Skip to content
MakerGain
← Back to project

In the Years Fathom Made Meeting Notes Free, the Founder Says It Lost \$50 per User per Month

Richard White built Fathom to keep the thoughts from his meetings: the personal tier offers unlimited recording and transcription for $0, while the Team tier charges for sharing and search and the Business tier charges for CRM write-back. A HubSpot example on the official site requires proposed field changes to be shown first and moves a stage only based on the buyer’s own words; White said Fathom lost about $50 per user per month in its early years, and in September 2026 Fathom became part of Superhuman.

Richard White found that after meetings, his ideas and decisions rarely made it into notes or follow-ups. He built himself a tool that recorded conversations, generated notes, and kept the action items. That origin is what Fathom’s About page describes. (About[1])

The Fathom that followed is an AI meeting notes tool that generates a summary as soon as a call ends, syncing notes, insights, and action items to Slack, Salesforce, HubSpot, Notion, and Asana. The personal version is $0, free forever, with unlimited recording and transcription. (pricing page[2])

A stranger arriving here will ask: if it is unlimited and free, where does the money come from? In an interview in May 2026, White gave a number—in Fathom’s first few years, it lost about $50 per user per month. That is his account of a historical stage, not today’s unit cost, and it does not imply profit. (interview[3])

“Only Move the Stage If the Buyer Actually Says So”

Start with the work that happens after a meeting. After a sales call, updating the deal record in HubSpot is the first thing to get delayed. Fathom’s website has a HubSpot work example that breaks this down in detail: after the call ends, ChatGPT or Claude reads the conversation through Fathom, finds the matching deal in HubSpot, proposes updates to the deal stage, next steps, pain points, and tasks for both sides, and then drafts a follow-up email for review. (HubSpot example[4])

The rules are more interesting than the features. Before writing to HubSpot, every proposed field change must be shown, and pain points must be supported by the customer’s own words; before creating a new deal or contact, the system asks first; most important—only move the deal stage if the buyer actually said words that support the stage change.

That rule leaves the decision with people: “I think we can move this forward” and “the customer agrees to move forward” are two different voices in a meeting, and who gives the nod is decided by someone who understands the context. It does not sell “fully automated AI” as the draw; instead it adds a confirmation step to the process.

This is the vendor’s own example, with no independent customer results and no validation of field-write accuracy. The FAQ on the same page says any plan, including Free, can read calls in ChatGPT or Claude via MCP; Fathom’s existing HubSpot integration syncs call summaries over, and this example adds field updates and a follow-up draft on top. The Business tier on the pricing page has a separate native “CRM field sync” that automatically updates records after meetings—a different path from this MCP example. (pricing page[2])

Image: Fathom official showcase[5], accessed 2026-11-10; original interface preserved.

The Free Tier Makes Recording the Default; the Team Tier Carries the Paid Case

Fathom’s pricing turns “recording” into the default action: the personal Free tier is $0, free forever, with unlimited recording and transcription, instant AI summaries, clips and playlists, and cross-call search. Above the personal tier is the paid Premium tier, which adds advanced summaries, AI action items, and a custom meeting bot. (pricing page[2])

The Team tier has a clearer boundary. Team is priced per seat, with a minimum of 2 people, and lets a team put call records in shared folders, search and ask questions across multiple calls, and includes team security controls and HIPAA compliance under a standard BAA. The pricing page lists Team as $19 and $15, with monthly and annual options, but the visible text does not fully label which number corresponds to which. The Business tier adds CRM field auto-sync, Deal View, and coaching scorecards. For a sales team, the reason to pay is not “it can record” but “the records can be shared across the team, queried, and written back to the CRM.”

In the same interview, White said the team did things in order of risk: first onboarding and activation, then acquisition, then referrals, and only last monetization. He also recalled that the team made a pitch deck first, saying they were “selling what amounted to a roadmap.” (interview[3])

There are also two moves that do not look like standard SaaS: they set a principle of reserving 15% of every funding round for user investment; on distribution they opened several channels—eligible startups affiliated with VCs or accelerators get up to two years of free Team, nonprofits get 10 free seats, and those switching from products such as Gong get Business free for the term of their contract, along with data migration. (pricing page[2]) The channels are laid out here; public materials do not say which one actually brought in how much paid business.

