AI Training, GRADtrain
Training on your own cases.
We offer three courses, from foundational knowledge for all employees to the team that builds and owns agents, always based on your documents, processes and questions.
Three courses
From foundations to an agent team.
AI Foundation
From fundamentals to safe use of AI.
We explain how language models work, practise prompting with measurement, put risks, law and ethics in context, and find your use cases plus your own RAG notebook.
- 1 day
- all employees
- no prerequisites
- 5 modules
- measurement sheet per person
| Module | Artefact | What you use it for |
|---|---|---|
| 1Fundamentals: understand how language models answer and where they fail | Reviewed list of approved AI tools | Know which tool is approved for which data |
| 2Opportunities & prompting: demo circuit, build prompts with CRAFT, measure time saved | Measurement sheet and tested prompt | Prove time saved for AI Strategy, reuse prompts team-wide |
| 3Risks: check results, apply the data protection traffic light, decide borderline cases | Traffic-light ruling, six borderline cases | Classify inputs before sending, using the data protection traffic light |
| 4Ethics & society: bias, copyright, law at a glance (FADP, EU AI Act), drafting a code of conduct | Code of conduct, version 1 | Record rules for everyday AI use in the team |
| 5AI in your own business: rate use cases, build your own RAG notebook | Use-case list and RAG notebook | Candidates for AI Strategy, query documents with sources cited |
AI Strategy
From current position to launch-ready pilot.
We determine your position, sharpen the vision, prioritise use cases with a business case and set up the pilot, backed by decisions on data, technology, roles and law.
- 2 days
- leadership and business units
- 6 modules
- 6 exercises
- no coding
| Module | Artefact | What you use it for |
|---|---|---|
| 1Diagnosis: assess maturity, system landscape and shadow AI with evidence | Position assessment with maturity profile | Evidence the starting point, justify bottlenecks, cost inaction |
| 2Vision: name the bottleneck, write a guiding statement, set exclusions | One-page vision | Test proposals against the guiding statement, reject with reasons |
| 3Portfolio: find use cases, calculate the business case, prioritise sequence | Prioritised shortlist and business case | Request start sequence and budget for year 1 |
| 4Implementation: draft pilot charter with stop thresholds and 90-day plan | Pilot charter and 90-day outline | Launch pilot, optionally implement it as a process agent in AI Orchestration |
| 5Data and technology: operating option per data class, buy, partner or build | Operating-option decision per data class | Prepare architecture and contract talks with providers |
| 6People and accountability: anchor roles, policy, law and change | Roles, policy and risk register | Assign accountability, adopt the policy, manage risks with owners |
AI Orchestration
From charter to secured agent.
We set autonomy and charter, build tools via MCP, state and approvals, measure with evals, orchestrate via A2A, secure with guardrails and plan the rollout.
- 2 days
- development and architecture
- 7 labs plus capstone
- solid Python
| Module | Artefact | What you use it for |
|---|---|---|
| 1Foundations & autonomy: choose architecture, set autonomy, write the charter | Agent charter | Set limits and owners for an agent, optionally for your AI Strategy pilot |
| 2Tools & MCP: build, attack and secure an MCP server | Secured MCP server | Connect internal systems to agents securely |
| 3Memory & state: implement approvals, crash and resume with checkpoints | Agent with approval | Execute write actions only after human approval |
| 4Evals & observability: read traces, categorise errors, set eval gate in CI | Error taxonomy and eval gate | Find errors systematically, block releases below threshold |
| 5Orchestration & A2A: design a team, connect via A2A, measure against single agent | Justified multi-agent architecture | Prove whether a team justifies its cost over one agent |
| 6Guardrails & governance: attack the system, harden it, document accountability | Canvas and registry | Show which agent acted, why, and who is accountable |
| 7Rollout & scaling: manage costs, plan rollout, decide on go-live | Readiness review | Decide go, no-go or conditions for production |
How we train
On your cases, not off slides.
With your documents and processes, a measurement sheet per person and exercises instead of long slide lectures.
The next step
Which course fits your team?
In the intro call we settle audiences, cases and dates.