Customer call analysis: Surfacing patterns across 100+ customer calls

TL;DR: Manually reviewing pitch and customer calls for recurring themes is impractical and introduces memory bias into investment decisions. Voice of Customer (VoC) mining solves this by using folder-level AI queries to surface pricing objections, feature gaps, competitor mentions, and churn signals across your entire call archive. Granola's Chat with folders feature lets you ask questions across hundreds of meetings at once, with inline citations pointing back to specific conversations. You jot rough notes during calls, Granola enhances them, and your team queries the whole folder for patterns that would take days to find manually.

Many investment teams focus on deal flow volume while the insights from hundreds of calls remain unanalyzed. The bottleneck is rarely a lack of data. It's the inability to query it. This guide explains how to structure call data for analysis, run folder-level queries that surface what you actually need, and move those findings directly into IC memos and Notion databases.

Manual review: Missing critical deal insights

Three pitch meetings a week, plus board meetings, portfolio check-ins, reference calls, and LP updates, generate hundreds of hours of conversation containing pricing signals, competitive intelligence, and founder red flags. The workflow itself fails structurally. Even disciplined post-call review habits cannot reliably surface cross-call patterns. A founder who mentioned burn rate concerns during diligence and a portfolio CEO who flagged the same issue in a separate board meeting are telling you something together. Reviewing them in isolation misses the signal entirely.

Hours lost to disconnected notes

The time cost compounds quickly because each call produces a transcript, a set of rough notes, and often a follow-up summary scattered across separate documents. Retrieving what a specific founder said about their go-to-market motion six weeks ago means searching email threads, Notion pages, and memory simultaneously. That friction explains why IC memos are frequently written from recollection rather than based on evidence.

When call notes live in individual documents, the knowledge belongs to whoever took them. When an associate leaves, the mental maps of hundreds of founders, market conversations, and deal patterns leave too. Shared folders with folder-level queries change this. Every call captured in a shared folder becomes part of an institutional archive that any team member can query, with citations from specific conversations according to the Granola help center.

How memory gaps weaken deal memos

Writing a comprehensive IC (Investment Committee) memo requires specific founder quotes, market claims, and risk assessments. Associates spend hours hunting through transcripts to find the single line where a CEO described their net revenue retention. If that line was not captured accurately during the original call, or if the note-taker was focused on maintaining rapport rather than typing, it is gone. The memo becomes weaker, and the deal presentation becomes less convincing.

Query past calls for instant deal insights

Granola's Chat with folders feature handles questions across all your meetings with detailed analysis, rebuilt from the ground up to be agentic. Rather than surfacing a generic summary, it answers specific questions and attributes each finding to the original conversation with an inline citation you can jump to directly.

The practical difference is significant. You can ask "Which competitors were mentioned most frequently across Q2 diligence calls and in what context?" and get a synthesized answer with links to every relevant meeting. That would take hours to compile manually.

"I love that I can use Granola for absolutely everything... Being able to turn those notes into content assets, reflections, and new ideas is priceless. For the price, Granola very quickly went from 'nice-to-have' to one of the most essential tools in my stack." - Christel C. on G2

Organizing call data for query search

The quality of a folder query depends on what is in the folder. Three folder structures work well for investment teams:

1. By deal stage: Separate folders for Screening, Active Diligence, Passed, and Portfolio. This lets you ask "Why did we pass on SaaS deals in Q3?" without noise from irrelevant meetings.
2. By sector: "FinTech Diligence 2026" or "Climate Tech Screening" enables sector-specific comparisons across dozens of pitches.
3. By meeting type: "Pitch Calls," "Reference Checks," "Board Meetings," and "Portfolio CEO 1-on-1s" as distinct folders, lets you query within a specific conversation context.

Folder-level chat is available on Business plans and above, and you can query any folder you have access to.

Folder strategies for rapid review

Concrete folder naming saves time when setting up a query. Useful naming patterns include:

• "Q2 2026 SaaS Series A Pitches"
• "Active Diligence: [Company Name]"
• "Hiring Loops: Head of Engineering Candidates"
• "Portfolio Board Meetings: [Company Name]"

Shared team folders mean every partner and associate sees the same archive. When a founder references a conversation from two months ago, anyone on the team can pull it instantly.

Pinpoint critical customer feedback

Once your folder is populated, the query does the analysis. Questions that surface the most useful signal include:

• "What are the top three reasons founders cited for choosing us over [competitor] in pitches this quarter?"
• "Summarize all mentions of pricing concerns across Series A pitch meetings."
• "Which friction points came up most often in customer diligence calls in the last 90 days?"
• "List all commitments made in board meetings this quarter with the name of who made them."

Each answer comes with inline citations linking back to the specific meeting, so you can verify the original context before putting it in a memo.

