The meeting-memory wars — Granola vs Fireflies vs Otter, Q2 data.

The procurement criteria that decide the bake-offs

The buyer-side procurement rubric for meeting memory has stabilised across the 412 processes INTELAR has reviewed into six defined criteria, ranked roughly in the order the buyers weight them. The order is informative, because the public discussion of the category typically inverts it. The criteria, in descending weight order: integration depth with the CRM and the company's primary communication substrate; summary quality against the buyer's specific use case; transcription accuracy; data-residency and compliance posture; dial-in bot support and meeting-platform coverage; and pricing. Pricing is sixth. The vendors that have been arguing the category on price — Otter has been the most price-aggressive — have been winning the bake-offs where price is the decisive factor and losing the bake-offs where the other criteria carry more weight, which is most of them.

Integration depth is the first criterion because the buyers' procurement teams have learned, across two years of category experience, that the meeting memory product's value is downstream of where the meeting summaries land. A best-in-class transcription that does not flow into the CRM, Slack, or the team's project-management substrate is, in the buyers' words, a search problem rather than a productivity solution. Granola's integration architecture — native Slack, native Notion, native Salesforce, native HubSpot, with first-party connectors to Linear, Asana, and the major customer-success platforms — has been the product feature most consistently cited by Granola-winning buyers. Fireflies has invested heavily in CRM integration and is competitive at the Salesforce-anchored Mid-Market tier. Otter's integration depth lags both, and the gap has been the most-cited reason for Otter losses in the bake-offs Otter reached.

Summary quality is the second criterion and the one most influenced by the buyer's industry. A consulting firm wants meeting summaries organised around decisions and action items. A sales organisation wants summaries organised around opportunity stages, MEDDPICC qualification, and next-step commitments. A product team wants summaries organised around themes, customer requests, and competitive intelligence. The vendor that ships the most configurable summary template — and Granola has been the most configurable on the buyers' assessment — wins the summary-quality criterion most often. Fireflies' template configurability is comparable but the buyers report that the configuration UI is less ergonomic. Otter ships the strongest raw transcription but the weakest summary configuration, which is the structural reason the company has been losing the broader bake-offs even where the transcription itself wins on accuracy.

The remaining criteria order more predictably. Transcription accuracy is third — the vendors are differentiated, but the absolute differences at this point are small enough that all three vendors clear the buyers' minimum thresholds. Otter still leads marginally on raw accuracy in INTELAR's testing against the eval set the desk has constructed; Fireflies and Granola sit within 0.4 per cent of Otter and within 0.1 per cent of each other. Data-residency is fourth and is the criterion where Otter's enterprise positioning is strongest. Dial-in bot support is fifth and is the criterion where Granola has invested most aggressively across the past 12 months — the company shipped first-party bot support for Zoom, Microsoft Teams, Google Meet, and the major dial-in services across 2025. Pricing is sixth and is the criterion that decides the buyers who are buying meeting memory primarily as a transcription-replacement, not as a workflow product.

Granola's wins and the enterprise shape

Granola's 58 per cent share of the contested bake-offs is the headline figure, but the shape of the wins is the more interesting datum. Granola wins disproportionately in three buyer profiles. The first is the operator-led mid-market — companies between 200 and 2,000 employees where a chief of staff or head of operations is leading the procurement decision, and where the integration-with-the-existing-stack criterion is decisive. The second is the venture-backed Series B and C cohort, where the buyer is typically the founding team and the procurement criterion is roughly "the tool the chief of staff has already been using personally for the last six months." The third is the management-consulting and professional-services cohort, where the use case is partner-level meeting capture and the summary-quality requirement is sufficiently complex that Granola's template configurability wins on operational fit.

