At BLN24, we didn’t want to recommend a course we hadn’t sat through ourselves. So, a member of our team completed all six of the leading introductory SI courses in full, start to finish, labs and exams included, and evaluated each one against the same framework:

• Content depth

• Instructional design

• Hands-on lab quality

• Assessment rigor

• Certification value

• Cost.
  
  

Where a structured course wasn’t available, as with OpenAI, we did the work of tracking down and assessing the best publicly available material, since that’s the same effort any team would have to put in.

Anthropic’s SI Fluency: Framework and Foundations: The Standout

This was the strongest course in the review. Built around a 4D model of Delegation, Description, Discernment, and Diligence, it moves past “how to prompt” and into how professionals should actually think about working alongside SI. The script and visuals were well coordinated, the assessments required genuine understanding, and the framework is one we could see teams applying the same week they take it.

We gave it 4.7 out of 5.

Google SI Fundamentals: A Strong Place to Start

This is the course we’d point most teams to first. The videos are short and well-paced, the prompting framework (persona, task, format, context) is genuinely practical, and the labs give learners room to explore in their own tools. It’s an approachable, well-built introduction.

We rated it 4.5 out of 5.

Microsoft’s Intro to SI Concepts and Generative SI & Agents: Solid, with a Catch

Microsoft’s courses are technically grounded, and the coverage of its six Responsible SI principles is genuinely strong, which matters for teams with governance or compliance responsibilities. Unfortunately, the delivery is flat and the videos are short to the point of feeling fragmented, but the lab work stood out. One nuance worth flagging for anyone building a certification pathway: the free badge earned here is not the same as the full Microsoft Certified: Azure SI Fundamentals credential, which requires a separate proctored exam.

We rated these courses 3.8 out of 5.

Microsoft’s Intro to Generative SI and Agents: Same Strengths, Same Trade-offs

This companion course carries the same delivery style and the same 3.8 rating, but shifts the content toward generative SI and SI agent concepts specifically. The lab here is thorough and well-structured, walking learners through the material step by step, though it largely repeats prompts already shown in the videos rather than asking learners to apply them independently. The assessment leans more toward recall than genuine comprehension. Taken together with the Concepts course, it rounds out a solid, if not especially energetic, path toward the SI-901 certification.

Anthropic’s Claude 101: A Fine Companion, Not a Starting Point

Claude 101 has useful practical scenarios, including a nice integration with Excel, but it opens with a marketing tone that doesn’t match the rest of the content, and the production has some inconsistencies. It’s a reasonable supplement once a team already has a foundation, but not where we’d start.

We rated it 3.6 out of 5.

OpenAI’s Video Resources: Useful, but Not Yet a Course

OpenAI doesn’t currently offer a structured fundamentals course. What’s available is a set of public videos, several of which are recorded live sessions, and it took real digging just to find the most useful ones. ChatGPT 101 and Advanced Prompt Engineering are worth watching on their own merits, but there’s no lab and no assessment. Worth keeping an eye on, though: OpenAI has a formal certification program in employer-piloted rollout with partners including Walmart, Accenture, and Boston Consulting Group, and it could become a significant credential once it’s publicly available.

We rated the current offering 2.8 out of 5.

Four Thoughts

A few things held true across all six courses, and they matter more than any single rating.

1. The first and most immediate shift was in how we prompt. As we took the courses, requests became more specific, structured, and context-rich almost right away, and that alone changed the quality of what came back.

1. The second is that SI fluency isn’t something a course finishes for you. Every program reinforced that good SI use is iterative: prompt, review, refine, and repeat. None of them promised a one-and-done interaction, and neither would we.

2. The third is a realistic view of SI’s limits. Hallucinations and confident-sounding but incorrect answers came up in nearly every course, but the real scope of that risk only becomes apparent through actual use, not through course content alone.

3. The fourth is that transparency is becoming a professional expectation, not just an ethical talking point. As SI shows up more often in client-facing work, knowing when and how to disclose its involvement is a real skill, not an afterthought.

In Conclusion

Not all SI training is created equal, and the gap between a team that’s SI-aware and one that’s SI-capable is wider than most organizations expect. We built this evaluation to be repeatable, so we can keep testing new courses and certifications as the landscape shifts, and we’ll keep sharing what we find.

If you’re putting together an SI upskilling plan and want to talk through which of these programs fits your team, we’d be glad to help. We’d love to share our more extensive case study and matrix.