# Acephalt

Aiolum story: https://aiolum.com/story/acephalt/ · added 2026-11-05 · updated 2026-11-06

**In short.** Acephalt (AI agents that run due diligence for venture and private-equity investors, Canada): Beta users at family offices and venture firms, November 2025. Evidence grade E: no revenue figure was checked. Team: Founded by Winnicent Zuo; a CFO writes on the company blog in 2026; headcount not published. Launched April 2025; still running.

AI agents that run due diligence for venture and private-equity investors · Canada · AI-native · team of 2 to 10 · still running · group: Finance, risk and compliance

- Outcome: Beta users at family offices and venture firms, November 2025 (evidence grade of the headline number: E; no revenue figure)
- Launch: April 2025; months from launch to the outcome: 7
- Team: Founded by Winnicent Zuo; a CFO writes on the company blog in 2026; headcount not published
- Outside funding: None found in the sources
- Main factor: Niche
- Meters, 1 to 4: evidence strength 1, copyable pattern 1, capital needed 2
- Tags: AI-native, Small team, Bootstrapped, Running, B2B, Investment software

Links:

- Visit acephalt.com: https://www.acephalt.com/

**The product.** A web workspace for venture capital (VC) and private-equity (PE) deal teams. The investor uploads a company's whole data room, and several AI agents read every file, check the financial model, score the deal against the firm's own criteria, and draft an investment-committee memo. Each figure in the output links back to the file it came from, and the work exports to Excel, Word or PDF.

**Why it worked.** Due diligence on one private company can mean hundreds of files, and small investment teams do most of the reading and spreadsheet checking by hand. The founder studied computer science and financial engineering and built a tool that splits this work across specialised agents, sold to lean VC firms and family offices.

Evidence grades: A = disclosed by a public company, a filing or a payment processor, or verified by one; B = two or more independent reports agree, or a journalist confirmed it; C = a founder claim that no one else verified; D = reported, not verified: one founder claim or a third-party estimate with no second source; E = no revenue figure found, so the headline is a non-money fact and the other facts carry their own grades; calc = computed on this page from graded facts.

## Summary

Acephalt is an AI platform for due diligence, the checking that an investor does on a company before investing: reading its contracts, financial statements and market claims. Investors receive these files in a data room, a shared folder that the company opens to possible investors. Acephalt runs several AI agents, programs that call a language model in steps and use tools without a person typing each prompt, in parallel over the data room, and it drafts an IC memo, the written case that a deal team sends to its investment committee.

Winnicent Zuo founded it; the company is Acephalt Inc. at 250 Wellington Street West in Toronto. UpNext, a founder-media site, wrote on 7 November 2025 that the platform "has already attracted design partners and beta users across North American family offices and venture firms" and that early adopters report "saving over 1,000 hours" of diligence time. The company's own blog describes one investor turning a 370-file data room into a full IC memo in under 24 hours.

No revenue figure, price list, customer name or funding round was found. This page is a grade E story: the business is real and live, and what it has achieved in money is not known.

## Pricing

Acephalt prices read on 5 November 2026 from the company's home page (grade C): no list price is published

| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Get Started or Talk to us | no list price |  | sign-up in the web app or a booked call; terms not published |

The home page offers "Get Started", which leads to the web app at app.acephalt.com, and "Talk to us", which books a call. No plan names, prices or limits appear on any page that was read, and the site has no pricing page.

## How much revenue does Acephalt make? The numbers, graded

A Toronto company selling AI diligence software to investors, with beta users in November 2025 and no published revenue

| Fact | Value | Grade | Note | Sources |
|---|---|---|---|---|
| Company | Acephalt Inc., 250 Wellington Street West, Toronto, Ontario, Canada | C | Terms of use and the security and privacy page. The terms are governed by the laws of Ontario and of Canada. | [7, 8] |
| Founder | Winnicent Zuo | C | UpNext profile, 7 November 2025: he studied mathematics and computer science and later earned a master's degree in financial engineering. | [1] |
| Earliest dated company page | terms of use last updated 3 April 2025 | C | No launch date is published. The Internet Archive's earliest copy of the home page found is from 19 January 2026. The launch month on this page is the date of the terms. | [7, 9] |
| Users, November 2025 | design partners and beta users at North American family offices and venture firms | C | UpNext profile, 7 November 2025. No firm is named and no count is given. | [1] |
| Time saved, November 2025 | "over 1,000 hours" of due diligence time | C | UpNext profile: "Early adopters report saving over 1,000 hours of due diligence time." Not a money figure, and not checked. | [1] |
| Case study | a 370-file data room to a full IC memo and model review in under 24 hours | C | Company blog. One investor, not named. Another company post dates the case study to 6 December 2025. | [4, 5] |
| Team | founder plus a CFO, Nicolas Frendo, who signs a company post; headcount not published | C | Company blog post "Introduction to Acephalt", written by the CFO. No source gives the number of people. | [3, 1] |
| Prices | not published | C | Home page read 5 November 2026: Get Started and Talk to us only. | [2] |
| Partner model for spreadsheets, 2026 | Tetra, a spreadsheet agent from DealGlass, embedded by API | C | Company benchmark post: "Acephalt is partnering with DealGlass to embed Tetra as the spreadsheet agent for VC and family office due diligence, via bespoke API." | [6] |
| Outside funding | none found | C | No funding round is named in any downloaded source. | [1, 2, 3] |
| Time from the terms date to the beta-user report | 7 months | calc | Calculated: April 2025 to November 2025. | [7, 1] |

