Notion and Decagon: an AI support agent, 34% faster resolution and 2x deflection
Notion deployed Decagon's AI customer-experience agent to automate repetitive inquiries and route the rest to the right expert. Per Decagon's case study, Notion saw a 34% improvement in ticket resolution time, a 2x increase in deflection, and a 3.4% ask-for-human rate. Every figure is vendor-published; Notion's Global Head of Customer Experience is named and quoted.
| Metric | Before | After |
|---|---|---|
| Ticket resolution time | pre-AI baseline | up to 34% improvement |
| Deflection | 2x increase | |
| Ask-for-human rate | 3.4% |
The problem
A fast-growing SaaS company accumulates support volume faster than it can hire, and most of that volume is repetitive. Per Decagon’s case study, Notion wanted to automate the repetitive inquiries and route the rest intelligently, consolidating a stack of support tools and freeing its customer-experience team for higher-value work (source).
What was built
Notion deployed Decagon’s AI customer-experience agent, chosen after what its CX lead describes as a rigorous evaluation. Notion’s Global Head of Customer Experience, Emma Auscher, is quoted on record: “We conducted a rigorous” selection process (source). The agent automates repetitive inquiries, does intelligent routing to connect customers with the right expert faster, and consolidates previously redundant platforms.
The outcome
The headline metrics. Per Decagon’s case study, Notion saw a 34% improvement in ticket resolution time, a 2x increase in deflection, and a 3.4% ask-for-human rate (source). In the vendor’s words: “34% improvement in ticket resolution time, 2x increase in deflection, 3.4% ask for human rate”, and “With an average ask for human rate of 3.4%, what once required multiple steps and manual intervention is now” handled by the agent (source).
How this was verified
This case is verified on a single impeccable source. Every figure comes from Decagon’s own case study, marketing material, and no independent outlet reports these numbers. What clears TIN’s bar is that the client is named and its Global Head of Customer Experience, Emma Auscher, is quoted on the record: an identifiable executive who could refute the numbers if they were wrong, published by a vendor with real reputational stake. The standing caveat is that this rests on that identifiability and refutability, not on independent corroboration, which has not been located.
Sources
Cited in this case file. Tier 3 = vendor or first-party. Each figure was checked against the live source on 2026-07-08.
- Decagon case study, “Notion” (Tier 3, vendor-published with the client CX head named and quoted). https://decagon.ai/case-studies/notion · archived
Decagon AI customer-experience agentIntelligent routing; support-tool consolidation
- Status
- verified
- Method
- Single-sourced to Decagon's own case study, but it clears the impeccable bar: the vendor has real reputational stake as a venture-backed CX-AI company selling into other enterprise buyers, and the claim is anchored to a named, on-record client executive, Notion's Global Head of Customer Experience Emma Auscher, who is quoted directly and could refute the numbers if they were wrong. Figures (34% faster ticket resolution, 2x deflection, 3.4% ask-for-human rate) are sourced to that case study and attributed to Auscher's team. No independent second source has been located; this rests on the vendor case study's identifiability and refutability, not corroboration.
- Verified on
- 2026-07-22
- Provider
- Decagon, AI customer-experience agents
- Client
- Notion · SaaS / productivity software (US)
- Disclosure
- named
What results did Notion report with Decagon?
Per Decagon's case study, a 34% improvement in ticket resolution time, a 2x increase in deflection, and a 3.4% ask-for-human rate.
Why is this case file verified on a single source?
The single source is Decagon's own case study, but it names and directly quotes Notion's Global Head of Customer Experience, a person who could refute the figures if they were wrong. That identifiability and refutability, plus the vendor's reputational stake, clears TIN's bar for a single impeccable source. No independent second source has been located.