What founders consistently ask for

Pitch calls contain rich product and market intelligence, but that intelligence only becomes useful when you can query it across many conversations at once. A single founder mentioning that their enterprise buyers ask for SSO before signing is interesting. Twenty founders saying the same thing across a sector folder is a pattern you can act on.

Deal insight: Identifying feature gaps

A query like "What capabilities do customers request that are not yet built?" surfaces the gaps that define market opportunity. When the same feature request appears across ten diligence calls in a sector, it is a signal worth tracking in your investment thesis documentation.

For portfolio company support, the same logic applies. Querying "What have portfolio CEOs flagged as the top product friction point in customer conversations this year?" gives you advisory context before a board meeting without reading through every call note.

Identify top feature requests

Feature requests from customer calls tend to cluster around three types: workflow gaps, integration requests, and pricing structure concerns. Running a query like "Categorize all product feedback from customer calls into workflow, integration, and pricing themes" produces a structured breakdown you can share directly with a portfolio company's product team.

Turning requests into product features

The path from a folder query to a product decision involves one additional step: exporting the structured findings. A Granola folder query result, with its inline citations, can be copied directly into an IC memo, a Notion database, or a Slack message to a portfolio founder. The citation trail means the portfolio CEO can trace every feature request back to a specific customer conversation rather than trusting a summary.

Addressing deal-breaking pricing issues

Pricing is one of the most reliable predictors of deal friction, and it is almost always discussed in specifics during calls rather than in generic terms. Conversations contain concrete details about price points, contract structures, and the thresholds where deals stall.

Querying founder pricing feedback

A query like "At what price points did Series B deals stall this year?" returns a synthesized view of the market that would take days to compile manually. The citations let you verify the exact language a founder used about pricing objections rather than relying on memory.

Pinpointing deal-killer pricing thresholds

The most useful insight is the specific number or contract structure that causes a deal to stop moving. Queries like "At what contract value did enterprise customers in these diligence calls say procurement got involved?" or "Which pricing objections appeared in more than three calls in this folder?" surface the thresholds that should inform how you advise portfolio companies on their pricing strategy.

Distribute pricing insights to deal teams

Shared folders make these insights available to every partner and associate simultaneously. Rather than the insight living in one person's notes, the folder query output can be shared via Granola's Slack integration to push the summary into a deal channel, where the full team sees the same evidence before a partner meeting.

Pinpoint win/loss drivers from calls

Understanding why deals were won or lost is only possible when you can query across the conversations that led to each outcome. Individual call notes do not reveal patterns. A folder of passed deals from a sector does.

Querying competitor mentions

A query like "List every mention of [competitor name] across diligence calls in this folder and describe the context" surfaces whether founders are positioning against a competitor as a strength or trying to explain away a weakness. The same query run against a "Passed Deals" folder reveals whether a specific competitor keeps appearing in conversations that did not convert to investments.

Understanding why deals are lost

Passed deals contain as much signal as closed ones. Building a "Passed 2025" folder and querying it for "What were the most common reasons we passed on deals and how did founders respond to our concerns?" produces a calibration tool for refining the investment filter. Source-linked citations mean you can always read the original conversation to check whether your pattern interpretation is accurate.

Uncovering seasonal win/loss drivers

Querying a year's worth of pitch calls for "What market conditions did founders cite as headwinds in Q1 versus Q3?" reveals whether your deal flow is affected by macro cycles or seasonality in a specific vertical. Patterns like this are invisible from a single call and clear from a comprehensive folder.

Identifying churn red flags in customer calls

For portfolio company support work, one of the most valuable applications of folder-level queries is early churn detection. Customer calls contain language patterns that precede churn: repeated complaints about the same bug, unanswered questions about a specific use case, or language signaling that the buyer is evaluating alternatives.

Surfacing early churn signals

A query like "What complaints or frustrations appear more than twice across customer calls in this folder?" surfaces issues that are systemic rather than one-off. A single customer mentioning an onboarding problem is noise. Multiple customers in a quarter mentioning the same onboarding step is a retention risk.

Natural language processing enables this kind of analysis by identifying the semantic content of transcripts rather than just exact keyword matching. The query captures variations in how customers describe the same issue, not just identical-phrase matches.

Detecting pre-churn indicators

The indicators that precede churn are often softer than an explicit complaint. Sentiment shift across sequential calls with the same account, declining engagement signals, or repeated references to unresolved issues are the patterns worth watching. A folder query focused on accounts that appear in multiple calls can surface which ones show consistent tone drift over time.

Early intervention for at-risk companies

Board members using this workflow can walk into a quarterly review with a portfolio CEO already knowing which customer segments are showing friction, based on querying the folder of recent customer call notes. The conversation moves from "how are customers feeling?" to "here are three specific accounts showing warning signs and here is the exact language they used."

Creating structured call data in Notion

Moving from a folder query to a structured Notion database closes the loop between call intelligence and deal tracking. The Granola Notion integration connects your meeting notes directly to a Notion database with a single click from each note.

