From the Archives asked
answered May 7
[I'm extending my AI agent product to tackle customer support scenarios]. Looks like a real wedge that could extend to other YC companies. Would love any advice you've got. [Editor: some additional company-specific context was cut]
Philip's answer
I know nothing about AI support agents at all, but here are some random thoughts from someone with very little knowledge in the domain:
- Naively, it seems like there is already a ton of energy and funding going into AI support agents and speeding up customer ticket resolutions.
- That said, the domain is almost definitely not winner-take-all. The trick is probably finding an angle. For instance, the standard super-cheap support agent likely flips through a flowchart of statements that can be easily automated. That sort of person doesn't need "context" from elsewhere/etc, because there are few decisions to be made. But Tier 2+ support begins to be manned by expensive engineers who do need context in order to make decisions. The good news about automating Tier 2+ is that it's the one where serious dollar savings can be realized.
- It's always better to start with a super-concrete use case or a niche product before generalizing. The whole 1000 True Fans line of thinking.
Above all, remember that you live in a time when moats are eroding. Anyone else can copy the code you've written very quickly -- and imagine how fast they could do that, say, six months from now when AI is even better at producing code. So you must rush to lock in other business moats: things like scale efficiencies, network efficiencies, or brand. The last is a great example: Calendly is a dead-simple product anyone can build over a weekend now. But they've locked in the brand, so competition no longer matters.
Have you read Seven Powers? https://www.amazon.com/7-Powers-Foundations-Business-Strategy/dp/0998116319. I wouldn't recommend spending your time on reading the whole book unless you're very fast -- instead, just ask AI to summarize it for you. Useful to think about when it comes to building a sustained business.
The best ideas are "contrarian and right" (Thiel). Your idea is currently not contrarian, but it certainly is right (i.e. there's a real need). Remember also that Warren Buffett often reminds people that you can make many customers happy, and be in a high revenue industry, but still make no profit -- like airlines. It's far better to be in a small business that's per-employee highly profitable than a large scale business that makes little money.