Two CFOs of mid-sized businesses raised an interesting point about what falls in their strike zones when it comes to AI:

“We don’t have the resources to compete with the giants. We try to develop AI stuff in-house that we feel like is proprietary…but we’ll kinda be like a fast follower for the big, complex stuff like revenue management and buy that off the shelf when someone like Marriott figures it out.”

- CFO, Hospitality Management Company

“We are focused on small wins. We have a contest right now: whoever can come up with a new use for Perplexity that drives the most business value by the end of September will get 2,500 bucks.”

- CFO, Home Furnishings Retailer

The first CFO stressed the importance of triaging AI opportunities. The second CFO was incentivizing their team to focus on the work that would drive immediate value. Put them together and we start calibrating the sweet spot for in-house builds.

There’s no shortage of AI advice out there, but most of it is geared towards the behemoths with 10,000+ employees with layers upon layers of technical experts. When a 50-person business tries to follow along, it’s not a fair fight.

We think size is an advantage. A 50-person company might be willing to pay someone $2,500 if they figure out how to get Perplexity to save one person 3-5 hours a week. A massive conglomerate has to chew bigger pieces.

So the big guys go after the big, gnarly, complex builds because they have the resources to throw at it. Revenue management in the hotel industry is a perfect example because it’s complicated, requires time to solve, and only pays off at scale. They may fail the first few attempts but be fine in the long run. Our clients don’t have that luxury.

The good news is that these big problems don’t remain unresolved for mid-sized businesses. Someone with a large balance sheet cracks the code and then the solution trickles downstream in the form of a software license or product that smaller players likely already pay for.

Through our client work and demos, we’ve seen mid-sized businesses struggle with projects that have one or more of these themes:

1. Commodity products: when the whole industry shares the pain, a vendor can sell a solution thousands of times so someone is likely already on it.

2. Future problems: when companies are proactively automating for a problem that they expect to grow into, they end up missing out on the immediate needs.

3. Good but not great: when the solution or software is good enough, move on. This is most true for third-party software because these vendors should always be improving.

The sweet spot works both ways. Bite off more than you can chew and you won’t see results. Go too small and the needle will barely move so you won’t see results either.

Here’s the immediate action for operators to take:

• Get your company or team’s AI wish list and for each item ask whether a software company could white label it to a thousand other businesses like yours. If yes, remove it. If no, decide if it’s a problem for a 10-person, 100-person, or 1,000+ person company.

Crossing out ideas is rarely fun, but what’s left on the list should be workflows that are specific to your business needs and size. That’s the Goldilocks segment.