Waiting for AI’s magic - Have I become a cynic?

A man crosses the stage and pulls a rabbit from a top hat – the audience is silent and unimpressed.

A man in a cape crosses the stage, mutters some magic words and, with a flourish, pulls a white rabbit from a top hat. In the audience, there is an awkward laugh and a smattering of polite applause. It’s been done before.

A man in a cape crosses the stage, mutters some magic words and, with a flourish, pulls a Ninja™ CREAMi® Deluxe 11-in-1 Ice Cream and Frozen Margarita machine from a top hat – the audience goes wild, a 5-minute standing ovation ensues, at intermission the lobby bar sells out of tequila.

I am always searching for the next big showstopper. In our industry today, that means AI. 'Innovation' and 'New' sell, and I want to find the best acts that will support the best show in town. What can I say – I am a little competitive.

Part of the process is critically analyzing the vendors, tools and services that are available, assessing constructively if there is a strong commercial use case for integrating them and, if so, ensuring they will go on to ultimately improve what value we offer our clients.

Sometimes, it feels like there is a new AI super-tool hitting the market weekly, or some previously underutilized way that machine learning or LLMs can be applied to our industry being relaunched just as frequently. It is overwhelming at times.

And yet I feel like I am missing a trick somewhere – but where - I don’t know.

AI once seemed like a broad frontier of fantastical futures, but more recently, for me, has narrowed down to small areas of niche added value. It feels like the initial hype is waning and that the one big AI showstopper might not materialize.

I think: “There are legitimate reasons why the AI Magic show may never get on the road.” But then I catch myself, and start to worry, have I stopped being a critic and become a cynic instead?

The Pledge

Before AI can perform, however, it must prepare. It must have a rather large and overly intricate script, learn its lines, have its sparkly costume just-so, a glamorous assistant (or 10), and then it must rehearse and rehearse some more. There is no guarantee it will go right on the night, and, frankly, the resulting show can lack the authenticity of a real and experienced performer.

From the audiences’ perspective, we are not seeing anything new in fact, we are seeing something with potentially less “wow” factor than before.

Many AI use cases are singular in application – once AI has learned the ‘Rabbit out the Hat’ it can’t jump, unprepared, into the ‘Bullet Catch’.

We serve clients across many industries, markets and customer profiles all with unique goals and realities – the ability to comprehend and evaluate these nuances is vital. Not being able to easily pivot and retrain these tools creates a huge barrier when it comes to investing teams and resources.

In an industry that depends on speed and cost efficiencies, the amount of prep work and investment for what amounts to, maybe, a lukewarm performance could be why we are seeing a struggle to place these tools in a more prominent billing on our product rosters. AI has great potential, but the majority is either falling short on relevance or, maybe more so, unfeasible in costs, time resource investment and perceived value-add.

The Turn

What I think is so challenging about AI is that it is a paradox of progress. Machine learning and AI unlocks the potential to do much at speed, ultimately moving us forward. At the same time, paradoxically, it is also preserving the thinking that already exists. It is not innovative in itself. What is taking time to develop is how we advance our capabilities as researchers.

I believe it is a good thing that, in order to run, we must learn to walk first. So rather than seeing the lack of immediate commercial application for new tools as a bad thing, see it as an opportunity for the creation of something improved. Try and see it as a vehicle for us to expedite our process and then challenge ourselves to elevate our findings.

In the pursuit of efficiencies, we sacrifice nuance in our methodologies, either in interview real estate or analysis. AI could help us as it has the potential to review swathes of data, build out the connected or synthetic data space’s offerings and transform the industry. The latent potential in consumer online profiles and behaviors could be the key to unlocking the insights for future products or services.

We as insight specialists will need to be able to weigh the relevancy and import to offer broader, actionable and more impactful insights.

The Prestige

I think it is easy to fall into the trap of innovation inertia, with the fear and pressure of needing to deliver something groundbreaking holding us back. Instead, we stagnate. There is unmet disruptive potential, but to recognize this, it’ll need a more involved, timely and academic review period than previously thought.

In the meantime, I will focus on keeping an open mind to the possibilities, leveraging a critical eye (with a healthy amount of cynicism) on what is out there, but will also lean into collaboration, adoption and learning opportunities in research, celebrating all advancements.

By taking small incremental steps, reinventing a wheel or two, challenging the norms, I think I can still find my showstopper. For now, however, I may just have to accept an under-stated twist on a classic – besides, what’s wrong with a good ol’ rabbit out of a hat anyhow?