
NAUT Ask AI is the chat box on naut.so that answers deal questions off NAUT's own directory, and NAUT's data shows it handled 354 questions across 252 live chats from 34 users between 4 October and 8 November 2026. Those answers pulled 139 company records covering 41 distinct companies, with 0 lookups failing to resolve to a NAUT page, and 3 of the 354 answers ended in an error.
How these figures were read: on 9 November 2026 NAUT read its own chat and message tables end to end. That is 253 chat rows, of which 252 are live and 1 is redacted, and 708 messages split evenly at 354 user questions and 354 assistant answers, an average of 2.8 messages per chat. Every number in this article is NAUT's own stored record of its product, so none of it was put to an outside source. The questions below are stored chat text, printed as the user typed them, and the company counts are lookups the assistant ran while answering, not page views.
What is NAUT Ask AI, and what does it answer?
NAUT Ask AI is a crypto research assistant that reads NAUT's directory of companies, investors and people, then writes the answer with the records it found. The 354 logged questions sort into 7 kinds of job, and 6 of the 7 are crypto deal research jobs, with 1 a definition lookup. Each example below is stored chat text from the log, and the last column is the record NAUT answered it from.
| Job | Example | Answered from |
|---|---|---|
| Decision-maker research | Who decides listings at |
Company pages and their team records |
| Outreach drafting | Draft a Telegram to the Head of BD at |
A person record plus the channels on file |
| Live funding screen | Which projects outside the top 100 raised money in the last 30 days? | Dated funding rounds |
| Investor-backing lookup | Which funds backed |
Round backer records |
| Trending and attention | What small-cap projects are getting the most attention on X this week? | Stored X intel snapshots |
| Company explainer | What is |
The company record and its sector tags |
| Service-provider list | List the market making firms on NAUT | Service provider records |
The 2 jobs that repeat most across the log are a decision-maker pull at a named company and a funding screen with a people step bolted on, as in a request for founders and BD people at projects that raised a Seed or Series A in the last 30 days. Both end at a page a reader can open: OKX, Kraken and Hyperliquid each carry their team on the page itself.
What did people ask NAUT Ask AI most?
OKX is the most asked company on NAUT Ask AI, with 20 of the 139 company lookups in the window, ahead of Wintermute at 14 and Kraken at 11. The table below ranks the 12 most looked-up records by the number of lookups the assistant ran against them between 4 October and 8 November 2026.
| Rank | Company | Lookups | How NAUT lists it |
|---|---|---|---|
| 1 | 20 | Exchange and wallet project | |
| 2 | 14 | Trading firm with a venture arm | |
| 3 | 11 | Exchange project | |
| 4 | 8 | Investor | |
| 5 | 7 | Stablecoin infrastructure project | |
| 5 | 7 | Layer-1 project | |
| 5 | 7 | Perp DEX and Layer-1 project | |
| 5 | 7 | Investor | |
| 9 | 5 | Investor | |
| 9 | 5 | Exchange project | |
| 11 | 4 | Gaming infrastructure project | |
| 11 | 4 | Modular data network project |
The chart holds the 4 records above 7 lookups; the other 8 in the table sit at 7, 5 or 4. Those 12 records took 99 of the 139 lookups. A further 3 companies sat at 3 lookups each, and the remaining 26 of the 41 distinct companies took 31 lookups between them. 4 of the 12 most asked records are investor pages, which is the signal a founder raising should read: the same chat box gets used to check a fund as to check a project.
How fast were the answers, and how often did they fail?
351 of the 354 answers finished with a done status and 3 ended in an error, and the average answer took 1.7 seconds end to end. 305 of the 354 answers called at least 1 NAUT data tool, at an average of 1.59 tool calls and 1.22 agent rounds per answer, with 4 rounds the most any single answer needed.
| Metric | Value | Note |
|---|---|---|
| Chats | 253 | 252 live, 1 redacted |
| Messages | 708 | 354 user questions, 354 assistant answers |
| Distinct users | 34 | 4 October to 8 November 2026 |
| Average messages per chat | 2.8 | Across all 253 chat rows |
| Quick mode answers | 311 | Against 43 deep-mode answers |
| Company lookups | 139 | 41 distinct companies, 0 unresolved |
| Answers that errored | 3 of 354 | 351 finished with a done status |
| Average tool calls per answer | 1.59 | 305 of 354 answers used at least 1 tool |
Quick mode carried 311 of the 354 answers and deep mode 43, so most of the work is a single pass over the directory with 1 or 2 record pulls behind it. The 354 questions landed on 26 days that saw any activity in the window, and the 2 busiest of those days carried 76 and 63 questions, which is what a working week of deal research looks like when a whole team is on it. Across the window the assistant read 4,117,033 tokens of context and wrote 483,353.
