Huddle 003:
AI and Strategy

We gathered a collection (what is the collective noun for strategists? a disagreement? a noodle? a discombobulation?) of our members together this week to chat AI.

How are we using it in our practice?
What opportunities do we see?
What risks does it present?

I asked AI to summarise the notes, and you can see that version of here, but I thought I'd do a human take too. You can let me know which you think is more valuable.

Accelerating inputs.

"AI is good at avoiding the blank page paralysis, it's almost like sending your interns or junior strategists on a treasure hunt."

Lots of us were using AI to accelerate much of the day to day work we do - going out there to scour across vast amounts of research, find what others are saying and doing, automating much of the "filling" of our brain, finding stimulus and and inputs to the process.

LLMs are great at that first pass, gathering lots of materials, creating initial hypothesis, initial steps or drafts, the first stab at things. Hallucinations seems to going away, and tools are better at citing sources, and finding references which would otherwise be missed or take hours to hunt down.

Tasks such as researching large datasets, drafting proposals, or finding secondary research sources—can now be handled more efficiently by AI tools, saving time for higer-value tasks.

Bland

However, it isn't neccessarily great at finding the gold. Most folk agreed that if you were asking it to surface new, interesting and different ideas - it would often regress to the mean, due to the way in which LLMs work.

"AI is good at delivering expected results, but it's bland... it doesn't crack the challenge."

AI tends to surface the expected, the obvious, what's already known and understood. In many ways, people felt it's a great tool for making sure you're not sharing the obvious. If an AI is highlighting something, then perhaps that's something to avoid? Going beyond the obvious is essential for most of our work.

That said, AI can be challenged, and never gets tired of being pushed to refine or repeat and come at things from another angle. But some of our crew also felt, the longer the conversations you had, the poorer results.

"I've noticed that it starts drifting, which is very dangerous. There’s diminishing returns the longer the chat becomes."

Good enough

There was a sense that there's a risk that AI could be 'good enough' to remove the need for junior strategic work, or even clients using the tools to do significant portions of the work themselves, and if not having a deep expert to interrogate or challenge it's outputs, that responses will be good enough to use.

"AI is a good start, but unless you're an expert in the category, it won't solve the challenge."

It was highlighted that this problem of 'good enough' isn't AI's fault though, and the concept of something being good enough has been eating away at the strategic and creative process for much longer than we've had ChatGPT - and when the market is in a bad place, it creates commercial pressures on everyone to accept good enough.

Strategist + AI

This opens the opportunity for the powerful combination of Strategist + AI tools, rather than replacing strategic specialism, supercharging it. Knowing how to use the tools, how to prompt, how to challenge, knowing when to use the tools, and when to return to human experience.

"AI is like Photoshop for the rest of us... it doesn't replace expertise but helps with the craft."

AI is being seen as an enabler, a supercharger, still requiring expert intervention to get the very best from it.

Ethical and Environment Concerns

Questions around the environmental impact of AI were discussed to - the volume of water and electricity usage for even the simplest of queries were of concern to the group, and whether it was morally acceptable to use for the most mundane/outsourcing of tasks, and as their adoption become even more widespread - what questions must we ask of when, where and how it should be used.

Summary

Overall, there's optimism and excitement for AI, we're on varying levels of use, some are going further and creating and training their own AIs on their body of work and approach, and using it as a powerful coworker.

And for independent strategists, who often work on their own, having a second strategist to bounce ideas off, to challenge our own thinking (or to stop us overthinking), sounds invaluable.

To be able to scale up the work we do working independently, so we can deliver more, and to create opportunities to play the role of trusted advisor for clients who are using AI, but still need expert navigation and curation of its outputs, strategists leveraging AI will be heads above those who are not.

Outside Perspective: Huddles

Huddles are a series of conversations with independent strategists and industry experts on topics ranging from strategy to business, culture to creativity, categories to commercials.

We assemble a group of independentspecialists, and let them debate and discuss for an hour.

Interested in hosting a huddle for your business, team or project? Get in touch.

</div>