THE NEXT STEP AFTER TFIS FLUENCY

Give AI a job.
Keep the
judgment.

Move from asking for answers to directing agents that use tools, remember context and do useful work—with boundaries you can verify.

Choose your access →

12-week learning path. Your pace. No coding prerequisite.

An agent loop: your brief, model decision, tools and observation, with your approval before a commitmentYOUR BRIEFDECIDE NEXT STEPUSE A TOOLRead · calculate · draftYOUR APPROVALBefore a commitmentObserveContinue
For nontechnical professionalsEnglish · self-pacedReal API PlaygroundEvidence-linked curriculum

A useful agent
needs more than a prompt.

Learn the whole loop through one practical case. Each week, build another part of a client-enquiry assistant—then apply the method to your own work.

Understand the system

See what models, memory and tools actually do. Recognise when a simple workflow is the better choice.

Practise with evidence

Inspect actions, calculations and failure states in the Playground. Keep a portfolio of work you can review.

Keep authority

Set permission boundaries. Test the outcome and the route taken. Know when to stop and ask a person.

Your 12-week path

Plan for 90–120 minutes a week, including practice and two short return visits. All weeks remain available while your access is active.

Learn. Try.
Return. Improve.

A spaced learning rhythm gives you time to use each idea. Recall it before reopening the lesson, then test it against a changed case.

First visit

Read the explanation, inspect a worked example and use the Playground with fictional business records.

Two days later

Answer a short recall question without notes. Compare it with the lesson and correct what you missed.

One week later

Try a different case. Save the evidence, then build on it in the next week.

Built on a clear source trail.

The primary source is An Illustrated Guide to AI Agents by Maarten Grootendorst and Jay Alammar. Original TFIS explanations and workplace exercises connect its concepts to practice.

Secondary research is restricted to arXiv papers from the specified AI labs and universities. The course Audit links learning outcomes, source references, limitations and assessment criteria.

Independent TFIS course. No affiliation with or endorsement by the book authors, research labs or universities is implied. The source book is not redistributed.

Choose your access.

Complete TFIS Fluency first. Both plans include the course, Playground, Audit and portfolio tools. Provider API usage is billed separately by your chosen provider.

One-time access

US$1,000

  • One payment; no recurring course fee
  • Full 12-week curriculum
  • Return to lessons and saved work

Monthly access

US$42 / month

  • Renews monthly until cancelled
  • Full access during the paid period
  • Cancel through your billing portal

A subscription, not an instalment plan. Cancellation ends access after the paid period.

Pay through Stripe, then sign in with the email used at checkout. Applicable taxes, if any, are shown at checkout.

Before you begin

Do I need to code?

No. You need TFIS Fluency and a task you can explain and verify. The coding-agent week teaches you to brief and review a small practice project in an approved coding tool.

Is this a live cohort?

No. It is an on-demand course with a suggested 12-week rhythm. Lessons are written, interactive learning materials with examples, labs and quizzes. You can move at your own pace.

Which API can I use?

Use an OpenAI-compatible Chat Completions endpoint and a model available to your key. MyToken is the default. Agent mode requires tool-calling support; image exercises require a vision-capable model.