Structuring call data for Notion

The integration maps meeting data to Notion database fields, including title, attendees, date, and summary. This turns each call into a structured row in a deal tracker or portfolio database rather than a freestanding document buried in a folder hierarchy.

One important detail: Granola exports meetings as Notion database rows, not as standalone Notion documents. This means the data is immediately queryable within Notion's filtering and sorting tools, which is more useful for deal tracking than a document that must be opened individually.

Querying to Notion database rows

The workflow is: capture the call in Granola, enhance the notes, and send to Notion. Each row in your Notion database then contains the structured summary, and the Granola Chat citation links back to the original enhanced notes for full context. For teams that want this to happen automatically without sending each note manually, the Granola Zapier integration supports a trigger that pushes any note added to a specific folder to Notion automatically.

Notion export: Row-only insights

Because the export creates database rows rather than pages, teams can build dynamic deal tracking boards that filter by sector, deal stage, meeting date, or attendee. A "Q2 SaaS Diligence" view in Notion that surfaces every call with a specific company, sorted by date, gives the team instant context on a deal.

"With Granola I don't have to worry anymore about taking meeting notes, I can just write down things I really care about and let Granola take care of the rest. Love that I can easily share my notes with my colleagues as well, and that we can all chat with the meeting transcript so everyone can see the full context of the meeting, even if they weren't there." - Jess M. on G2

Streamlining IC memos with call insights

The IC memo problem is not a writing problem. It is a retrieval problem. The quotes that make a memo compelling, the specific founder claim about net revenue retention, and the exact pricing objection from the enterprise customer reference call are all in the transcripts. Getting them out quickly is what separates a strong memo from a weak one.

Export formats and options

Granola supports several paths for moving call intelligence into your existing tools:

• Notion: Database rows via the Notion integration, or automatic sync via Zapier
• Slack: Auto-post summaries to deal channels via the Business plan Slack integration
• CRM: Direct connection to Affinity, HubSpot, or Attio for deal tracking
• Zapier: Connect to 8,000+ apps including Google Sheets, Asana, and others

The Granola pricing page details current plan features and integration availability.

Building IC memos from call patterns

A folder query like "Summarize the three strongest signals across all diligence calls for this company, with citations" produces a paragraph you can paste directly into the IC memo. The inline citations give the IC audience a way to verify the claims against the original transcript rather than trusting the memo author's interpretation alone.

Reliable knowledge transfer to teams

Daversa Partners introduced Granola across 136 of 150 employees after their president described traditional bots as "intrusive" for the sensitive executive search conversations their work requires. Bot-free capture became a practical necessity rather than a preference. The same logic applies to IC discussions, reference calls, and any conversation in which a visible recording participant would change what is said.

Granola maintains strong retention among busy professionals by removing friction: no bot announcements, no visible participants, just a notepad that enhances what you write. When an associate leaves, their folder archive stays. Any team member with folder access can query the same conversations and get the same citations.

Try Granola for free: download the Mac or Windows app, connect your calendar, and run your next pitch call through it. Create a shared folder for your active deal diligence, add the calls, and run your first folder query before the next partner meeting.

FAQs

How many calls do you need for actionable insights?

Folder-level queries work best with enough calls to surface recurring themes. A small set may reveal initial patterns, while larger archives provide richer context for the AI to identify consistent signals with source-linked citations.

Can I query across different call types in the same folder?

Yes, you can mix pitch calls, reference calls, board meetings, and customer diligence calls in the same folder and query them all simultaneously. The citation links in each answer identify which specific meeting the finding came from, so you always know the source context.

Do I need to manually send each call to my Notion database?

No, the Zapier integration can automatically push any call added to a specific folder to Notion as a database row without manual sending. You can also send notes individually with a single click directly from each note in Granola.

How is this different from generic transcription tools?

Generic transcription tools produce individual meeting transcripts that you read one at a time. Granola's Chat with folders feature lets you query across an entire archive simultaneously, returning synthesized answers with inline citations to specific conversations. The difference is between a filing cabinet and a queryable database.

Key terms glossary

Voice of Customer (VoC) mining: The practice of systematically extracting and analyzing customer expectations, preferences, and pain points from raw conversation data using text analytics and sentiment analysis. VoC mining applied to call archives surfaces patterns across hundreds of conversations that would be invisible from individual transcript review.

NLP (Natural Language Processing): A field of computer science that applies linguistic and statistical algorithms to text in order to extract meaning from human language at scale. AI tools powered by NLP can identify themes, sentiment shifts, and recurring signals across large volumes of meeting transcripts without keyword-only matching.

Folder-level queries: The capability within Granola's Chat feature to ask analytical questions across all meetings in a shared folder simultaneously, with each answer attributed to specific source conversations via inline citations. Folder-level queries are available on Business plans and above.