
Granola's customer experience team has found a way to write help documentation ten times faster by using the company's own transcription features. The workflow, shared on LinkedIn, combines engineering meetings with automated transcript processing to produce ready-to-edit articles.
The approach relies on Recipes, a feature Granola introduced last September that transforms meeting transcripts into specific document types. For help documentation, the CX team schedules a meeting with the engineer who built a feature, runs the Recipe during or after the call, and receives a structured draft that includes feature descriptions, use cases, step-by-step instructions, and suggestions for screenshots.
How the Workflow Works
The process starts with a conversation between CX staff and engineering. Rather than the CX team researching features independently or waiting for written specifications, they talk directly to the person who built the functionality. Granola records and transcripts the discussion automatically.
After the meeting, the team runs a Recipe specifically designed for documentation. The Recipe analyzes the transcript and outputs a draft article structured for end users. This includes what the feature does, why someone would use it, how to use it step by step, and which parts of the interface should be shown in screenshots. The team then makes light edits to refine the draft before publishing.
Granola's team reports this cuts documentation time dramatically, though they have not published specific metrics. The company has made the Recipe available for other teams to use, and included a link in the LinkedIn post comments.
Recipes Beyond Documentation
Recipes are pre-configured prompts that process meeting transcripts into specific outputs. When Granola launched Recipes, the company positioned them as a way to standardize how teams extract information from meetings. The feature works within Granola Chat, which can search across all of a user's meeting notes.
Beyond help docs, Recipes can generate product requirement documents, interview summaries, sales call follow-ups, and other formatted outputs. Each Recipe is designed for a particular use case, with prompts tuned to pull relevant information from transcripts and format it appropriately.
The documentation Recipe appears to focus on translating technical explanations into user-friendly language, structuring content pedagogically, and identifying visual aids. For teams that write documentation reactively after features ship, this could compress timelines significantly.
Context for Documentation Teams
Documentation has long been a bottleneck in software development. Engineers build features faster than support teams can document them, leading to gaps in help centers and increased support load. Traditional approaches require documentation specialists to reverse-engineer features, test workflows, and write from scratch.
Granola's approach shifts documentation earlier in the development cycle. By capturing the engineer's own explanation during a structured conversation, the Recipe has access to information that might never make it into a spec: why design decisions were made, what edge cases exist, what users should watch out for. This oral knowledge transfer, combined with automated structuring, could produce more accurate initial drafts than documentation written in isolation.
The workflow also distributes documentation work differently. Rather than CX teams doing all the research and writing, engineers contribute their expertise in a low-friction format, a meeting, while automation handles the initial structuring. The CX team's role shifts toward editing and refinement rather than creation from zero.
For teams already using Granola for meeting notes, the Recipe represents a repurposing of existing transcripts. Meetings that would have produced only notes now also produce documentation drafts, increasing the return on time spent in conversations. Whether this workflow scales beyond small teams or specific feature types remains to be tested by the broader user base.
Sources
1 checkedHow we cover tool news: Create With's tool desk drafts these reports with AI from the sources listed above and checks them against those sources before publishing.




