Deccan AI Experts’ Post

I genuinely think the difference between getting average results from AI and getting genuinely useful work from it usually comes down to one thing: context. When you give very little information, you tend to get generic answers. When you properly explain what you are trying to do, share the background, the constraints, and what a good result looks like, the quality of the output improves a lot. It starts to feel less like a chatbot and more like working with someone who actually understands what you are trying to do. The main skill is not writing clever prompts but it is explaining problems clearly to the AI. If you can explain the goal, the situation, what you have tried, and where you are stuck, AI becomes much more useful.

Peer review isn’t just for humans anymore. It applies to AI too — and honestly, I’m enjoying it. Reviewing AI output has become its own learning loop. Every pass reveals a new insight, a new angle, a new way of thinking. Turns out, when you peer‑review a machine, you end up upgrading your own clarity, structure and reasoning. AI produces fast. Peer review produces wisdom. And the combination is surprisingly powerful

Doing a session on CoPilot today to the Applied AI Members! This really applies to all AI use...People first AI! The One Rule That Applies to Every Session, Every Time Copilot produces. You decide. The output is a starting point, not a destination. Your expertise, your relationships, your knowledge of the room, and your professional accountability do not transfer to the AI. They stay with you. Always. The fastest way to lose trust in your organization is to let an AI tool produce something that goes out under your name without your full review. The fastest way to build credibility with these tools is to use them well, review them honestly, and make every output better than it was when it came back from the prompt. That is the skill. That is what these sessions are building.

I recently came across a short film created by Léo Cannone using Google’s AI tools — and it’s impressive. The film showcases how much can now be done with the help of AI, really reducing the barriers to entry. BUT tools alone cannot produce high quality content. To stand out from the “average”, that the AI tools generate, the creative direction still needs to come from a human. Full documentary in the comments.

My biggest concern about AI. As it becomes a daily tool, I have seen what I believe to be errors in the proper use of AI. Perhaps it’s the Gen X in me, but it reminds me on how newer generations started to lose their ability to do basic math due to dependency on calculators. I believe the problem is that people rely on AI solving everything without propper directions. In order to get the best out of AI, you need to understand what you need done, direction of analysis or expectations on results so that AI can do the heavy lifting for you. But if you just input data without guidance, AI will give you options that may not be aligned to what you really need. Thoughts?

Strong content makes strong AI. And if you want AI that supports your organization instead of surprising it, the work starts with building that foundation.

How would we know if the L&D field is actually getting better? Serious question. We talk a lot about where the field is going. More AI. More personalization. More skills. More workflow learning. More strategy More value More seat at the table But what would count as evidence that the field itself is progressing and advancing?

AI doesn’t fail. Systems around it do. The model does exactly what it was designed to do: predict the next token. The real problems start elsewhere: – inconsistent data – missing context – no ownership of decisions And then we hear: “AI made a mistake.” No. The system wasn’t ready for AI. Until we treat AI as part of a decision system. not a magic tool we’ll keep scaling errors instead of value. I’m not learning AI as a tool. I’m learning the systems where AI actually makes sense. #AI #LLM #SystemsThinking #AIImplementation

I have been noticing an interesting trend. A lot of companies are talking about AI. Very few are using it in a way that actually improves how they work. The assumption is that AI will make things faster and more efficient. But in practice, it often highlights what is already unclear. Unstructured inputs. Inconsistent processes. No clear ownership. AI does not solve these problems. It follows them. And when it does, it tends to amplify the gaps rather than fix them. The teams that see real value are not starting with AI. They are starting with clarity. How work flows. Who owns what. What good output actually looks like. AI works best when the system around it already makes sense.

In an attention economy, what if learning looked more like a text thread than a course? The Department of Labor just released a simple, text-based way to build AI basics: https://lnkd.in/gjbN9gA4 It’s not complex and that’s the point. A few things stand out: • Meets people where they already are with zero friction • Builds skills and habits through small, repeatable interactions • Makes AI feel accessible, not intimidating • Feels more like a conversation than a lesson This isn’t comprehensive, but it’s a signal. Lower the barrier. Keep it simple. Build momentum. Is this how we make America AI ready? #AI #FutureOfWork #LearningDesign #EdTech #WorkforceDevelopment #DigitalLiteracy #Innovation #Upskilling