BotDojo: Revenue, sale and how it was built
BotDojo (Platform to build, trace and evaluate AI agents for business work, United States): Acquired by Mitratech, September 2026. Evidence grade E: no revenue figure was checked. Team: Founder and CEO Paul Henry with CTO Edwin Woo; Artificial Lawyer understood the team to be around 10 people or fewer at the sale in September 2026 (headcount not published). Launched May 2024; sold 28 months after launch.Added 8 November 2026
- Outcome
- Acquired by Mitratech, September 2026
- Evidence
- E no revenue figure was checked
- Launch
- May 2024
- Status
- Sold, 28 months after launch
- Team
- Founder and CEO Paul Henry with CTO Edwin Woo; Artificial Lawyer understood the team to be around 10 people or fewer at the sale in September 2026 (headcount not published)
- Outside funding
- No funding round found in the sources (not verified)
- Main channel
- Not stated
What is not verified
No revenue figure and no sale price were found. We looked for a payment-processor profile, a founder interview with a dated figure, a funding announcement and a buyer filing; Mitratech's release and Dealroom's report give no price, and Artificial Lawyer only guessed that the deal was moderate in scale. The team size of around 10 people or fewer is Artificial Lawyer's understanding, not a published headcount, and no source names a funding round or rules one out. The usage counts are BotDojo's own, as reported on the AI Engineer profile. A reader could look for a statement by the founders or a later Mitratech disclosure. This story is grade E: the business is real and its product, price and customers are described from public pages, but no revenue figure was checked. Aiolum shows it as an idea worth studying, not as proof that it makes money.
BotDojo was a platform for AI agents, programs that use a large language model to read a request, decide on steps and act inside a company's own systems, such as a help desk or a CRM. It combined a visual agent builder with tracing, a record of every step an agent took, and evaluations, batch tests that run an agent over many saved questions and score the answers.
Paul Henry, formerly chief technology officer of the shipping-software company Auctane, founded it in 2023, with Edwin Woo as chief technology officer. The product was online by May 2024, and BotDojo's pricing page in April 2025 listed a free plan, a Starter plan at US$89 a month and a Pro plan at US$499 a month. On 28 September 2026 the legal-software company Mitratech Legal announced that it had bought BotDojo to build agents into its products for corporate legal teams and law firms.
This is a grade E story: no source gives BotDojo's revenue or the sale price, which Dealroom reports was not disclosed. Artificial Lawyer, a legal-technology news site, understood the team to be around 10 people or fewer at the sale. The page shows what a very small agent company built and who bought it, not how much money it made.

How much did BotDojo make, and what was it sold for?
Each row shows the evidence grade and the source numbers from the list at the end of this page. The grades are explained in How Aiolum verifies.
