Humans are still doing the work behind some AI agents
- ID
- 41514
- Status
- summarized
- Published
- 31 Oct 2026, 8:00 PM
- Fetched
- 31 Oct 2026, 8:54 PM
- Provider
- CNBC Technology
- Category
- technology
- Original URL
- https://www.cnbc.com/2026/10/31/ai-agents-humans-dazzle-duckbill-meta.html
- Source URL
- https://www.cnbc.com/id/19854910/device/rss/rss.html
Summary
- Score
- 6.5
- Created
- 31 Oct 2026, 8:55 PM
- Tags
- Audience
- developersvibe_codersai_ml_learnerssaas_founders
What happened
CNBC reports that AI agents built on frontier models from Meta and OpenAI still can't reliably complete real-world actions across the web, especially anything involving a phone call. Startups including Marissa Mayer's Dazzle AI, Duckbill and Wajo AI are therefore selling human workers behind the scenes — Mayer's example is using frontier models to generate a good list of home-renovation contractors, then handing off the actual booking and scheduling. The framing is that an agent negotiating your cable bill may be a person haggling with customer service.
Why it matters
If you're shipping an agent that touches real-world outcomes, budget for a human fallback step in the workflow rather than assuming the model closes the loop — that changes your unit economics (a per-task human cost) and your UX copy (you need to disclose when a person takes over). It also opens a practical build option: use the model for the suggestion/candidate-list stage, which the article says works, and route the brittle action stage to a human queue.
Discussion angle
Where in your agent's flow would you draw the handoff-to-human line, and what does that do to your cost per completed task versus a fully automated flow?