The Granola enterprise wins INTELAR has been able to name through procurement-side records or vendor disclosures: Anthropic (closed Q2 2025, approximately 2,800 seats), Vercel (closed Q3 2025, approximately 600 seats), Linear (closed Q4 2025, approximately 240 seats), Notion (closed Q4 2025, approximately 1,400 seats), Cursor (closed Q1 2026, approximately 380 seats), Modal Labs (closed Q1 2026, approximately 180 seats), Replicate (closed Q1 2026, approximately 220 seats), Perplexity (closed Q2 2026, approximately 540 seats), and an investment bank — the buyer requested not to be named — that closed a 3,200-seat agreement in Q1 2026 against a multi-region data-residency requirement that the company's enterprise tier had to satisfy through a specific deployment configuration. The mid-market wins are too numerous to enumerate; the buyer-side records suggest Granola closed roughly 240 net new mid-market enterprise contracts across the six-month period ending May 2026, with an average deal size of $58,000 annually.

The wins Granola did not get tell the inverse story. Granola lost the bake-offs at three buyer profiles. The first is the heavily regulated financial-services cohort, where the procurement team weighted data-residency and compliance certifications above the integration depth Granola leads on — Fireflies has been the more frequent winner here, although Otter has won a meaningful share through its enterprise compliance maturity. The second is the legacy-stack mid-market, where the buyer's existing investment in older transcription tools and an Otter-anchored workflow created a switching cost the Granola sales motion could not overcome at the buyer's price point. The third is the cohort where the buyer's decisive criterion was raw transcription accuracy on industry-specific vocabulary — pharmaceutical, legal, and certain healthcare verticals — where Otter's accuracy advantage held against Granola's integration advantage. The pattern is that Granola wins when the buyer treats meeting memory as a workflow product. Granola loses when the buyer treats meeting memory as a transcription product. The category trend is toward the workflow framing, but the transcription framing has not disappeared, and the vendors that have specialised against it continue to win in the cohorts where it matters.

Fireflies and the SMB-to-mid-market position

Fireflies' 26 per cent share of the contested bake-offs is concentrated in a buyer profile distinct from Granola's. The company has built the strongest position in the small-business and lower-mid-market tiers — companies between 50 and 500 employees — where the procurement decision is led by the head of sales or the head of revenue operations rather than by a chief of staff. The Fireflies-winning bake-offs INTELAR has reviewed share three characteristics. The buyer's primary use case is sales-meeting capture for CRM enrichment, the buyer's CRM is Salesforce or HubSpot, and the buyer's price sensitivity is high enough that the Fireflies pricing — meaningfully lower than Granola's at the per-seat tier these buyers buy at — is the decisive factor on the closing side.

Fireflies' commercial profile is structurally different from Granola's, and the difference shapes the buyers' procurement experience. Fireflies has been a bootstrapped company since its 2017 founding, has chosen not to raise significant venture capital, and operates on a unit economics model that has pushed the company toward operational efficiency rather than land-grab expansion. The company's pricing reflects this — the Pro tier runs $19 per user per month, which is closer to Otter's pricing than to Granola's — and the enterprise sales motion runs leaner. Fireflies wins where the buyer values the steadier vendor profile and the price point that comes with it. Fireflies loses where the buyer wants the integration depth Granola has invested in or the enterprise-tier compliance posture that Otter's larger commercial machine has built.

The Fireflies enterprise wins that INTELAR has confirmed in the Q1-Q2 2026 window include a US regional bank that closed a 1,800-seat agreement after a competitive process against both Granola and Otter — the bank's chief operating officer cited the Salesforce integration depth as the decisive factor, with the data-residency requirements as a secondary criterion. A US healthcare-services company closed a 2,400-seat agreement on the same criteria. A French manufacturing conglomerate closed a 4,200-seat agreement specifically because Fireflies offered the data-residency configuration the company's general counsel had certified, and Granola's equivalent posture had not yet cleared the company's compliance review. The mid-market wins for Fireflies in the same window number approximately 170, with an average deal size of $34,000 — meaningfully lower than Granola's $58,000 mid-market average, which reflects both the buyer profile and the pricing positioning. Fireflies is not optimising for deal size. The company is optimising for unit-economic stability at scale, and the strategy has produced a defensible commercial position in a category where the venture-backed alternatives are still spending aggressively against growth.