| Fact | Value | Grade | How we know | Sources |
|---|---|---|---|---|
| Status, 28 September 2026 | acquired by Mitratech Legal | A | The buyer's press release, datelined Austin, Texas: "today announced the acquisition of AI-native startup BotDojo to accelerate the evolution of Mitratech's ARIES AI". The BotDojo home page now says "Now a Part of Mitratech ARIES". | [1, 6] |
| Sale price | not disclosed | C | Dealroom: "Terms were not disclosed." Artificial Lawyer guessed the deal "would have been rather 'moderate' in scale". | [3, 2] |
| Founded | 2023 | C | AI Engineer organization profile, checked by its editors on 27 August 2026. The domain showed a parked page in November 2021, so no earlier company used it in a source we read. | [4, 10] |
| Product online | by 6 May 2024 | C | Earliest archived product page: "BotDojo - AI enablement platform", "From idea to production, BotDojo provides tools for shipping AI Applications using Large Language Models." Artificial Lawyer: "formally launched back in 2024". The terms of service carry an effective date of 3 January 2024. | [10, 2, 6] |
| Founders | Paul Henry (founder and CEO); Edwin Woo (CTO) | C | AI Engineer profiles. Paul Henry joined ShipStation when it had about ten employees and later was chief technology officer of Auctane, before he founded BotDojo. | [4, 5] |
| Team at the sale | around 10 people or fewer | B | Artificial Lawyer: "is understood to be around 10 people or fewer". No headcount was published by BotDojo or Mitratech. | [2] |
| Prices, April 2025 | Free; Starter US$89 a month; Pro US$499 a month; Enterprise custom | C | Archived pricing page of 21 April 2025, with usage charges above the plan limits. | [7] |
| Prices, April 2026 | from US$499 a month, usage-based | C | Archived pricing page of 19 April 2026: "Plans start at $499/month with usage-based pricing, not per seat." | [8] |
| Usage, August 2026 | more than 2 million interactions and more than 1 million call minutes | C | Company figures as reported by the AI Engineer profile. The archived voice page of April 2026 also showed "1,000,000 +" AI voice minutes. | [4, 9] |
| Funding | no round found | C | None of the sources names an investor or a round. This is an absence in the sources, not a statement by the company. | [1, 2, 4] |
| Buyer's customer base | more than 12,000 legal team relationships | A | Mitratech's release. Artificial Lawyer wrote "more than 8,000 legal teams" for the same company. | [1, 2] |
| Time from product online to the sale | 28 months | calc | Calculated: May 2024 to September 2026. | [10, 1] |
How to read this table: grade A is the buyer's own announcement; grade B is a journalist's report; grade C is BotDojo's own pages, a conference organizer's profile or a data provider's summary. A "calc" value was computed on this page. No revenue figure exists in any source we read.
| Plan | Per month, billed yearly (US$) | Per month, billed monthly (US$) | Limit |
|---|---|---|---|
| Free, April 2025 | 0 | 1 developer account, 1,000 runs | |
| Starter, April 2025 | 89 | 10 developer accounts, 5,000 runs | |
| Pro, April 2025 | 499 | unlimited developer accounts, 15,000 runs | |
| Enterprise, April 2025 | custom | custom limits | |
| Platform, April 2026 | from 499 | usage-based, not per seat | |
| Enterprise, April 2026 | custom | partner-led rollout and onboarding |
BotDojo's April 2025 page listed usage above the plan limits on Starter and Pro at US$0.02 per additional flow request, US$0.80 per additional 100,000 stored dimensions, US$0.005 per additional debug trace and US$0.05 per minute of web-scraping compute. No yearly prices were shown. The April 2026 page replaced the plans with one usage-based price "not per seat" and an Enterprise offer with hands-on onboarding.
When did BotDojo launch, and how did it grow?
From a parked domain to a purchase by a legal-software company, in about two and a half years. Each entry carries its grade.
- November 2021CThe domain botdojo.com shows a parked page with no product. [10]
- 2023CPaul Henry, a former chief technology officer of Auctane, founds BotDojo, with Edwin Woo as chief technology officer. [4, 5]
- 6 May 2024CThe earliest archived product page calls BotDojo an AI enablement platform for shipping applications built on large language models. Henry presents "BotDojo Launch" at the AI Engineer World's Fair in 2024. [10, 5]
- 21 April 2025CThe pricing page lists Free, Starter, Pro and Enterprise plans for building, evaluating and deploying AI. [7]
- April 2026CBotDojo sells specialized agents with usage-based pricing and voice agents for inbound and outbound calls. [8, 9]
- 28 September 2026AMitratech Legal announces the acquisition. The price is not disclosed. [1, 2, 3]
Why did BotDojo work?
Four causes, in the order they mattered. Each one is a reusable lesson.
Test agents with real questions
Henry's method, as the AI Engineer profile describes it, was to run historical customer questions against prompts, retrieval methods and models before release. Traces of each step show whether a failure comes from missing context, an unreliable tool or the model itself.
