AI at Work
Granola Review for Work Notes: Personal Capture, Shared Folders, and History Access
A fictional one-on-one becomes a reviewed project handoff, then a three-person team compares free shared folders with paid history and separate workspace costs.
Granola is worth evaluating when your own rough meeting notes need structure and older conversations need to remain searchable. Start with Basic if its 30-day in-app history meets your needs: shared folders already appear in the free plan. Business becomes a stronger candidate when longer history or paid features matter, but a company workspace bills active members together. Paying for yourself does not automatically upgrade colleagues or authorize sharing a sensitive conversation.
This LandOffer.ai review checks Granola’s current documentation and pricing on October 27, 2026. The worked notes, people, and decisions below are fictional editorial examples, not captured meetings or Granola-generated output. No account, recording, permission, subscription, or retention setting was changed. That boundary matters because a convincing written note cannot demonstrate transcription accuracy or prove who can access a real workspace.

What Granola captures, and what you still need to review
Granola combines your rough notes with a meeting transcript and calendar information to create enhanced notes. Its documentation describes a way to inspect source context behind a point. That makes verification part of the workflow: check the sentence that turned a discussion into an apparent decision, especially when dates, owners, or approval conditions matter. Verify acceptance in the source before treating a tidy summary as an assignment.
Desktop transcription uses microphone and device audio without adding a meeting bot. Granola says it does not save meeting audio or video, while a transcript is retained. This is not a claim that processing stays entirely on your computer. The web application can display notes but cannot capture or transcribe meetings. Confirm the supported native application and your employer’s recording and participant-notice requirements before using it for work. A quiet capture method does not establish permission.
Choose around the record you need to use next: a personal reminder, an approved handoff, or older project context. Someone who needs occasional personal reminders has different requirements from a team maintaining approved project history. Before comparing subscriptions, identify the source material, intended readers, history period, and editing responsibility. These are requirements you can assess without assuming an AI summary will be correct.
Worked note: turn a one-on-one into an accurate handoff
Tessa and Ravi discuss a documentation migration on October 14, 2026. Tessa owns two redirect mappings; Ravi coordinates content review. Their source conversation also contains a vendor-budget aside that has no place in a broader project handoff. All names and details are invented. The following source excerpt and rough notes are written for this article, not transcribed by Granola.
Source excerpt: Ravi says, “The old /start and /install pages need mappings. The destinations are still subject to the content owner’s FAQ review. I’ll ask that owner tomorrow. Can you draft the two mappings by Friday?” Tessa replies, “Yes, October 16 for the draft. October 23 could be our target for the change, if the content owner clears the FAQ and the release owner agrees.” Ravi answers, “Agreed on the draft. We have not approved deployment.”
Tessa’s rough notes: “2 redirects /start /install; FAQ gate; Ravi tomorrow owner; mine Fri16 draft; maybe23 change; deployment not approved; vendor budget aside separate.”
A useful enhanced record must preserve the difference between the accepted draft deadline and a conditional implementation target. It should not expand “mine” into responsibility for release approval. Tessa’s finished editorial version reads:
Cedar documentation migration — reviewed handoff, October 14
Purpose: Prepare two redirect mappings for the old
/startand/installpages. This record covers the mapping draft and its dependencies.Accepted action — Tessa: Produce the two draft mappings by October 16, 2026. This commitment covers preparation, not deployment or approval of destination content.
Accepted action — Ravi: Ask the content owner to review the destination FAQ on October 15, 2026. The conversation contains no promised response date from that owner.
Conditional target: October 23 is a proposed change date. It depends on content clearance and agreement from the release owner. Neither condition is recorded as complete.
Open question: What date can the content owner commit to for the FAQ decision? Ravi will bring the answer back to the project discussion.
Release status: Deployment is not approved. Obtain the release owner’s decision through the team’s existing process.
This completed editorial note records agreement; it leaves the underlying project actions pending. It states what was agreed without filling gaps with plausible answers. Tessa checks the two paths, dates, and ownership against the supplied excerpt. The fictional exchange supports the draft commitment; it does not prove that the draft was later delivered or that the release occurred.
The budget aside is excluded from this handoff. However, removing a sentence from an enhanced summary does not establish that the original transcript, rough notes, or contextual chat cannot expose it. Tessa therefore keeps the mixed one-on-one source private and proposes a separate, minimal handoff record containing only approved project material. She does not use a cleaner-looking summary as evidence that the underlying conversation is safe to distribute.
Shared folders do not settle every permission question
Granola distinguishes private notes, directly shared notes, folders, and workspace spaces. Notes are private by default. Under My notes, folders are shared explicitly; folders in a company team space are visible across that workspace. Check whether the folder sits under My notes or a company team space before choosing its audience. Calling a folder “restricted” does not make a broadly visible space restricted.
Directly sharing a note with someone does not itself make that person a workspace member or create a billable seat. Granola also distinguishes Viewer and Collaborator roles. Folder membership does not automatically give someone editing permission on every note; an owner can explicitly grant Collaborator access to the relevant note. Do not buy a subscription merely to solve an access problem that depends on the owner’s sharing decision.