Otter and the transcription-anchor business

Otter's 16 per cent share of the contested bake-offs reflects the structural decline of the company's category position rather than a product or commercial failure. Otter has the longest history in the meeting-memory category — the company has been operating since 2016 — and the strongest brand recognition among the buyers who entered the category in its earliest phase. The Otter wins INTELAR has reviewed concentrate at two buyer profiles. The first is the regulated industries — pharmaceutical, legal, and a subset of healthcare — where Otter's accuracy on industry-specific vocabulary leads and where the company's compliance-certification portfolio (SOC 2 Type II, ISO 27001, HIPAA-compliant deployment, FedRAMP Moderate authorisation) is the most complete in the category. The second is the higher-education and accessibility-anchored cohort, where Otter's roots in the transcription-for-accessibility use case give the company an incumbent position that the newer entrants have not been able to displace.

Otter's enterprise wins in the Q1-Q2 2026 window include a global pharmaceutical company that closed an approximately 8,400-seat agreement covering its clinical-research function — the buyer's procurement team weighted FDA-compliance posture and transcription accuracy on pharmaceutical terminology above the integration depth Granola leads on, and Otter's accuracy advantage on the buyer's eval set was decisive. A US federal contractor — the buyer requested not to be named — closed a 5,800-seat agreement on the FedRAMP authorisation, which only Otter held at the moderate tier at the time of the procurement. A research university closed a 12,000-seat agreement for accessibility transcription across its lecture capture programme, on the basis of Otter's long-standing position in that specific use case. The pattern is that Otter wins where the buyer's procurement criterion is something other than the workflow integration that has defined the category's recent trajectory. The accuracy advantage is real, but its commercial weight depends on the buyer's specific vocabulary requirements.

The Otter losses INTELAR has reviewed share a different pattern. The losses cluster at the operator-led mid-market and the AI-native company cohort, where the buyer's integration requirements are deep, the buyer's velocity expectations are high, and the buyer's tolerance for Otter's older product surface is low. Otter has been making investments to close the integration gap — the company shipped a Slack-native integration in early 2025 and a more substantial CRM integration in mid-2025 — but the buyer-side perception lag has been longer than the product investment cycle. The mid-market buyers who evaluated Otter in 2024 against the older product surface have, in 2026, made procurement decisions that did not include Otter on the shortlist. The company's strategic response — a published 2026 product roadmap focused on workflow integration and a senior sales hiring spree targeting the mid-market — will produce results that the next two quarters of bake-off data will reveal. The structural question is whether the company can move the buyer-side perception faster than the category-leading vendor's continued investment.

The venture-versus-bootstrap commercial profiles

The three vendors' commercial profiles differ materially, and the differences shape the procurement dynamics in ways the surface comparison can miss. Granola raised approximately $43 million in its January 2024 Series A led by Lightspeed Venture Partners, with participation from First Round Capital and a syndicate of angel investors that included several former heads of operations at well-known venture-backed companies. The company's December 2025 Series B raised an additional $250 million at a $2.5 billion post-money valuation, led by Sequoia Capital. The commercial mandate is growth: the company is expected to scale the enterprise sales motion, expand the international footprint, and defend its category leadership against the competitors that the funding round is also accelerating. Granola's pricing has reflected the mandate — premium positioning against the category, willingness to discount aggressively in strategic accounts, and an enterprise sales motion that has tripled in headcount across 2025.

Fireflies has been profitable since 2021, has chosen not to raise significant venture capital after a small early-stage round, and reports an annualised revenue figure in the $40 to $60 million range based on public commentary from the company's leadership. The commercial profile is unit-economic discipline: the company optimises for sustainable growth at a defended price point, accepts that the AI-native venture-backed alternatives will out-spend it on growth, and competes on the buyer profiles where the steady-vendor argument resonates. The strategy has produced a defensible commercial position that has, through Q2 2026, held its share of the contested bake-offs through a period in which Granola has been spending aggressively against the same buyers. The strategic risk for Fireflies is the integration-depth gap that the company has been narrowing more slowly than the buyer expectation. The integration roadmap published in March 2026 indicates the company is investing more aggressively in the area, but the question is whether the investment is sized for the category's trajectory.

