Enstra Consulting and Training’s view on BOOST25: The mind-set change needed to successfully reap the benefits of AI

We had the pleasure of welcoming Peter Franklin of Enstra consulting at our BOOST25 event in Antwerp. As one of our sponsors and long-term partners, Peter shares his reflections on the conference, exploring how AI is reshaping the utilities industry, why a mindset shift is essential to harness its full potential, and how these insights inspired Enstra to identify new ways of applying their expertise in logic-based thinking and training:

BOOST 25, Ferranti’s conference in Antwerp on the future of utilities across Europe, proved to be both a successful social event, and a thought-provoking learning adventure.

On the social side it brought together partners, customers , and the Ferranti team. On the thought-provoking side, the speaker and panel contributions highlighted both how AI is changing the industry, but also how it requires a cultural shift to harness its power successfully.

We, Enstra Consulting, were there as sponsors of the conference, having for many years provided Ferranti with business development support services and a monthly update report on developments in the energy and water markets in the UK which are relevant to Ferranti and its customers. This information enables Ferranti to ensure that its MECOMS offering is aligned to both current and future needs of the marketplace.

AI’s Growing Influence on the Utilities Sector

Whilst  all the speakers were excellent, one topic provoked more food for thought than the others. That was the impact of AI on suppliers and customers. It became clear that AI would have, or more correctly is having, a significant impact on all aspects of the industry…..and, in what was an eye opener for me, would require a mind-set change on the part of users.

A Required Mind-Set Shift: From “The Computer Is Right” to “AI’s Best Guess”

That mind-set change is in the realm of Large Language Models (LLMs). What is needed is  a shift from a belief that “the computer says it – so it must be right “ which is correct in the absence of programming bugs – to “this is AI’s best guess”. It was the Co-pilot lead from Microsoft that made the point that all LLMs do is look for “the word that has the highest probability of following another in multi-dimensional space”. This means that LLMs can find themselves guessing incorrectly and hallucinating. What is needed is a human to check the answers.

The focus will be on how humans interact with the LLMs, how they both construct the questions posed, the prompts, and how they verify the answers.

Microsoft’s Ninja Method for Crafting Effective Prompts

On the former, Microsoft shared the Ninja method of writing prompts – which is to give the model some context without asking for an answer and then to ask it to interview the prompter regarding the questions that need to be asked to answer the question that needs to be answered. You then provide the answers as context and get the AI to craft the answer. So, for example, if you wanted a marketing plan you  would ask the AI to interview you regarding the questions that need to be answered to create a robust and credible marketing plan. You assisted by the AI would answer those questions and only then do you ask the AI to write the marketing plan.

On the verification side, humans will be needed to fact check the answers.

That made us think.

As well as consulting we run a training business in which we train executives in corporates and consultancies how to craft compelling communications using logics crafted using Pyramid  Thinking Plus, an evolution of the Pyramid Principle developed by Barabara Minto of McKinsey in the 1980s. Once you know these rules of logic, you can use them to check the coherence of the logic of the LLM output – and hence its ability to convince your audience with the communication it has created for you.

Following BOOST, we now have a new “Use-Case” for the deployment of the logic rules we teach. Testing AI outputs for logical consistency. This will allow humans to write subsidiary prompts to refine the AI output and make the communications produced truly compelling to their audience.

Thank you, Ferranti, BOOST was the catalyst for triggering our realisation of the importance of Pyramid Thinking Plus for managing LLM outputs.