Did OpenAI Just Kill Glean?

If you found value in this newsletter, please consider supporting via subscribing or upgrading. It helps make independent AI journalism sustainable.

In ChatGPT, you can now do “Deep Research” across all your private documents in Google Drive, Dropbox, etc, with data connectors and MCPs. Previously, the agent could only do research with public web data.

This means you can spin up Deep Research to write entire QBR reports or RFPs using various Google Sheets and other internal docs. Best of all, these report writing jobs can run in the background.

Considering writing MBRs and QBRs is like 30% of a middle manager’s job, this is massive.

So clearly, ChatGPT is on track to become “one stop shop" to write research reports on private plus public data. Since research and writing are the top business use cases of AI apps, this will have enormous competitive implications.

Note, this feature was hinted for ChatGPT enterprise tier even back in February, but to see it happen for the consumer tiers is something else.

But will this kill enterprise RAG apps, like Glean? And what does this mean for all vertical AI SaaS startups which are basically agentic RAG apps at their core?

Join the discussion on LinkedIn

It all depends on these three questions:

• How fast can ChatGPT bundle features such that it commoditizes all other enterprise AI apps?

• How much can OpenAI use its models as a lever to make their product experience superior to the purpose-built apps?

• Will OpenAI ever release a slick workflow automation runner & manager?

A few months back, I observed that ChatGPT wants to be “Cursor” for all your apps: a root-level chat assistant that sits on top of all your apps. This allows ChatGPT to be the only app that has context across multiple apps, while other apps are mostly limited to knowing what’s going on inside their apps.

Recent releases by OpenAI in their productivity suite seem to confirm that trajectory. It’s bundling everything and anything. In particular, it’s working on a major redesign to put collaboration features at the center. Essentially, it’s going for Microsoft 365 and Google Workspace’s throat.

But as OpenAI goes after Microsoft and Google, it may also streamroll AI startups. OpenAI is aggressively bundling AI features (e.g. OpenAI just released a meeting recorder). This basically weeds out and crowds out the market of software that only has table stakes features.

OpenAI could put a final nail in the coffin to many startups if it introduces a workflow design and automation app for enterprise (which I haven’t yet heard any rumors about).

Workflow automation is practically the only area that AI startups seem to be heavily invested into, that ChatGPT lacks.

But fundamentally, it’s not a hard product to build, and I am pretty sure OpenAI will eventually roll one out on their own.

So to revisit the original question: did OpenAI just kill RAG startups? Not yet, but the bar will be much higher to win customers, and good luck if you are trying to sell a RAG app at $299 per month per seat when ChatGPT enterprise only costs $50 per month.

Join the discussion on LinkedIn