There is one revenue number to be careful with. The interview page title says Fathom reached $30M ARR, while the body quotes White saying that in the first three years of monetization it went from zero to $30M, and a field on the same page labels it as Company revenue. This is an interview self-account and a page label, not audited financials; paid seat counts, ARPU, churn, and free-to-paid conversion are all undisclosed. (interview[3])

The Record Itself Is Hard to Stand Alone as an Endpoint

On September 14, 2026, White announced that Fathom had officially become part of Superhuman. He said that on that day users’ apps, logins, and product experience were unchanged, and that the Fathom connector in Superhuman Go was already live, able to query summaries, transcripts, action items, and recordings from past meetings. (announcement[6]) The announcement did not include the deal price, structure, or post-integration product roadmap.

This ownership somewhat echoes the product’s original form: meeting notes themselves are hard to stand alone as an endpoint; they are more like a piece of source material that others read when making judgments. In the announcement, White wrote that the next step was “less about searching and more about acting”—letting the right context find the people who need it, rather than people digging through call records. The homepage also says meeting data can be used in external assistants such as ChatGPT and Claude. (Fathom website[5])

There is another approach on the market. Granola explicitly does not use a meeting bot, relying on computer system audio and the microphone, and can handle scheduled, ad hoc, online and offline conversations; it processes audio for transcription but does not record or retain audio. Not using a bot means it is not automatically visible to other participants, so it provides automatic meeting chat messages and video watermarks to let participants know Granola is transcribing; reminding users to obtain consent is another mechanism—each time they start taking notes, Granola gives a reminder. On price, the Basic tier is $0, with meeting history retained for only 30 days; the Business tier is $14 per user per month, with unlimited notes and history and includes MCP and API. (Granola explanation[7], pricing[8])

These two are not mutually exclusive categories; Fathom itself also has bot-free capture, though the pricing page labels it as a beta feature on Mac. “Whether audio is retained” is a real choice; Fathom’s public materials do not lay out complete audio and transcript retention periods or deletion policies, so the comparison can only go this far. (pricing page[2])

White initially just wanted to keep the thoughts that would disappear from his meetings. Fathom turned recording into a free default action, and made team sharing and CRM write-back the reasons to pay; whether a human has to give the nod when writing back, the MCP example clearly says yes: every field change is shown first, then saved. Business’s native sync only says it automatically updates records; public materials do not say whether it also stops for confirmation. That calibration is closer to the real office than the “fully automated AI” slogan.

Continue reading

AI productivity
Klipy, the sales AI that waits for your approval

Klipy takes over follow-up after sales meetings. It captures customer conversations from email, meetings, and instant messaging, drafts follow-ups and CRM updates, and prepares for the next meeting. Follow-up messages go out only after the salesperson approves them. Founder Jung Hong Kim is a Korean serial entrepreneur who took on debt when COVID sank his retail venture. He met his cofounder while consulting and launched Klipy.ai in three months. It went live in November 2024 as an AI CRM that logged communications automatically. In June 2025, it shifted toward carrying out next steps for sales teams and tied billing to specific actions users approve.

AI productivity
The founder reported \$8,000 in 2024 revenue, 85% collected only once: NodeLand turns long reports into maps

Rio de Janeiro-based Gabriel Xará runs NodeLand as a side project. Canvas extracts material from long reports, video transcripts, or meeting notes and generates structured maps; people still decide which nodes to keep, whether relationships are correct, and who the map is for. Built in 2023 and launched in 2024, it acquired early users through platforms such as AppSumo. A January 2025 interview disclosed \$8,000 in 2024 revenue, 85% from lifetime licenses, with costs and profitability undisclosed. He had focused almost entirely on the product that year, with limited marketing.

AI productivity
After closing a product nobody wanted, they broke entrepreneurship into “the next step”

David Heikka and his brother Felix launched their first AI form product and found zero demand. They tried ads, landing-page changes, renaming, and added features, then closed it after six months. They went on to build Buildpad, breaking business creation into stages that indicate what to validate next. Its early and current billing mechanisms differ, and the product was renamed aicofounder before changing back to Buildpad.

AI productivity
The client owns the code and pays the AI bill: SpaceAI360 builds automation you can take with you

Founded by Shahdin Salman alone, SpaceAI360 builds portable AI workflows inside customers' own accounts. It handles parsing, OCR, field extraction, database checks, and ERP/CRM writes after a form, email, or file arrives, routing uncertain items to human review. Rather than generic AI consulting, he starts with the highest-friction bottlenecks, assigning deterministic steps to n8n and judgment steps to Claude.