Granola's customer story about Docker describes a specific order of events. More than fifty people had already signed up before Docker's IT team assessed the tool, and a rollout for 600 employees followed an evaluation of shared analysis, security and operational support. Two different things can spread through a company that way: a personal habit and a shared working system.
This note asks what turns the first into the second. We use "team habit" in a narrow sense: a colleague who did not attend a conversation repeatedly gets useful context from it. That is our definition, not Granola's activation metric. Sending a note, creating a folder and buying seats are different events, and they do not have to happen together.
Docker's account has limits. Rollout coverage is not the same as 600 people using it, and a vendor-published story shows what happened at one company, not how often it happens.
What Granola is, and what we looked at
Granola is an AI meeting notepad. A person writes what matters to them while the app transcribes the conversation and enhances their notes. In its Series A post of 23 October 2024, the company dates its public launch to May 2024 and describes its original emphasis as supporting the person's judgment, not replacing it.
We read Granola's public pages, documentation, announcements and three customer stories, checked on 28 October 2026. We have not used a team account, seen internal data or spoken to anyone at the company. We do not know how well its team adoption performs, and our research does not establish a team-adoption problem. Pricing and documentation change, so check the live pages before relying on a detail.
The author and the colleague need different things
The person who records a conversation gets value straight away. They pay attention without documenting every detail, and they can find what was said later. A colleague who missed the conversation needs something else: context they can find, understand, trust and use.
| Job | What the author gets | What a colleague also needs |
|---|---|---|
| Capture | Attention on the conversation, without documenting every detail | Agreement on which conversations can appropriately be captured and shared |
| Recall | Finding what was said in their own meetings | Relevant context made available to someone who was absent |
| Synthesise | Comparison across several conversations | Knowing which conversations are included, what is missing and what the evidence supports |
| Act | A prepared follow-up or agenda | Use of the context in a recurring shared task, with a named person accountable for interpreting it |
The table is our model. A note can be useful to its author without being relevant, understandable or appropriate for anyone else.
The author's reward arrives immediately. Preparing useful context for colleagues can take extra effort, and the colleague then needs a recurring reason to go and retrieve it. We think that gap is worth investigating. Public evidence does not show that it is Granola's largest growth bottleneck.
What Granola has already built
Granola has been building toward teams for some time. Granola 2.0, announced on 14 May 2025, introduced shared team folders and questions across meetings, and described Slack distribution and folder templates.
| Capability | Documented behaviour |
|---|---|
| Sharing notes | A note can be shared, added to folders and opened by recipients in a browser, according to the sharing documentation |
| Recurring meetings | Adding a recurring meeting to a folder can enable automatic filing of later instances, per the Spaces and Folders documentation |
| Existing tools | Folder rules can route newly added notes to integrations, per the same documentation |
| Reusable analysis | Recipes are saved prompts that can be shared. Sharing a recipe does not share the author's notes |
| Recipient access | A dedicated Shared with me view, added in a September 2025 update |
This list matters for what the note does not propose. Automatic filing of recurring meetings, a shared inbox and shareable prompts already exist, so a proposal to build them would duplicate the product. Access also differs by role and by surface. The documentation separates browser viewing from desktop collaboration and full-transcript access, and folder roles from editing rights on individual notes, so we make no claim that every recipient can see every transcript.
Three events that need not happen together
| Transition | Evidence of progress | A metric that can mislead |
|---|---|---|
| Personal habit | A person returns to capture and retrieve useful context | Downloads, or a single meeting recorded |
| Shared working habit | An authorised colleague repeatedly uses another person's meeting context in real work | Invites, link opens or automatic Slack posts alone |
| Paid organisational expansion | More paid workspace use, or policy requirements that justify a purchase | Treating each recipient as a billable seat |
The last row follows from how Granola describes sharing. Its documentation says sharing a note does not add the recipient to the workspace or count as a billable seat. Its billing guide says subscriptions apply to workspaces, with paid seats for active members, one plan per workspace and no mixed tiers within a workspace. On 28 October 2026 the pricing page listed Basic as free, Business at $14 per user per month and Enterprise from $35, with shared folders on every plan and 30 days of accessible history on Basic.
