Wix AI customer support: built own, doesn't work
Wix tried multiple off-the-shelf AI customer support tools and built its own, but admits none work reliably yet.
Snapshot · The Twenty Minute VC (20VC): Venture Capital | Startup Funding | The Pitch
20VC: Wix's Founder on What Wall St Gets Wrong About AI and Wix | Will Base44 Win the Vibe Coding Wars | The Truth About the Economics of Vibe-Coding | The Buyback Disaster: Lessons Learned with Avishai Abrahami
Explore episode
Jul 13, 2026
Where this was said
Why Every AI Customer Support Startup Keeps Failing (According to Wix)
At 32:00 · chapter starts 31:00
This chapter is the philosophical heart of the episode. Avishai resists both the techno-utopian and the dismissive framing of AI. His position: AI is genuinely transformational for specific, well-scoped tasks — having a natural language conversation, managing a schedule, reaching new customers for a gym owner. These are real capabilities that didn't exist five years ago and will reshape small business operations. But the cultural trust we've placed in AI accuracy is dangerously ahead of actual performance. [1] — Avishai Abrahami "AI is incredible at reframing data but terrible at reasoning from scratch. Avishai tested Claude on a safety protocol — six required gates …" 35:00 He describes using Claude to generate a safety protocol: it produced six mandatory test gates. Under interrogation, five of them collapsed — 'great question, you really caught me, I might have overemphasized that.' This is the 'but' he keeps returning to. For SMBs, he believes meaningful AI-powered productivity gains for things like scheduling, customer outreach, and order management are coming within a year. For complex reasoning, simulation, or research, the limitations remain profound.
Wix total headcount ~3,500
Wix employs approximately 3,500 people, with customer support being the single largest department.
Wix operates in 192 countries
Wix provides services in 192 countries, one key reason a small AI-only support team won't work.
Why AI Customer Support Startups Keep Failing
Wix tested every AI customer support startup on the market before building its own — which also doesn't work reliably. The problem: young companies prove a quick demo, hire salespeople, then sell into enterprises with better engineers and more complex support needs.
AI Is Amazing, But the But Is Big
AI is incredible at reframing data but terrible at reasoning from scratch. Avishai tested Claude on a safety protocol — six required gates became one after interrogation. We've built a massive cultural trust in AI accuracy that isn't yet warranted.
Similar snapshots
20 downloads/day at launch
PropGPT averaged 20 downloads per day right after launching on the App Store through influencer marketing.
4 months of rebuilding with no marketing
Eyal and Yali shut down all marketing and spent 4 months completely rebuilding PropGPT from scratch.
Over 40,000 downloads total
PropGPT has accumulated over 40,000 total downloads since launch.
LLM costs only $20/month
PropGPT's large language model (AI) operating costs are just $20 per month, and the cost is continually falling.
Ad monetization for games
Ad-based monetization works well for game apps where users spend extended time in-session, as seen with Grid and Wordle.
Ads don't work for tool apps
Tool-focused apps like PuffCount are poor candidates for ad monetization because users don't stay in-session long enough.
Hard paywall is unskippable
A hard paywall is a screen that blocks all app features unless the user pays or starts a free trial — it cannot be dismissed.
Two core monetization models
Mobile apps are primarily monetized through either ads (best for games) or in-app purchases/subscriptions (best for tools).
YouTube beats every other platform
According to the episode, YouTube outperforms every other social platform for building trust and driving SaaS conversions.
