Andrew Jacobs explores how creatively thinking about technology and what we want from it can enhance not only your use of AI, but your strategy too

I like to tinker.  It comes from when I was a kid; I used to take toys apart and tried to put them back together, or more likely, integrate them somehow. Think Toy Story’s Sid Phillips but much less violent.

As I’ve got older, I’ve updated laptops, built PCs, and played with settings on things to see what other purpose I can use something for. It’s something I’ve carried on into my working life – I play with blog posts, articles, and podcasts over many revisions before I publish them. And then I still want to change them.

No AI is going to be able to help until the learning profession thinks differently about data

I think it’s a way of learning I have; putting things together to compare and contrast and create a frame of reference to help me understand. I do it in conversations with people; you’ll introduce a concept and I’ll search for something to link it with. I do it with people too – mention a name and I’ll seek out a connection. It’s called bricolage and confirms the best ideas aren’t born; they’re grown.

I’ve been doing the same with ChatGPT recently. I was interested in Jo Cook’s approach to testing out how effective ChatGPT was (in this AI-written article and this commentary), and the results were, well, a little underwhelming.

I wrote about this recently for People Management magazine:

“If I was to offer you a way of working which might cut your time spent on some of your work by half, you’d most likely jump at the chance to use it. In November 2022 that happened, and a single tool was launched which has cut through like no other recent technology.

It was when Open AI released version 3.5 of their platform ChatGPT, and it changed the world. By January 2023, 100 million users had signed up and since then we’ve seen it become commonplace in workplace conversations about learning, creativity, performance, and the speed of delivering content.“

As I said above, taking tools and seeing how they work together is the way I learn and develop but it’s clear that not everyone is the same as me. I’m incredibly lucky to have the space to think these things through and, to save you time, here are a few ideas about what you need to do to use these technologies well.

Understand the data you need

Learning and training professionals have masses of data, but we rarely investigate it fully to make sure we get the best value from it. This means taking a different approach to data and no AI is going to be able to help until the learning profession thinks differently about data. For example, you have a 75% completion rate. That’s a meaningless metric out of context and simply a headline figure. Have a look at things like:

• Ratio of mandated users and voluntary entries into the learning experience to the completion rate. If people are ‘forced’ to undertake the learning, do they complete it?
  
  
• Pace and cadence of completion – was the content completed over a longer period or rushed at the end? The pace users complete the material might lend itself to a view of how useful, challenging, relevant, etc the learning is
  
  
• What relationships exist between test scores and completion data? If people can complete an assessment without completing the material, what value is there in the material
  

Understanding data is a critical core skill which learning functions need to be better at, and no AI technology is going to make it work if you don’t have a reason to be using it.

More and more content

It seems that tools like ChatGPT, Gemini (formerly Bard) and other Large Language Models can do a reasonable job of producing content quickly.

There are, however, a couple of issues with this. Firstly, the ability to produce reams of content and place it in front of people isn’t the way which people want to learn. You can try it yourself; go to ChatGPT and ask it to produce a piece of learning content about time management. I’ll wait.

I did this using ChatGPT 3.5, 4.0 and Gemini and got three very different and simple responses. All valuable content if applied correctly but these kinds of prompts and content are being used as ways of crafting internal content.

Think about the end we want – improved performance – and use these tools to achieve that. I asked each of these tools to craft an assessment rubric, based on performance in the workplace to develop skills in time management. I’ll let you try this and see how different the responses are.

Again, three very different but much more useful responses. We can’t use these tools until we understand how to use them properly.

The three B’s of strategy

I’m lucky to work with people and organisations who appreciate the need for designing learning in a strategic way. That means being able to describe the way you want to support performance through learning now, next and in the future.

One of the common questions I’m asked is which technologies to buy to meet strategic needs. My response is to always highlight that to Buy should be the last step in the process and 2 other B’s come first.

The first step in developing the technology as part of your strategy is to identify what you can Borrow at a nil/neutral/low cost. In many cases, it’s about switching on some features in your existing systems. For example, if you want to develop communities in your organisation, why not use Microsoft Viva Engage (formerly Yammer)?

The second B in your strategy should be to Build. Like all good buildings you’ll have a proposed architecture and understand how the thing you need to build connects with the things you have and what needs to be put in its place.

The last B is Buy. Now you can decide what you need and are in an place to make an informed choice instead of an educated guess about what technology or support you need.

Where do we go from here

As I said in People Management:

“The use of AI tools in workplace learning represents an opportunity and a challenge. To successfully navigate these challenges, L&D professionals must adopt an approach that considers the characteristics of their employees and teams.“

AI tools can help – the italic text above was written by AI – but we need to get the basics right before we jump into using AI tools.

---

Andrew Jacobs is CEO at Llarn Learning Services