That structure has a commercial reading, which is ours. Letting colleagues consume shared context for free may help adoption before it produces more paid seats. That can be a rational choice, and it would become an economic problem only if the cost of supporting free recipients outweighed paid growth or other value. No source we read shows that. One contributor can also support a genuine team habit if several colleagues benefit, so asking every reader to become a note-taker would confuse seat growth with customer value. Whether a shared working habit makes paid expansion more likely is a separate question, and the test below measures the two apart.
What published rollouts describe
Three customer stories give a sequence to work from. All three are published by Granola, so they show reported successes and not how often an individual installation becomes a team habit.
- Docker: individual sign-ups came first, then IT assessment, then a company-wide rollout, as described above.
- Brex: existing employee use led IT to formalise adoption, through a Business trial and then Enterprise, with departmental templates and sharing. The story reports over 50% adoption among knowledge workers within six months.
- Zapier: a design leader found internal teams already using Granola. The story describes shared collections, with retrieval taking the place of some recap meetings, and reports 52% adoption within four months.
User pull and an organisational decision both appear in these accounts. The Zapier account is the closest to our definition, because the benefit comes from colleagues retrieving context and not from the author's own summary. We do not use the time-saved estimates in these stories, because we found no method behind them.
Sharing needs control
An update on Granola's site explains a distinction that matters here. The visible note is editable, but chat draws on the underlying transcript, so removing something from the note can leave it discoverable through chat. The company added deletion of transcript sections to address this.
So automatic sharing of everything is a poor starting point, and it would be equally wrong to say Granola lacks controls for this distinction. A team habit depends on contributors knowing what they are sharing and on recipients knowing what the material supports.
That gives two failure modes to test. They are not reported findings.
- Too little usable sharing: the author benefits, but a colleague still needs a separate recap, cannot find the relevant notes or cannot tell what a shared collection covers.
- Too much indiscriminate sharing: recipients receive a stream of irrelevant summaries, or contributors lose confidence in how their meeting context will circulate.
The decision is not whether to make sharing easier. It is whether the next sharing step creates useful, repeatable collaboration at an acceptable review and control cost.
Is this a Granola problem or a team problem?
Granola is not the only product that does this. Notion's documentation describes meeting transcription, summaries, human notes that inform the summary, and optional sharing with internal meeting participants. So the competitive question is whether Granola adds enough retrieval, cross-conversation analysis and workflow fit to justify a separate tool next to a team's existing stack.
The other alternatives are ordinary ones: a summary in Slack, decisions in a shared document, or asking the person who was there. Those can be enough when the information need is rare. A team with no recurring need for another person's conversations is unlikely to form this habit however good the setup is, so the recurring need comes first.
A test with one change
Segment. A small product team that already holds a weekly customer-feedback review, with an established Granola user who conducts relevant conversations and a colleague who uses that feedback to prioritise. Sharing must already be allowed by the organisation's policy.
Screen first. Check whether recipient-led setup already happens in the participating teams, for example because the colleague already brings a standing question to the author. Teams that work this way are left out, or the test would compare two versions of the same practice.
Control. A standardised folder setup based on Granola's shared meeting-library guide, which already recommends organising folders around questions and querying across conversations. We add generic example questions to the control. That is our implementation of the comparison, not an onboarding flow Granola prescribes. The permissions, review expectations and integrations are the same as in the treatment.
Treatment. The colleague supplies a real question from their next scheduled review, and the setup is organised around that question. An illustrative question is which onboarding difficulties appeared in the interviews available for that review. That is an example, not a report of a real request. The colleague states the question, not the facilitator, and it is written down before setup so that it cannot be chosen to fit the evidence. The guide already recommends organising folders around questions and querying across conversations, so the treatment adds one thing: the question comes from the recipient's own next decision, not from an example. It can start as a small assisted onboarding test, and it assumes no new Granola feature.