How to read this table: a grade C value comes from the company's own pages or from a founder profile written from his statements. A "calc" value was computed on this page. No value is grade A or B, and no revenue figure exists.

## When did Acephalt launch, and how did it grow? Timeline

From terms of use dated April 2025 to a benchmark study of AI agents in 2026. Each entry carries its grade.

- **3 April 2025** (grade C [7]): Date of the terms of use of Acephalt Inc., Toronto. The earliest dated page of the company that was found.
- **2025** (grade C [1]): Zuo gives an UpNext PopUp interview at Elevate Festival 2025.
- **7 November 2025** (grade C [1]): UpNext publishes a founder profile: design partners and beta users at North American family offices and venture firms, and more than 1,000 hours of diligence time saved, by the users' own report.
- **December 2025** (grade C [4, 5]): The company publishes a case study: one investor turns a 370-file data room into an IC memo in under 24 hours.
- **12 March 2026** (grade C [8]): Security and privacy overview of Acephalt Inc. last updated.
- **2026** (grade C [6]): The company tests three AI agents on a cap-table model and announces a partnership with DealGlass to embed its Tetra spreadsheet agent.

## Why did Acephalt work?

### One agent per kind of work

Diligence mixes financial, legal and market analysis. The founder told UpNext that most AI tools treat these as one task, so Acephalt gives each part to a specialised agent and then joins the results into one report.

### A source for every figure

An investor has to defend each number in front of an investment committee. Acephalt links every claim and figure in the memo back to the file in the data room, which a plain chat answer does not do.

### The firm's own scorecard

A firm sets its investment criteria once, and the software scores each new deal against them and lists the open questions for the team. This makes the output fit the firm, not a general template.

## How it was built

These details come from the company's own pages and the page source of its site. They are grade C.

- A web app at app.acephalt.com; the marketing site is a Next.js site.
- Parallel agents read the data room, build or check the financial model and draft the memo, with exports to Excel, Word and PDF.
- Connectors bring in data rooms, call transcripts, inboxes and CRM records; the home page says 20 or more connectors are available on request.
- For spreadsheet work it embeds Tetra, an agent from DealGlass, through an API. Which language models run the other agents is not published.
- The terms say that AI outputs, including memos, risk flags and confidence scores, remain subject to verification by the user.

Sources: [2, 6, 7]

## Who tried this before Acephalt, and why did they fail?

Three companies in our database of failed startups sold data and analysis to investors and failed. The values below are from that database and are not independently verified.

The failed startups come from our database of dead startups, retold here in our own words; Its entries are AI-assisted summaries of public sources, so they carry no evidence grade.

### Mattermark (2012 to 2019; funding US$17.7M; cause: Competition)

- Promise: Private-company data and growth signals for venture investors.
- Why it died: It could not set itself apart from CB Insights and PitchBook, and the cost of collecting data grew faster than revenue.
- Record: 

### Monitor110 (2005 to 2008; funding US$17M; cause: Unit Economics)

- Promise: Early warning signals for hedge funds from blogs, news and social media.
- Why it died: Processing the data cost more than clients paid, and the market was not ready for it.
- Record: 

### Pit.AI (2017 to 2020; funding US$5M; cause: Competition)

- Promise: AI-driven insights to manage investment portfolios.
- Why it died: It could not show better results than traditional asset managers, and regulation asked for transparency that its models did not give.
- Record: 

## What changed

The three predecessors sold data or signals and had to pay to collect and process that data themselves. Acephalt does not sell data: the investor brings the data room, and the software does the reading and checking that analysts did by hand.

Language models that can read hundreds of documents and write a structured memo became available only after 2022, so one founder can now build what needed a data team before. The open question is the same one that ended Mattermark: whether a small company can stay different from large data vendors and from general AI tools that add the same features.