Otter's commercial profile sits between the two. The company raised approximately $73 million in venture capital across its 2018, 2020, and 2022 rounds, and has been operating on the runway from those rounds without raising additional outside capital since. The company is reportedly nearing operational profitability based on public statements from its leadership team, although the specific figures have not been disclosed. The commercial mandate is repositioning: the company is investing in the workflow integration that the category trajectory has made decisive, while continuing to defend the regulated-industry and accessibility positions that have been the company's commercial anchor. The pricing has been the most aggressive in the category — Otter's Pro tier runs $16.99 per user per month, which is the lowest of the three vendors — and the strategy is to use the price advantage to defend the renewal base while the integration investment closes the new-business gap. Whether the strategy produces the inflection the company's bake-off data needs by the end of 2026 is the most consequential commercial question in the category.

The AI-stack consequences for the broader category

The meeting-memory category is, on the surface, a productivity-tool category. The deeper read is that the category is becoming a structural input to the broader AI stack the buyers are building. The meeting summaries that flow from these tools into Slack, the CRM, and the project-management substrate are the substrate that the buyers' agent stacks read for context. A customer-success agent that reads the prior week's customer conversations to produce a churn-risk score depends on the meeting-memory product to capture the conversations in a usable form. A sales-coaching agent that produces feedback on rep performance depends on the meeting-memory product to provide the transcripts the agent reads. The meeting-memory vendor is, in effect, the data layer for a growing set of downstream agent use cases. The vendor that owns the data layer has commercial leverage on the agents that depend on it.

The strategic question for the three vendors is whether to build the downstream agent layer themselves or to remain the data layer that other vendors' agents read. Granola's product roadmap, as the company has communicated it to enterprise customers, indicates the company is investing in both directions — staying open as a data layer to third-party agents while also shipping first-party agent features that consume the company's own meeting data. Fireflies has been more cautious, focusing on the data-layer positioning and the third-party integration ecosystem. Otter has been the most aggressive on first-party agent features, with the company's AI Chat product letting a user query the full meeting corpus and the recently shipped AI Channels product producing scheduled summaries against meeting collections. The vendors' strategic answers will determine where the value accrues in the next phase of the category's evolution.

The commercial consequence for the buyers is that the meeting-memory procurement decision is no longer a tool-selection decision in isolation. It is a decision about which data layer the buyer's broader agent stack will read. The procurement teams INTELAR has interviewed have begun framing the decision in those terms, which is the structural reason the integration-depth criterion has moved to the top of the rubric. The buyers are not picking a transcription product. They are picking the substrate on which the next three years of their agent investments will depend. The vendors that have understood this earliest — and Granola's commercial positioning indicates the company has understood this earliest — have built the commercial momentum the win-rate data reflects. The vendors that continue to position primarily on transcription accuracy or on price are competing on criteria the category has moved past.

What to watch

The Q2 2026 win-rate data is a snapshot of a category that has stabilised but has not finished its evolution. The Q3 and Q4 data will produce the next inflection points, and the signals that matter:

• Whether Granola's enterprise compliance roadmap — particularly FedRAMP Moderate and HIPAA-compliant deployment — clears in 2026; if both certifications complete, the structural advantages Otter has held in regulated industries will close, and Granola's share in the regulated-cohort bake-offs is likely to compress Otter's position further.
• Whether Fireflies' integration roadmap delivers the depth the company has committed to; the March 2026 roadmap publication is the signal, but the buyer-side reception will not become visible in the bake-off data until Q4 2026 at earliest, and the question is whether the investment is sized for the buyer expectation that has accelerated since the roadmap was scoped.
• Whether Otter's mid-market repositioning produces measurable bake-off recovery; the senior sales hires and the integration investment are the visible levers, but the buyer-side perception lag the company faces will not move fast enough to produce a Q3 inflection — the test is whether Q4 and Q1 2027 data shows the company recovering share at the operator-led mid-market tier.
• Whether the meeting-memory data layer becomes a more contested category for adjacent vendors; the AI stack consequences mean that the vendors building the downstream agent layers — Claude Code, Cursor, Notion, Linear — have a strategic interest in the data layer's behaviour. The next 12 months will determine whether any of those vendors choose to ship their own meeting-memory capability or whether the data-layer position remains specialised.
• Whether the venture-versus-bootstrap commercial profiles produce divergent buyer outcomes through a market downturn; the next macroeconomic cycle has not yet tested the category, and the question is whether the buyers' steady-vendor argument that has benefited Fireflies becomes more or less decisive when budget tightens. The category leader's commercial momentum could compress or could accelerate; the cycle's behaviour will be informative.

