Written By Adam T. Hark, Managing Member
Back in the summer of 2023 I had the opportunity to evaluate a small company who used machine learning to analyze the payment processing statements of merchants. Though this occurred roughly 6-months after Open AI released Chat GPT 3.5 – the version that went viral and jump started the AI race we’re still running in today – this company’s system did not use a transformer architecture with an attention mechanism. Instead it used a good “old fashioned” recurrent neural network for classification purposes. The company had built and trained a machine to ingest merchant statements of all shapes and sizes and formats, extract relevant data points, and map the same to standardized fields that would then be used to find pricing inefficiencies that competing payment processors could avail themselves of to win over that merchant’s payment processing business.
These founders created real AI, architected and trained properly to address real marketplace pricing inefficiencies through automation, analysis, insights and recommendations. Basically, this little AI company was the poster child for how to properly build and design AI for profit – work backwards from the problem that needs to be solved, build the right algo, train it with large quantities of the right kind of data, and sell it as a subscription-based solution to a very narrowly defined TAM (the same strategy Peter Thiel advocates for on behalf of startups in his book Zero to One).
Well, this past week that little company, Fee Navigator, became part of a much larger concern when it was acquired by payment processing market leader NMI. It was a highly strategic buy due to the efficacy of the AI solution and NMI’s scale with payment processing partners and their respective merchants.
Is there a particular point to this transaction aside from the fact I personally have been following Fee Navigator’s journey for a few years now?
Of course there is.
We live in a day and age of competing narratives regarding AI in payments and fintech. On one hand, there’s the ubiquitous hype around the advent and promise of agentic commerce, and the attendant panic among fintech and payment concerns that if they don’t embrace any form of AI ( agentic or non-agentic) in their respective orgs, they’ll find themselves looking into the maw of oblivion. And on the other hand, there’s the continuous flow of reportage and studies that show that many early adopters of AI have little to show by way of return on investment.
This ultimately leads founders (and investors and acquireres to ask, “what’s the right AI strategy for my payments or fintech company?” The answer in its simplest form is to first identify the problem, inefficiency, and/or challenge that AI could possibly make better, and work backwards from there! This of course also requires founders to understand the different types of AI solutions out there, what they can do, and more importantly, what their limitations are. Once the problem and solution are sketched out, then the design, build, and training of the AI solution to be implemented begins.
Fee Navigator is a great “AI in fintech” story because they identified a real problem, and built a real solution. Congrats to Fee Navigator’s founders and kudos to NMI for recognizing its value to their own payments platform (fyi…there were many major industry players who did not).
How do I know?
Well, in addition to following Fee Navigator’s journey since 2023, I (we at Wellesley Hills Financial) were honored to have represented them in this transaction.
Cheers to the true innovators. Cheers to those who build the real deal.