Usage-based plans, not seats
BotDojo's April 2025 pricing page sold plans by runs, stored data and traces, with charges for usage above the limits. Its April 2026 page kept usage-based pricing and stated that it was not per seat, which fits software that does work instead of software that people log in to.
Agents inside the systems a company already uses
The agents connected to CRM, telephony, tickets, documents, Slack and approvals. BotDojo also built on the <dfn>Model Context Protocol</dfn> (MCP), an open standard by which an AI assistant reads data from, and acts in, another program. Mitratech named two-way MCP connections as a main reason for the purchase.
Hands-on onboarding
The 2026 offer included onboarding, workflow design and optional employee AI training next to the software. Large customers got help to go live, which made the platform easier to sell to companies without their own AI engineers.
How it was built
These details come from BotDojo's own pages and the AI Engineer conference profiles. They are grade C. [4, 5, 6, 9]
- A visual agent builder with tracing, evaluations and human feedback tools. Agents run on the BotDojo platform, and developers connect them to their own applications through SDKs.
- An Interactive Agent SDK that uses the Model Context Protocol, so a website agent can see the page, highlight controls and fill in form fields. The company also publishes mcp-app-view, an SDK for building and embedding interactive MCP apps.
- Integrations listed on the home page include OpenAI, Anthropic, xAI, Hugging Face and Azure for models, and Salesforce, Zendesk, Slack, HubSpot and Microsoft Teams for business systems.
- Voice agents that run on the customer's existing phone numbers and routing; the voice page cites Deepgram's public Flux materials for its latency figures.
- A marketing agency, Kalungi, published a case study titled "How Kalungi Helped Botdojo Build a Scalable GTM Motion", listed on the AI Engineer profile.
Tools named in the sources
Each one opens its page in the Tools tab, with every story that uses it.
Who tried this before BotDojo, and why did they fail?
Three companies in our database of failed startups sold chat or support bots to businesses before large language models were available. The values below are from that database and are not independently verified.
Niles
The promiseA Slack bot that answered staff questions from the company's own knowledge.
Why it diedSlack added its own AI features, and faster and cheaper competitors followed.
Answerly
The promiseFAQ chatbots that answer support tickets so that people handle only the hard ones.
Why it diedLanguage understanding was too weak at the time, and Zendesk and Intercom already owned the support desk.
Vanti
The promiseConversational AI that let brands sell to shoppers through chat.
Why it diedNo product-market fit in a crowded market; the cost of winning each customer and the integration work were too high.
The failed startups come from our database of 1,749 dead startups, retold here in our own words. Its entries are AI-assisted summaries of public sources, so these cards carry no evidence grade.
What changed
The three predecessors sold bots that matched questions to fixed answers or scripted flows. Their bots could not handle questions outside the script, and the platforms they sat on (Slack, Zendesk, Intercom) could copy the simple parts. BotDojo started after large language models made agents able to read free text and call tools, so the hard part moved from understanding the question to proving that the agent does the right thing every time.
BotDojo sold that proof: traces, batch evaluations and human review. It also did not try to own the customer's data. It connected to the systems where the data already lived, and the company that bought it is one of those systems, a legal-software vendor with matters, invoices and documents from many years.
Can one person copy BotDojo now?
The platform, partly. The exit path, yes.
A general agent platform now competes with the model makers and with many funded companies. A small team can still build the testing and tracing layer for agents in one industry, and sell it, or sell itself, to the software vendor that already holds that industry's records.
Where it is still open: Industries where a few old software vendors hold the records and now need agents that can act on them: legal departments, insurance claims, property management and accounting.
- 1Pick one kind of work where mistakes are costly, such as invoice review or support tickets.
- 2Collect real past questions from the first customers and turn them into an evaluation set before release.
- 3Connect to the systems the customer already uses, through MCP and their existing APIs, instead of asking them to move data.
- 4Price by usage, and offer onboarding to customers who have no AI engineers.