In the fictional handoff, Tessa records these audience decisions before distributing anything:
| Object | Intended audience and responsibility | Recorded disposition |
|---|---|---|
| Original mixed one-on-one | Tessa; any additional access requires a separate decision | Remains private; no project-folder placement |
| Minimal project handoff | Tessa, Ravi, and Mina, the implementation reviewer | Proposed invited-only folder under My notes; actual access test still required |
| Handoff editing | Tessa owns the note; Ravi needs to correct the content-review status | Request explicit Collaborator access on this note; folder membership alone is insufficient |
| Wider company update | Other company colleagues | Share only after the project audience approves its content and destination; not yet authorized |
The documentation describes summary-only viewing on the web and fuller transcript access for collaborators using desktop. It also describes contextual interactions around shared notes. Treat these as product-specific access rules to verify, not a blanket promise that anything absent from the visible summary is private. In Tessa’s case, the editorial decision is to keep the sensitive source out of the shared object.
The final permission disposition is deliberately incomplete: the minimal record is ready for review, while folder placement, recipient visibility, and editing access remain untested. This separates a finished sharing design from an actual permission change. Before a real rollout, have an authorized person check the intended reader’s view and the owner’s settings without using confidential material as a test fixture.

Personal history and a team workspace have different costs
Granola’s current pricing lists Basic at $0, Business at $14 per user per month, and Enterprise from $35 per user per month. These are displayed dollar prices, not a verified checkout quote, currency selection, or tax calculation. Basic includes shared folders and 30-day history; Business lists unlimited history. Do not assume shared folders require an upgrade.
The billing documentation says Business is monthly only; annual billing is available through Enterprise contracts. It also says Basic’s older notes are stored rather than deleted when they fall outside the 30-day in-app view. A history upgrade and recovery from deletion are different questions. Avoid calling a note lost simply because it no longer appears in the free viewing period.
Consider Tessa’s inventory on October 27. An August 20 note is 68 days old, a September 29 note is 28 days old, and the October 14 handoff is 13 days old. The older August note contains a previous routing decision she now needs to inspect. It is clearly outside the documented Basic window; the other two are within it. These ages are calendar arithmetic, not an observed Granola cutoff test.
The workspace model changes the bill. Granola says a workspace uses one plan for all its active members, including administrators. Pending invitations are not billed until people join. An administrator is not automatically a free billing-only seat. Separate workspaces have separate subscriptions, and upgrading a personal workspace does not move or share its notes into the company workspace.
The completed budget record uses three active people: Tessa, Ravi, and Mina. Ravi is also the administrator. It assumes the displayed Business rate remains unchanged and excludes taxes, credits, proration, and contract terms:
| Arrangement | Monthly base amount | Twelve unchanged monthly payments | What it would establish |
|---|---|---|---|
| Tessa’s personal workspace only | $14 | $168 | Paid history in that workspace; no company-wide upgrade |
| One company workspace, three active members | $42 | $504 | Three paid members under one workspace plan |
| Both arrangements retained | $56 | $672 | Two separate paid workspaces, not one shared entitlement |
| Company workspace after a fourth person joins | $56 | $672 at four members throughout | An additional active seat; a pending invitation alone does not produce this amount |
The twelve-payment column is a planning multiplication, not an annual subscription offer. Nor can Tessa combine one Business member with two Basic members inside the same workspace to reproduce the three-person arrangement at $14. If the organization only needs shared folders within the free history window, Basic remains a legitimate candidate. If it needs older company notes, evaluate the company workspace rather than upgrading the wrong personal archive.
Recorded decision: Tessa proposes the three-person Business workspace for longer project history, with a $42 monthly base budget. Purchase remains on hold until Ravi confirms workspace ownership, the approved audience, and data policy. The personal upgrade is an alternative for personal history, not an additional default purchase. Mina’s future participation and the actual account quote still need confirmation; no subscription was bought.
Check retention and data use separately from history access
Unlimited history does not mean every underlying source survives forever. Granola’s transcript auto-deletion documentation distinguishes transcripts from notes: a transcript can be deleted while its note remains. That can limit later regeneration and source-based questioning. A history upgrade does not establish recovery of permanently deleted transcript material. Check the relevant retention policy before relying on an old summary as a complete record.
Data use is another independent control. Granola says it may use anonymized data to improve its services by default and provides an individual opt-out. Its documentation separately says AI-provider agreements prevent those providers from using the data to train their models. Enterprise can enforce an opt-out. These statements describe different parties and settings; they do not mean there is no cloud processing, no stored transcript, or permission to capture every work conversation.
For a real evaluation, start with an approved, nonsensitive meeting and a short review period. Check whether the enhanced note preserves conditions and ownership, whether the correct workspace holds the history, and whether the intended audience can see only the approved material. Keep those checks separate from payment authorization. Use the completed requirements-and-budget record to define an approved evaluation before buying; LandOffer.ai is the publisher of this review, not the provider of the meeting-note service.


