Recall.ai shipped an update this week. Core functionality stopped working. A paying customer tried to get help.
No support channel existed. No chat. No email to humans. No ticket system. Just silence.
This is not a Recall problem. This is an AI startup problem.
Most non-enterprise AI tools operate in four support tiers: nothing at all (maybe a Discord the founder checks weekly), an abandoned chatbot (configured once, never maintained), a contact form that auto-replies and never follows up, or actual support (rare, enterprise-focused only).
Pre-AI SaaS companies staff support teams. Squarespace provides real-time support for $20/month. AI tools? Even paying customers get silence.
The sales implications: AI startups are hiring engineers, not support operations. SaaS companies running 20+ AI agents report maybe four vendors with functional support. The rest offer nothing when tools break.
When support does exist, it is under-resourced. One auth platform took weeks to respond to a cancellation request, then provided instructions pointing to a UI element that did not exist.
The talent gap is real. AI companies prioritise product velocity over customer success infrastructure. No support managers. No operations hires. No escalation paths.
For sales teams evaluating AI tools: ask about support structure before buying. Who answers tickets? What is the SLA? Is there a human escalation path?
The answers reveal whether a vendor is building for enterprise scale or hoping inbound growth covers the support debt.
Pre-AI playbook: hire support as you scale. AI playbook: ship fast, deal with broken workflows later. The gap creates risk for any team relying on these tools in production.
Support operations roles remain unfilled because AI startups are not hiring for them. The skills shortage is not about finding talent. It is about companies choosing not to staff these functions at all.