Typewise

Customer service AI that resolves requests end-to-end.

5.2Score#63 of 999 agents#8 of 35 in Customer support agent

Why it ranks #63

Typewise, an AI customer support agent from YC S22, ranks #63 of 999 AI agents and #8 of 35 in Customer support agent, with a score of 5.2 of 10. Data as of 3 Oct 2026.

The score adds four parts, each 0 to 10, by weight. Every point comes from public evidence. How scoring works ·

Proof 5 × 30% 1.5
Scale 4 × 30% 1.2
Momentum 5 × 25% 1.3
Autonomy 8 × 15% 1.2
= 5.2

Proof
Do customers use it? 30% of the score
5
Claims named customers and a customer result.
- +3
  Named customers
  “We're YC S22 and work with companies like Superhuman and Unilever.”
  YC page
- +2
  A customer result
  “Across live accounts, that means 90% of requests resolved without human involvement.”
  YC page
Not found: revenue or growth (+4), a customer count (+2), work done at scale (+2)
Scale
Is there a real company behind it? 30% of the score
4
12 people.
- +4
  12 people on its YC page10 to 19 people
Not found: funding round in its news (+3)
Momentum
Is it shipping and growing now? 25% of the score
5
Shows a recent launch and hiring.
- +3
  Launched in the last 6 monthsTypewise Nova: The customer operator for fast-growing teamsLaunch YC post · 10 Sep 2026
- +2
  Hiring: 1 open roleFounding Sales Lead, DACHYC page
Not found: new traction claim in the last 6 months (+2), funding news in the last year (+2), a batch in the last year (it is YC Summer 2022) (+1)
Autonomy
How much of the job does it do on its own? 15% of the score
8
Autonomous worker: completes whole tasks; people handle exceptions.
- +8
  Level 3 of 4: Autonomous workerCompletes whole tasks; people handle exceptions.
What the model read
- Acts in other systems: 95% likely
- Finishes whole tasks: 79% likely
- Calls itself an agent: 73% likely
Level 3: Completes whole tasks; people handle exceptions.

About Typewise

Typewise resolves customer service requests end-to-end. Someone sends a request; it reads the message, looks up the real order or account in your own systems, takes the action, and replies. Anything it shouldn't touch goes to a human. It works across email, chat, WhatsApp and social, and is MCP-native.

Until now, that took a proper setup project, which is why it was mostly larger companies using it. Nova does the setup for you. You describe how you want things handled, in plain sentences:

- "Handle refunds under €50 yourself."

- "Anything about enterprise pricing goes to a human."

- "Turn bug reports into Linear issues and assign them to the Excellence team."

Nova connects the tools, configures the agents, tests them against your past conversations, and shows you what it got wrong before anything goes live. Nothing ships without your approval. First deployment takes 15 minutes. After that, it keeps going: it surfaces what customers keep complaining about, proposes fixes, and ships the ones you approve. We call it an AI Operator because it doesn't just answer customers; it keeps the entire operation working and improving. Across live accounts, that means 90% of requests resolved without human involvement.

Your team stays in control of what it may do on its own. Hybrid intelligence handles partial hand-offs, review-before-send, and context-preserving resumption.

We're YC S22 and work with companies like Superhuman and Unilever. Fortune 500 brands such as Unilever and DPD use Typewise to cut support effort by 50%+ while increasing CSAT and NPS. Run it natively in Typewise Ticketing or inside Zendesk, Salesforce or Freshdesk, with a CRM-embedded copilot for your human agents. EU-hosted, ISO 27001 certified.