What could go wrong for BotDojo?
The risks are the mirror image of the causes above.
What is not verified
No revenue figure and no sale price were found. We looked for a payment-processor profile, a founder interview with a dated figure, a funding announcement and a buyer filing; Mitratech's release and Dealroom's report give no price, and Artificial Lawyer only guessed that the deal was moderate in scale. The team size of around 10 people or fewer is Artificial Lawyer's understanding, not a published headcount, and no source names a funding round or rules one out. The usage counts are BotDojo's own, as reported on the AI Engineer profile. A reader could look for a statement by the founders or a later Mitratech disclosure.
Competition
ThinkHive, another story on this site, also traced and evaluated AI agents in production and was bought by Rootly in July 2026, so small agent-testing companies have been sold to larger software vendors more than once. In the release, Mitratech's chief executive described the market as "crowded with AI wrappers".
Dependence on model providers
The home page lists Anthropic, OpenAI, xAI, Hugging Face and Azure among its integrations, and Mitratech calls the combined product model-agnostic. The agents still need a model provider for every run, and the voice page bases its latency figures on Deepgram's Flux materials. A price or policy change by these providers changes the cost of each task.
Phone calls and service work
The April 2026 voice page offered "batch outbound campaigns" and calls for "reminders, collections, follow-up". We did not read a page that says whether these calls tell the person that an AI is speaking. The 2026 pricing page also included hands-on onboarding and a "Partner-led rollout", so part of what customers bought was service work next to the software.
Sources
The grade next to each source is the best grade that the source can support. "Read in full" means that the text of the page was downloaded and the numbers were checked in it.
- A Mitratech press release, 28 September 2026: Mitratech Acquires AI-Native Startup BotDojoRead in full. The buyer's own announcement. Source for the sale, its date, the buyer's customer count and the reasons for the purchase.
- B Artificial Lawyer, 28 September 2026: Mitratech Buys BotDojo For Agent OrchestrationRead in full. A legal-technology news site. Source for the 2024 launch, the team size of around 10 people or fewer and the guess that the deal was moderate in scale.
- C Dealroom News, 28 September 2026: Mitratech acquires AI startup BotDojo to build "System of Action" for legal teamsRead in full, from the copy the pipeline downloaded on 7 November 2026. A data provider's summary. Source for "Terms were not disclosed".
- C AI Engineer organization profile: BotDojo, Enterprise AI agents and platform (company sources checked 27 August 2026)Read in full. A conference organizer's profile. Source for the founding year, the founder and CTO, the product description and the usage counts.
- C AI Engineer speaker profile: Paul HenryRead in full. Source for the founder's background, his 2024 launch talk and his testing method.
- C BotDojo home page and terms of service, read 8 November 2026Read in full. Source for "Now a Part of Mitratech ARIES", the integrations list and the terms' effective date of 3 January 2024.
- C BotDojo pricing page, Internet Archive copy of 21 April 2025Read in full. Source for the Free, Starter, Pro and Enterprise plans and the usage charges.
- C BotDojo pricing page, Internet Archive copy of 19 April 2026Read in full. Source for the usage-based price from US$499 a month and the onboarding offer.
- C BotDojo voice agents page, Internet Archive copy of 19 April 2026Read in full. Source for the voice agents, outbound campaigns and the voice-minutes counter.
- C botdojo.com home page, Internet Archive copies of 28 November 2021 and 6 May 2024Read both copies. The 2021 copy is a parked-domain page; the 2024 copy is the earliest product page found.
- C BotDojo blog, 9 February 2025: Hard Lessons Learned from Deploying AI Agents, by Paul HenryRead in full (a short summary page). Confirms the founder writing for the company in 2025.
- C Database of failed startups, local exportLocal export. AI-written summaries, not independently verified. Source for the failed predecessors.
Run this company? Send corrections or a verified figure to aiolum@livyden.com