#248 Enterprise architecture and AI agents 6.1 min
TOGAF and ArchiMate: Where an AI Agent Fits, as Code
Where does an AI agent go on an enterprise architecture diagram?
Predict and run the check
Copy the code folder first. Open model.py and rules.py and write down the four problems you expect check to report. Then run python compare.py and compare your list with the output. You should see 13 elements, 12 relationships and 4 problems.
A hint
Look at the last relation in RELATIONS, the behaviour elements that nothing performs, the AI element, and the passive data.
Remove the planted layer mistake
In model.py, delete the line of RELATIONS marked as a mistake (Run the model uses Handle a refund request). Run python compare.py. You should see 11 relationships and 3 problems, and the layer problem gone. Put the line back afterwards.
A hint
Delete the whole line, including the comment at the end of it.
Make a second tool an AI element
In model.py, add ai=True to the Order tools element. Predict how many problems the check finds, then run python compare.py. You should see 5 problems, including a new one for Order tools, and after the repairs 1 problem. Then add a line to the repairs in compare.py so that Support team lead also answers for Order tools, and run again to get 0. Then undo both edits before the next task.
A hint
The repairs list in compare.py already has an answers for line for Support agent. Copy its pattern.
Drop one repair
In compare.py, remove the answers for repair line from the fixed list, keeping the list valid Python. Predict how many problems remain after the repairs, then run python compare.py. You should see after the repairs: 1 problems.
A hint
Move the closing square bracket onto the end of the previous line.
These tasks are for you to try on your own machine. Nothing is checked or saved.
What this does not show
A toy model of 13 made-up elements with rules written from Wikipedia summaries, not from the ArchiMate or TOGAF standards; the checks only test that a link exists, and a model that passes is consistent, not good.
Up next in Enterprise architecture and AI agents
More in Enterprise architecture and AI agents
- #229 A Reference Architecture for Enterprise AI Agents, as Code
- #230 Agent Gateway: One Front Door for Models, Tools, Agents
- #231 MCP Gateway: Per-Tool Policy for Agent Tools (Python)
- #232 Give Every AI Agent Its Own Identity
- #233 Delegated Authority: Let an AI Agent Act for a User, Safely
- #234 Authorisation for AI Agents With Policy as Code (Cedar)
Related videos
- #229 A Reference Architecture for Enterprise AI Agents, as Code Enterprise architecture and AI agents
- #241 C4 Diagrams for AI Systems, Written and Checked as Code Enterprise architecture and AI agents
- #242 Architecture Decision Records for AI Choices, With a Lint Enterprise architecture and AI agents
- #100 Fake but Realistic Test Data in Python With Faker AI security and governance