We would not bundle a dashboard, a new inbox, a referral scheme, company-wide sharing or a new summary engine. If the value cannot show up with existing capabilities and a real recipient, those additions would only hide the result.
Main outcome. Repeated, useful reuse by an authorised non-author across consecutive instances of the review, assessed after facilitation stops. Reuse means an attributable reference to evidence that answers a review question, prepares an agenda or clarifies a follow-up. Opening a page, or accepting an AI answer without checking it, does not count.
The trade-off to watch. Recipient gains have to be set against what contributors carry: the time they spend reviewing and correcting what they share, and any access confusion or unwanted disclosure. If access problems appear, we would tighten sharing, not widen it to improve the metric. Paid expansion is measured separately, because recipients of shared notes are not seats.
Randomise at team level where possible, since members of one review group influence one another. Start by observing a few teams to find where the setup fails, then size a larger test from baseline rates. The conditions held equal in both groups and the full measurement plan are in a methodology appendix.
Reading the result.
- Equal reuse in both groups: the test found no measurable additional benefit from recipient-led setup under its conditions. With small teams the uncertainty is large, so this does not prove it adds nothing.
- Reuse that stops once facilitation ends: continued use has not been shown. We would ask whether the need disappeared, source material was missing or the facilitation was carrying the behaviour.
- Clearly more reuse with the treatment: an encouraging result worth repeating with more teams, not proof of a durable habit.
- Too few completed review cycles to compare: the test is inconclusive, and we would say so.
If the real constraint is capture quality or missing source coverage, that should be fixed before onboarding is redesigned.
The case against changing anything
- Granola may already do this well. The customer stories make that possible. This note tests a generalisation of a successful mechanism, not a proven failure.
- A Slack or shared-document process may be enough. Granola has to earn its place beside it.
- Some notes should stay personal. Not sharing can be correct behaviour, not incomplete adoption.
- Recipients may gain without becoming users or buyers. Product value and paid expansion can arrive at different times.
- More reliance can amplify errors. Search and citations make checking easier but do not turn every transcript into an authoritative record.
What we do not know:
- How often an individual installation becomes a team habit, and what Granola counts as team activation.
- How often recipients benefit without paid membership, and how many invitations convert.
- What current in-app onboarding already does, and whether recipient-led setup is part of it.
- How much review and correction work contributors carry.
- Whether our proposal adds anything over the control.
Next step
Product Garage works through live product and go-to-market decisions with experienced operators. This note takes one of those decisions in public, while the thinking is still provisional.
Subscribe to Garage Dispatch to get the next working note.
FAQ
Does sharing a Granola note add the person to my paid plan? According to Granola's documentation, no. Sharing a note does not add the recipient to the workspace or count as a billable seat. The billing guide says subscriptions apply to workspaces, with paid seats for active members. Check the current billing page before you plan a rollout.
How much does Granola cost? On 28 October 2026 the pricing page listed Basic as free, Business at $14 per user per month and Enterprise from $35. Business is billed monthly, and Enterprise also offers annual billing. Basic keeps 30 days of accessible history, and shared folders are available on every plan. Plans change, so check the live page.
Can a colleague see my transcript if I share a note? It depends on their role and on whether they use the browser or the desktop app. The documentation separates browser viewing from desktop collaboration and full-transcript access, and folder roles from editing rights on individual notes, so we make no blanket claim. Check the sharing documentation for your setup.
Does Granola already have shared folders? Yes. Granola 2.0, announced on 14 May 2025, introduced shared team folders and questions across meetings. The documentation also describes filing of recurring meetings, routing to integrations, shareable recipes and a Shared with me view.
What does "team habit" mean in this note? A colleague who did not attend a conversation repeatedly gets useful context from it. That is our analytical definition, not a metric Granola publishes.