I saw the movie Le mystère Picasso when I was 11. The artist became my hero. The infatuation was short-lived, but I was briefly captivated by the master’s creative panache and his intense, passionate, dramatic artistic vision. I wanted to embrace it as my own.

The episode was one of those childhood learning experiences which in hindsight is vaguely embarrassing, but was both character-building and illuminating, and is therefore invaluable: I learned something.

I can’t sketch very well, so I quickly turned to impressionnisme abstrait (at least, that’s how my highfalutin self described my watercolour splatter). Even then I couldn’t stay within the lines, but I kept trying.

One day my previously encouraging but now just sympathetic father took me aside for one of those talks. “Guillaume,” he said, “everyone is good at something. One of the fundamental keys to success in life is to concentrate on those things at which we are good, and leave the things we’re very bad at to other people.”

I could see what was coming. “I believe, son,” my father continued, “that you are very bad at art. Perhaps focus on something to do with numbers, eh?”

I immediately decided to study business.

There’s a lesson here about MGAs. If any one of them has any hope of succeeding in a softening market, it will have to be very good at something. Maybe it’s their distribution, or their unparalleled underwriting skills in a particular niche class of business, or the levels of service they provide to brokers.

It’s the reason they’re in business in the first place. It’s what makes them a unique business, their USP.

Everything that doesn’t contribute directly to the USP, from telephony to sanctions checks, is just a cost. That means the more you automate any part of your operation which isn’t part of your USP, the greater your margins will be. It’s as simple as that.

If your USP is your underwriting expertise, for example, you cannot automate it. You may be able to improve your risk assessment marginally through automation, but the gains will be commensurately marginal. That would almost certainly be beneficial, and therefore worth doing, especially as competition hots up. However, a very much bigger prize is hanging low from the tree.

In our example, it’s the automation of all the non-underwriting tasks and processes. You won’t manage to hand over 100% to the bots, but if you can mechanise most, your costs will fall materially. The impact on your bottom line will be very much bigger than an effort to get AI to do your job for you.

I thought of all this after reading an excellent LinkedIn blog by Jo Zuk from Altamont Capital, The End of Acquisition Cost: Why the Next Great MGU will be Agentic. In it he argues eloquently that the NGMGU (specifically in the micro-SME coverage space) will be an “Agentic MGU” which “uses advanced AI agents to autonomously classify, underwrite, price, bind, and service standardized commercial products at near-zero marginal cost.”

Unfortunately it won’t be the NGMGU for long. The tech will almost certainly be  cloned swiftly, and deployed by another team that has one of the magic ingredients (in Joe’s example, almost certainly distribution).  

Herein lies another lesson. Your USP isn’t your technology. If it is, you need to get one of the other ones on board quickly. If you don’t, someone who has one of the three key unique characteristics is bound to get similar or better tech soon, and transform you into an also-ran.

Herein lies the simple secret to MGA success is simple. Do what you’re good at, and automate the rest. By implication, leave the tech to specialists.

I hope that’s clear, because you don’t want me to draw you a picture!

Guillaume Bonnissent is CEO of Quotech.