Is customer support automation an attractive market for AI companies to build in?

Carles Reina · Apr 11, 2026

Customer support automation is a boring, saturated pitch and the more interesting opportunity is building 'labs' partnerships with companies that generate new top-line revenue for them

An airline CEO told him he receives over 100 customer support pitches a week; most companies start with bottom-line optimization, whereas consulting-style partnerships that build revenue-generating agents are unproven but potentially scalable to hundreds of millions

25:49 Saturated pitch top line agent partnerships are the opening

Carles Reina · Apr 11, 2026

Customer support is not worth investing in as a venture bet, but it is a large and attractive space for already-established, fast-growing companies to enter

Most ElevenLabs customers start with customer support and the ROI for them is enormous

28:37 Not a venture bet but good for established scalers

David George · Dec 15, 2025

AI customer support is already better, faster and cheaper than the incumbent alternative with today's model quality and cost — no future model capability needs to be assumed

Reasoning capabilities and current token costs already deliver the functionality; the market pull is visible in enterprise briefing centers where portfolio company Decagon draws extremely high interest and usually converts to a deal

37:51 Already better cheaper today so the pull is real

Harry Stebbings · Feb 16, 2026

Customer support is hard to justify as an investment category because 14 players have been funded with over $100,000,000 in the last fifteen months, existing incumbents compete, and companies are building their own in-house

Too many funded entrants plus incumbents plus in-house builds by companies like Ariel/Ivan's and Jack's at Air

13:13 Funding glut plus in house builds make support uninvestable

Oren Zeev · Feb 2, 2026

He won't invest early in AI customer-support startups because there are thousands of them and he doesn't trust his intuition to pick the one winner

One of them will succeed but thousands won't, and at the very early stage he can't tell which

12:35 One winner exists but is unpickable at the earliest stage

Avishai Abrahami · Jul 13, 2026

Off-the-shelf AI customer support products do not work for a company at Wix's complexity; they repeatedly tested vendors and ended up relying on their own solution

Young AI agent startups prove something quickly and hire salespeople, but they are selling to companies with more and better engineers focused on that exact problem

31:53 Complex buyers build in house so support vendors cannot serve them