result you need. Who will use it? How will you recognise improvement? Consider changing the process without AI and using tools you already have.

• Options, benefits and costs — prepare a decision before investing.
• Choosing an AI architecture — when a ready-made tool is enough and when integration matters.

2. Find out what you already use

List tools and their specific purposes. Identify users, input data and who checks results. Mark missing information for follow-up.

• Build an AI inventory — the process and useful fields.
• AI use outside the organisation’s view — ask colleagues about actual practice and agree shared rules.

3. Agree rules and prepare people

Define permitted tools and data, checks on important outputs and a contact for questions. Identify the skills people need for their actual work.

• Rules and data — from an inventory to approved rules used in practice.
• People and learning — skills, communication and practice.
• Which laws may apply — assess the specific activity and your role.

4. Test a bounded use case

Before wider deployment, test quality, effort, cost and limitations on representative tasks. Agree who reviews the pilot and the conditions for continuing.

• Choose and implement a solution — brief, pilot, handover and operation.
• Templates and checklists — material to adapt.

How the steps connect

The AI-Native Entry Framework connects the business situation, rules, people, verification and implementation. Build on work you have already done. See the FAQ and glossary for common questions and terms.