## Can one person copy Acephalt now?

**Verdict: Hard for one person; the pattern is clear.**

Selling software to investment firms needs knowledge of how diligence works, trust with careful buyers, and security reviews. The pattern can be copied in other fields where a professional reads a pile of documents and must cite each claim.

1. Pick a document-heavy decision that a small team makes again and again.
2. Split the reading into specialised agents, one per kind of analysis.
3. Link every output figure to its source file, so the buyer can defend it.
4. Start with design partners who bring real files, before setting prices.

**What is still open.** Audit, lending, insurance underwriting and legal review, where small teams read data rooms of their own. The founder told UpNext he is looking at internal auditing.

## What could go wrong for Acephalt?

- **What is not verified.** No revenue figure was found. The only traction statements are the UpNext profile of 7 November 2025 (beta users and design partners, no names or count, and more than 1,000 hours saved by the users' own report) and the company's own unnamed case study. No price list, payment-processor profile, funding announcement or company filing was found. The launch month on this page is the date of the terms of use, not a published launch date. A reader could look for a named customer, a funding announcement or a founder interview with figures.
- **Competition and model providers.** Henry AI, already on this site, also sells AI-built underwriting work to finance professionals, and the company itself compares its product with general tools such as Claude and ChatGPT, which can read the same files. The company's 2026 benchmark tested Opus 4.7 and GPT-5.5 against its partner DealGlass's Tetra agent, which it embeds by API; which model providers run the rest of the product is not published, so its dependence on them cannot be judged.
- **Sample data and a produced profile.** The home page shows a demonstration dashboard whose rows name well-known companies such as Databricks, Figma and Notion with check sizes and partner initials; the page does not say that these are customers or real deals. The UpNext profile is a "Flashpoint POV" feature, which the page says is produced by Flashpoint Global, a thought-leadership studio that works with founders.
- **People mapping.** The home page offers to "Map the executives, decision makers, shareholders, and stakeholders behind every company" and to "Enrich every profile". It is presented as research on companies under review; the pages read do not say which data sources it uses for people.

## Sources

1. UpNext, 7 November 2025: The Future of Due Diligence (Flashpoint POV founder profile of Winnicent Zuo) — https://upnext.world/future-of-due-diligence-ai-investing/ (grade C). Read in full, from the copy downloaded by the pipeline on 6 November 2026. A profile written from the founder's statements. Source for the founder, the beta users and the 1,000 hours.
2. Acephalt home page, read 5 November 2026 — https://www.acephalt.com/ (grade C). Read in full. Source for the product, the connectors, the demonstration dashboard and the absence of prices.
3. Acephalt blog: Introduction to Acephalt, written by Nicolas Frendo, CFO — https://www.acephalt.com/blogs/introduction-to-acephalt (grade C). Read in full. Source for the CFO and the company's own description.
4. Acephalt blog: A Real Deal: 370 Files To A Full IC Memo In Under 24 Hours — https://www.acephalt.com/blogs/a-real-deal-370-files-to-ic-memo (grade C). Read in full. The company's case study of one unnamed investor.
5. Acephalt blog: Why Acephalt Outperforms Claude and ChatGPT in Data Room Due Diligence — https://www.acephalt.com/blogs/why-acephalt-outperforms-claude-and-chatgpt (grade C). Read in full. Source for the comparison with general AI tools and the date of the case study.
6. Acephalt x DealGlass: Benchmarking Agents for Cap Table Modelling — https://www.acephalt.com/blogs/benchmarking-agents-for-cap-table-modeling (grade C). Read the sections on the method and the partnership. Source for the models tested and the Tetra integration.
7. Acephalt terms of use, last updated 3 April 2025 — https://www.acephalt.com/terms-of-use (grade C). Read in full. Source for the company name, address, governing law and the earliest company date.
8. Acephalt security and privacy overview, last updated 12 March 2026 — https://www.acephalt.com/privacy-policy (grade C). Read in full. Confirms Acephalt Inc. and its Toronto address.
9. Internet Archive Wayback Machine availability record for acephalt.com — http://archive.org/wayback/available?url=acephalt.com&timestamp=20230101 (grade C). Read in full. The closest copy to 1 January 2023 is from 19 January 2026, so no earlier copy was found.
10. Database of failed startups, local export — /method/#dead (grade C). Local export. AI-written summaries, not independently verified. Source for the failed predecessors.

Contact: aiolum@livyden.com. Run Acephalt? Send corrections or a verified figure to aiolum@livyden.com (subject: Correction: Acephalt).