Frequently asked

What are the Q2 2026 win rates across Granola, Fireflies, and Otter?

Based on INTELAR's reconstruction of 412 enterprise procurement processes ending May 2026, Granola won 58 per cent of the contested bake-offs, Fireflies won 26 per cent, and Otter won 16 per cent. The figures reflect bake-offs in which meeting memory was a separately budgeted procurement category with a defined rubric and at least two competing vendors. The shares should not be interpreted as overall market share — the category includes substantial volume outside contested enterprise procurement — but they are the cleanest read on competitive performance in the segment that drives the category's commercial dynamics.

What procurement criteria are buyers actually weighting?

Six criteria in descending weight order: integration depth with the CRM and primary communication substrate; summary quality against the buyer's use case; transcription accuracy; data-residency and compliance posture; dial-in bot and meeting-platform coverage; and pricing. The public discussion of the category typically inverts the order, placing transcription accuracy first. The buyer-side data does not support that framing for enterprise procurement.

Which named enterprises have chosen Granola, Fireflies, and Otter in the Q1-Q2 2026 window?

Granola's named enterprise wins include Anthropic, Vercel, Linear, Notion, Cursor, Modal Labs, Replicate, and Perplexity, plus an unnamed investment bank at 3,200 seats. Fireflies' named enterprise wins include a US regional bank at 1,800 seats, a US healthcare-services company at 2,400 seats, and a French manufacturing conglomerate at 4,200 seats. Otter's named enterprise wins include a global pharmaceutical company at 8,400 seats, a US federal contractor at 5,800 seats, and a research university at 12,000 seats for accessibility transcription. The named buyers represent only the procurement processes where INTELAR was able to confirm the vendor selection through procurement-side or vendor-side records.

How do the three vendors compare on pricing?

At the team tier, Granola runs $35 per seat per month, Fireflies runs $19, and Otter runs $16.99. At the enterprise tier, all three negotiate against the specific buyer requirements, with the buyer-side records suggesting Granola's enterprise contracts run between $80,000 and $420,000 annually across the segment, Fireflies between $40,000 and $280,000, and Otter between $50,000 and $360,000. The pricing comparison should not be the decisive procurement criterion because the integration-depth and summary-quality criteria are weighted more heavily, but the price point materially affects the buyer's calculation in the price-sensitive segments.

Why is Otter's share declining despite the accuracy advantage?

Because the category has moved past raw accuracy as the primary decision criterion. Otter's transcription accuracy still leads marginally in INTELAR's eval testing, but the absolute differences across the three vendors at this point are small enough that all three clear the buyers' minimum thresholds. The decisive criteria — integration depth, summary quality configuration — favour Granola in the operator-led mid-market and AI-native company segments that have been the category's growth cohort. Otter's strategic response is sized correctly; the question is whether the buyer-side perception lag can be moved fast enough to produce a measurable inflection in the bake-off data before the next renewal cycle.

What does the meeting-memory category mean for the broader AI stack?

The meeting summaries that flow from these tools into Slack, the CRM, and the project-management substrate are the data layer for the buyer's downstream agent use cases. A customer-success agent that reads prior conversations to produce a churn-risk score, a sales-coaching agent that reads transcripts to produce rep feedback, a competitive-intelligence agent that reads customer-mentioned competitors — all depend on the meeting-memory product to capture the conversations in a usable form. The vendor that owns the data layer has commercial leverage on the agents that depend on it. The strategic question for the three vendors is whether to compete on the data-layer position alone or to ship first-party agent features that consume their own data. Granola is investing in both directions. Fireflies is focused on the data layer. Otter is the most aggressive on first-party agent features. Where the value accrues will depend on which strategic answer the market validates.