Taking a Chance on a Small Vendor
A young company can be worth the risk when the customer understands the dependency and has a workable way to continue.
A small vendor can be worth taking a chance on. The product may offer a substantial improvement, and working directly with the founder can help a difficult implementation get done.
I’d still want to understand how the customer keeps going if the company has a difficult month. That includes knowing which parts of the service depend on the founder being available.
I’d ask for financial information appropriate to the commitment and look at the obligations competing for the company’s resources. A recent fundraise would be relevant. The more useful follow-up is how long the available cash is expected to last and whether the team can support the customers it has already signed.
For an AI product, the providers behind it are part of the dependency. If the main model became unavailable, the existence of another model would be only a starting point. The vendor would need to show that its alternative works for this customer’s task, at a cost and review burden the customer can sustain.
Take a hypothetical buyer testing an exit from a 2,000-record workflow. The export and import take two hours, and 20 records arrive without the relationships needed to use them. The buyer needs to resume work within one hour with the complete record set.
| Exit test | Required | Observed |
|---|---|---|
| Time to resume | Within 1 hour | 2 hours |
| Usable records | 2,000 of 2,000 | 1,980 of 2,000 |
The transfer misses both requirements. Twenty records arrive without the relationships needed to use them.
I would not rely on that exit plan for the full workflow. The missing records need to be fixed and the transfer tested within the available window. A smaller deployment could be reasonable if the customer had a separate, workable way to keep that portion running.
The rehearsal matters because receiving a file and resuming work are different steps. The team may need the links between records, prior decisions, and a person who knows how to load the material into the fallback. Those details should be established while the customer still has choices.
Ordinary interactions can provide evidence of continuity as well. Can someone besides the founder resolve a technical question? Can another employee pick up an implementation without asking the customer to explain the history again? These are practical things a buyer can observe.
A small experiment using public data and replaceable outputs could justify much lighter diligence. If the company closed, the customer could repeat the work elsewhere at an acceptable cost. That judgment needs revisiting as soon as the product holds unique records or supports a recurring obligation.
The decision would come down to the value of the product and the cost of a disruption the customer understands. An exit promise nobody has tried leaves too much of that calculation unresolved. A workable fallback can make it possible to choose a younger company whose product is a better fit.
For bank buyers, the 2023 interagency third-party guidance addresses financial condition and transition planning in sections C.2.c and C.5.