What does NAUT Ask AI tell a BD team?
- The jobs that worked are narrow and named: a company, a role, a round size, a date window.
- A question that names a company resolves, with 139 of 139 company lookups landing on a NAUT page.
- A 2-step question is normal here, at 2.8 messages per chat: screen the rounds, then ask who runs the teams that came back.
- An outreach draft is a question the log shows people asking, as with a Telegram note to the Head of BD at Kraken.
- Screen the roundsAsk for projects that raised in the last 30 days, the screen behind several of the 354 logged questions.
- Name the decision makerAsk who decides at those companies, the job behind the most asked record, OKX at 20 lookups.
- Check the backersAsk which funds backed the company, the lookup behind Coinbase Ventures at 8 and Paradigm at 7.
- Draft the noteAsk for a Telegram to the named person, then open their NAUT page for the channels on file.
What this tells a founder raising
An investor page is as likely a target here as a project page. 4 of the 12 most asked records are funds, and Coinbase Ventures at 8 lookups and Paradigm at 7 outrank every project below Kraken. The questions behind those lookups ask who co-invested with a fund recently and who decides at the projects a fund has already backed, which is the shape of a warm-intro search. Binance Labs at 5 lookups drew the same question about the projects it invested in this year.
Related reading on NAUT: Crypto data API: 3 calls, 5 credits a pull, 14,350 people and Crypto data for AI agents: 1 API, 2 calls, 15 credits.
Can you pull this data through the NAUT API?
The chat log has no endpoint of its own, and the records behind every answer in it do. The most asked record in this article comes back with its team from GET https://api.naut.so/v1/projects/okx, and the combined company-and-people lookup that sits behind a decision-maker question runs as GET https://api.naut.so/v1/search?q=OKX, each with the header Authorization: Bearer $NAUT_KEY at 5 credits a call on the plans at naut.so/pricing, documented at naut.so/docs/api. Verified Telegram, email and LinkedIn on a person record unlock on any paid NAUT plan, so start a 7-day trial if you want the channels as well as the names.
NAUT Ask AI FAQ
What can NAUT Ask AI do?
NAUT Ask AI answers crypto deal questions off NAUT's directory of companies, investors and people, and NAUT's data shows the 354 questions asked between 4 October and 8 November 2026 sorted into 7 jobs: decision-maker research, outreach drafting, funding screens, investor-backing lookups, attention checks, company explainers and service-provider lists.
How many questions has NAUT Ask AI answered?
NAUT Ask AI answered 354 questions across 252 live chats from 34 distinct users between 4 October and 8 November 2026, an average of 2.8 messages per chat, with 351 of the 354 answers finishing and 3 ending in an error.
Which companies do people ask NAUT Ask AI about most?
OKX at 20 lookups, Wintermute at 14 and Kraken at 11 are the 3 most asked records on NAUT Ask AI between 4 October and 8 November 2026, out of 139 company lookups across 41 distinct companies.
Is this a crypto AI chat built on live data?
Yes: each of the 354 answers was written against NAUT's own records at the time of asking, 305 of them calling at least 1 data tool, at an average of 1.59 tool calls per answer and 1.7 seconds end to end.
Where can I find the decision makers NAUT Ask AI names, and their contact channels?
Every person the assistant named across those 139 company lookups has a NAUT page that says which channels NAUT holds, and the same records come back through GET https://api.naut.so/v1/people?q=head%20of%20bd at 5 credits a call. Verified Telegram, email and LinkedIn unlock on any paid NAUT plan, so start a 7-day trial to open them.
Sources
Every figure here is NAUT's own stored record of its Ask AI chat and message tables, read on 9 November 2026 and covering 4 October to 8 November 2026. No figure in this article was put to an outside source, because none of it exists outside NAUT's database. 1 of the 253 chat rows is redacted, so counts stated over live chats use 252 while the 708 message rows cover all 253.
OKX
Wintermute
Kraken
Coinbase Ventures