That's where I want to go with it is I call that prompt list, which is I want the AI just to chime in when it determines that it has relevant information and can help the team, even if it hasn't been summoned yet. But we need some model improvement for that, frankly. I mean, we'll be able to get there by GPT five, but that's totally where this is headed.View on YouTube
GPT‑5, a “GPT‑5‑class” multimodal foundation model from OpenAI, was released on August 7, 2025 and is now widely available via ChatGPT and Microsoft Copilot, satisfying the timeline condition “by the time a GPT‑5‑class model is available.”(en.wikipedia.org)
By that point, at least some enterprise chat environments do support effectively “promptless” AI behavior very close to what Sacks described:
-
Doti Autopilot for Slack is an AI assistant specifically designed to sit in Slack channels and listen in the background. Its documentation states that Autopilot:
- “monitors messages in enabled channels,”
- “detects valid questions,”
- “replies automatically when it can confidently help,” and
- does so “without requiring the user to tag
@doti,” adding that “users might not even know Doti is there – and yet their questions get answered within seconds.”(help.doti.ai)
This matches the prediction’s idea of an AI that “just [chimes] in” with relevant, helpful information in a team chat without being explicitly summoned.
-
Slack’s own AI/agent framework describes AI agents in Slack as “autonomous, proactive applications” that can operate inside channels and DMs, analyze conversational context, and take actions or respond based on automated triggers rather than only direct prompts.(docs.slack.dev) This supports the claim that the enterprise chat stack is moving toward embedded, always‑on agents.
Glue itself, the product Sacks co‑founded, has shipped extensive agentic AI features—MCP integration so Glue AI can act across external tools, multi‑step tool use, and AI Rules/Suggestion Rules that let it propose thread names and recipients and run complex actions once a user asks for help.(docs.glue.ai) However, the publicly documented behavior still appears to be user‑initiated (via @‑mentions, AI sessions, or explicit actions) rather than a fully autonomous “chime‑in on any conversation” mode.
Taken literally as a prediction about Glue’s own feature set, the claim is not yet clearly fulfilled. But as normalized—“enterprise chat tools like Glue will be able to support effectively promptless AI behavior by the GPT‑5 era”—the ecosystem now includes Slack‑based tools (e.g., Doti Autopilot) that deliver exactly this: background agents in enterprise chat that autonomously interject when they detect they can help, by or before the GPT‑5 timeframe. On that broader reading, the prediction has effectively come true.