Hey, have you heard about AI!? It’s all the rage, but then maybe it’s just another passing fad… kinda like the Pet Rock. Kidding aside, the thinking computer is here. It’s changing things as we speak. I am a bit skeptical of the predictions. It’s been suggested to me that consulting is dead and everyone is just going to use AI in the future. I suppose this colors much of my skepticism. I am skeptical of the quality of AI written documentation and the answers AI provides. I think it is appropriate to share Chalmers St.’s experience with AI. I’d like to hear your feedback.
I get plenty of questions about our experiences with AI, specifically, where we think it is useful. We’ve experienced AI in four different ways and I will explain each:
- AI developed marketing materials
- AI written documents
- AI developed training
- AI enhanced templates
The area where I have seen the most benefit is in creating marketing material. AI is best for the things that you are worst at. I am not good at marketing. Specifically, developing a flow of messaging and visuals to sell a concept. For example, I was recently told that my PowerPoint slides are content heavy with lots of bullets. I responded, “Thank you.” It was not intended to be a complement. To improve in this area of deficiency, I have successfully used AI to outline a flow of messaging and develop some slide decks. AI has also assisted me in taking a bulleted list of my thoughts and translating them into a full blown business case. This has been powerful both from saving us the cost of hiring a marketing firm and creating some materials that we would not have otherwise created. Marketing communications is an AI win for sure.
Let me clarify though, it is a win only on the messaging side of marketing communication. AI is not so great at building visuals. We needed humans to turn the information into graphics that conceptualize our offers into a cogent picture. I know AI can create pictures, but not well. Put that one in the lose column for AI. Let’s call marketing a mixed bag for the capability of AI.
Another AI application is building project documentation. I find that AI is great when you feed it a list of bullets and ask it to create a document. I have prompted my way to really well written executive summaries. It’s a time saver. Though our use in this area is still limited. I can tell when one of my team members has written a document using AI.There is a mechanical aspect to the writing. It’s always oddly broad and kind of thorough, but in a generic way. It is not human. Yes, I have seen all sorts of AI fakes. They are good, but I can tell and my guess is that you can too.
We’ve used AI to generate statements of work and executive summaries. We have not used it at all for more technical documentation. Frankly, if I can tell AI was used, that means my clients can tell. This removes the personal touch from our consulting. This is not acceptable. I can confidently say that AI won’t be writing my newsletter any time soon. When it comes to written documentation, technical or otherwise, it’s a time saver. It is useful for idea generation. However, it does not come close to replacing a human.
We’ve used AI to develop illustrations for training materials. We were amazed at AI’s ability to create illustrations for us. I was not directly involved in the creation, but my understanding is that our team prompted an image generator to produce “comic strip” illustrations of people in different work environments. One of our content developers created the storyline and the AI (along with a human prompter) created the illustrations. This was absolutely fantastic. I was told that we would have spent $50,000 on an illustrator to do this work. Now before someone complains that AI is stealing the job of starving artists, the reality is that the choice was not between an AI and a human. It was between AI and clip art or nothing at all. We would never have spent the money on an illustrator. This is an important point because technology driven creative destruction is over emphasized. AI technology is giving us products that would not have existed; it is not replacing anything. There is no destruction, just creation.
We have not experimented with AI to create training content. I suppose we should give it a shot, but given what I already know, I predict it will be an overly mechanical result. I don’t trust it to explain the concepts that we practice in Continuous Improvement.
We’ve integrated AI into our problem solving templates. Two examples are in our Problem Statements template where the AI checks to verify that the user properly answered the 5 Ws and 2 Hs and the 5 Whys template where the AI checks the logic of each Why verifying that it connects back as a symptom of the previous Why all the way up to the original problem statement. These are useful. They are beta tools. They are not good enough to replace the Black Belt facilitator that runs our CI Mastermind sessions. Case in point, I tested one of our standard examples and the AI correctly approved the logic. When I varied the logic the AI correctly disapproved. Great! I then tried the classic example of the Jefferson Memorial… nope. If you are not familiar with this example, look here. The AI did not like that logic. It’s not good because we’ve been teaching 5 Whys this way for 25 years. Not just me, mind you, but just about everyone teaching 5 Whys. Admittedly, it could have been me as a prompter, but regardless, the AI was not helpful in coaching on logic. It is not a good coach at this point.
So, I don’t buy the hype, but I do see the future coming. I see our current stage of AI similar to the internet in 1995. It’s interesting. It’s a big idea for sure. We’ve been talking about AI for decades and finally the potential is beginning to come into view. The next 15 years will bring a constant wave of new and amazing use cases. But the role of the human will not change overnight. This will take time.
At Chalmers St. we will continue to experiment. Our role is helping our clients navigate the changing landscape of Continuous Improvement. AI is a factor. In the next few months we are launching our most exciting AI experiment to date. I won’t provide details now, but you can expect a newsletter in January where I will share our work and what we found. Stay tuned and keep experimenting!