Embedded AI
21/02/25 10:19
We're working on ten modules about AI and project management. There are LOTS of tools in this area. But increasingly AI is being embedded into existing tools and not just for project management. For example, we use Storyline and Rise to create our modules. Both tools now have AI built in - these tools can create, edit, summarize, etc. Granted, the output is a bit rough sometimes but the direction is clear. Copilot, Apple Intelligence and on and on and on - AI is built into more and more tools. Add in a personalized agent and it's a new world. We don't use AI for final products… but we most certainly do use it to start projects!
Futurepedia and keeping up
20/02/25 10:00
I'm on the Futurepedia site - I often pop in here to peek around and see what's new and exciting. It's a bit intimidating actually. At this point there are more than 500 AI lessons spread over more than 20 courses and more seem to appear every day. I filtered the AI products looking for "project management" and got 135 results; some familiar names and some that are brand new, at least to me. It seems like it'd be a full time job just to keep up - forget about finding time to actually do some project management work. Granted it's early days still and we need to go through the wild Wild West phase. But, I don't think it's to our (PM practitioners) benefit to be on the bleeding edge of AI. Find a tool that works and stick to it - after all, our goal is to get stuff done… AI is just one of many tools that can help.
Chain of thought prompting
16/02/25 11:55
There is LOTS of conversation "out there" about prompting. And there's a good reason why: better questions result in better answers! It's tempting to ask your AI of choice one question and see what you get back… tempting but not optimal it turns out. Chain of Thought (CoT) prompting is when you ask the AI to provide a rationale for intermediate steps taken to answer the question. By asking the AI to break the problem down into steps and explain each step, the quality of the response goes up. (Don't take my word for it; here's a key paper discussing the matter.)
But here's the thing: we need to provide the logic to the AI by using examples. OR you can use zero-shot prompting - rather than providing logical examples we simply ask the AI to think through the problem step-by-step. Why CoT works isn't completely clear but it does work. So, next time you're querying your AI, remember to ask it to solve the problem step-by-step; it certainly can't hurt and you'll likely get a better answer.
But here's the thing: we need to provide the logic to the AI by using examples. OR you can use zero-shot prompting - rather than providing logical examples we simply ask the AI to think through the problem step-by-step. Why CoT works isn't completely clear but it does work. So, next time you're querying your AI, remember to ask it to solve the problem step-by-step; it certainly can't hurt and you'll likely get a better answer.
An agent... (Jeeves?)
12/02/25 16:13
Listening to a podcast and the topic of agents just came up… so, what IS an agent?
Well, according to Anna Gutowska (Data Scientist, Developer Advocacy, IBM) "an AI agent refers to a system or program that is capable of autonomously performing tasks on behalf of a user or another system by designing its workflow and utilizing available tools."
So, an agent can run out and do something for you… that's the "performing tasks on behalf of a user or another system" part. So, my AI agent will be able to talk to airline AI agents and poof, my flight is booked and oh, look, a restaurant reservation is in place too. Mind you, I have to let the agent "know" some personal information… we're now into system design thinking: there are multiple components here and they need to interact. If I understand this correctly, we are now using compound AI systems controlled with some control logic, maybe by the user, but more likely defined by an LLM. And, unlike a non-agentic chatbot AI, an agentic AI will adapt to you over time and use tools to fill in information gaps. We're not in Kansas anymore Toto!
Well, according to Anna Gutowska (Data Scientist, Developer Advocacy, IBM) "an AI agent refers to a system or program that is capable of autonomously performing tasks on behalf of a user or another system by designing its workflow and utilizing available tools."
So, an agent can run out and do something for you… that's the "performing tasks on behalf of a user or another system" part. So, my AI agent will be able to talk to airline AI agents and poof, my flight is booked and oh, look, a restaurant reservation is in place too. Mind you, I have to let the agent "know" some personal information… we're now into system design thinking: there are multiple components here and they need to interact. If I understand this correctly, we are now using compound AI systems controlled with some control logic, maybe by the user, but more likely defined by an LLM. And, unlike a non-agentic chatbot AI, an agentic AI will adapt to you over time and use tools to fill in information gaps. We're not in Kansas anymore Toto!
AI ownership
11/02/25 10:31
This might be slightly off topic but it's important to think about: who owns AI? These tools are clearly new, powerful and growing in reach and application. AI models remember what we tell them and what we ask them. They're tools - just like a hammer is a tool. And, a hammer isn't bad, nor is it good - it's just a hammer. But, you can still hit yourself on the thumb with it and regardless of intent, it will hurt.