Fall 2026 · Gladia Research Report

2,100 users surveyed across 6 countries

Fall edition

2026 AI Meeting Assistant Market Map

Aka Gladia’s “Magique Square” for meeting assistants.

Featuring

  • Gong
  • Glean
  • Granola
  • Wispr Flow
  • Fireflies
  • Otter.ai
  • Krisp
  • Fyxer
  • Jamie
  • Leexi
  • Noota
  • Claap
  • Jiminny
  • Attention
  • Circleback

What this report is

What is this?

A map of the meeting assistant market in 2026 that explores how products are differentiating, and where the category is heading. Our framework is built around four “moats” of defensibility, and our POV is informed by product testing, survey data, and conversations with the investors and founders shaping the market.

How do we define “meeting assistants”?

Any AI-powered platform that (magically) turns a conversation into an artifact. Notepads, notetakers, operating systems, ambient AI… We don’t really care what they call themselves, so long as voice is the input for some sort of useful output.

Who’s it for?

Primarily founders and product teams operating in (or adjacent to) this space. But honestly? Anyone trying to make sense of the market. That includes investors, enterprise buyers, everyday consumers, and anyone else who cares about how AI is reshaping how work gets done.


New insights
The user’s POV We asked 2,100 users across the US and Europe what they use, what they trust, and what would make them switch.

In May 2026 we published the first edition of this report: our “Magique Square” for a category we think is one of the most interesting in software. Our view hasn’t changed, but the market around it has… even faster than we expected.

Here’s what’s new in the last few months:

1. Meeting notes are now table stakes

Everyone wants and expects meeting notes now. That’s why many employees are bringing their own tools, with a third of them choosing their meeting assistant themselves.

Demand is so high, that even adjacent products are adding notetakers of their own, from enterprise search (Glean) and dictation (Wispr Flow) to CRM (Attio) and email (Fyxer).

The next frontier is intelligence. How much context does it collect beyond the meeting itself? How deeply does it connect to the systems where work actually happens? Can it recall what was said three months ago, across hundreds of conversations? And can it help in real time, while the conversation is still happening?

The leaders are already competing on these.

Otter launched live in-call coaching in July, Fireflies launched a sales suite, an email assistant, and voice agents in just over two months, and nearly every vendor shipped an MCP server so its meeting data can flow into other AI tools.

2. Ambient capture is going mainstream

When our first edition came out, we described bot-free and portable, always-on capture as the least enterprise-friendly corner of the market.

Two companies proved the model works anyway. Granola built its product around bot-free capture and was valued at $1.5 billion in March. Plaud built its business on hardware and has now passed 2 million users and $100 million in ARR.

Now others are following. Fellow now lets users choose bot-free recording straight from Google Meet, and Google brought AI notes to in-person meetings. We expect that to continue.

Devices are getting harder to spot, too. TechCrunch reporter Ivan Mehta recently tested the Vocci Ring, a recorder that could pass for an ordinary ring or fitness tracker. For all intents and purposes, it’s invisible to the people it could be recording.

This is a tension that will continue to define (or disrupt) the market: (in)visibility versus compliance and governance. Users want capture they never have to think about, and organizations need capture they can control. The vendors that deliver both will have a serious edge.

3. The standalone notetaker has two exits: get acquired or expand

In May, we argued that meeting assistants are the wedge into something bigger: a true AI workplace assistant that understands what’s happening inside an organization and acts on it. It’s now clear that there are two ways forward, which we explore in more detail here.

The first is to plug into someone else’s product.

In June, Aircall bought Piper to turn calls on its phone platform into CRM updates. In September, Superhuman bought Fathom after testing its own notetaker and finding it harder to build than expected. Those deals follow a year of similar moves: lemlist bought Claap, Workday acquired Sana, and Salesforce bought Momentum to feed conversation data into its AI agents. Each buyer already owned a workflow or a system of record, but not the conversations feeding it.

The second is to outgrow the notetaker label. That’s a product challenge and a marketing one. You have to first build the assistant, then convince the market you’re more than a notetaker. That’s why Otter hired a new CMO in August to take the company beyond the notetaking category, and why Granola is now pitching itself as an enterprise context layer rather than a notes app.

Want to learn more about recent acquisitions and funding rounds? Check out this article.

TLDR: The notetaker was never the end product, and what comes next will be something far more capable and useful than most of us have ever imagined.

What we do know is that voice will remain the primary input. And it’s about damn time. It’s been the most underutilized data source in every organization for far too long.

Trust us. We bet the company on it.

Jamie was capturing meetings before ChatGPT had even launched. Since then, it’s grown from a note-taker into a meeting intelligence platform for deal teams and regulated industries.

We spoke to Benedikt Böringer, CEO and Co-Founder of Jamie, about why context matters, how bot-free capture and governance can coexist, and what this market might look like in five years.

Q: You’ve described automating meeting notes as the first step on a longer journey. What did that journey look like at the start, and where is it heading?

A: Taking the manual work out of note-taking was always the first step. It was never the destination. We discovered quite quickly that the real value isn’t in the notes themselves, it’s in the context that comes with them.

Today, that context is something you can build on. You can use it in workflows, your agents can draw on it, and it stays useful long after the meeting ends.

Where it’s heading is something like a super brain for your organization: one that doesn’t only capture and hold knowledge, but can act on it. The user always stays in control, so nothing happens without you unless you want it to. Once that’s in place, you can eliminate all kinds of busywork around meetings and get far more out of every conversation.

Q: You’ve mentioned context a few times. Why does it matter so much, and what makes it hard to get right?

A: Without it, knowledge stays with whoever was in the room. I spend a large part of my day in meetings. If I weren’t using Jamie, all that context would only stay with me. Now I can reuse it and make it accessible to my team and my agents.

That compounds over time. If a partner learned something five years ago, a new hire can access it on their very first call, the moment it becomes relevant.

Getting there is harder than transcription, though.

Reflecting what was said word for word is only the start. The challenge is contextualising it across the conversation. A point raised at the beginning of a call might come back at the end, and a useful summary has to connect those moments rather than list them in order. The same problem then repeats at a larger scale: connecting what was said in one meeting to what was said in another, weeks or months apart, by different people. And without a bot, you’re also working out who said what in a room, not just on a call.

Getting all of that into a format that’s actually useful is a real challenge, and it’s one we’ve been solving for years.

Q: Bot-free capture and enterprise governance can look like they’re at odds. How does Jamie handle that?

A: Jamie has been bot-free from day one, unlike others that have shifted to it since. There were two reasons.

The first is atmosphere. When a bot joins a call, it changes the conversation, especially a personal one. People feel constantly watched, and that can stop them talking openly. We’ve all joined a call where the other person isn’t there yet but their bot already is. Nobody likes that experience.

The second is coverage. A bot limits you to online meetings. As soon as a conversation is hybrid or in person, it can’t tell who said what. For the conversations our customers have, that would leave out a lot of what matters.

But bot-free doesn’t mean invisible. Jamie is very open about what it does, and users and admins control how people are notified that Jamie is in use. That can be a consent email before the meeting, a compliance notice in the chat, or admin controls that enforce it across the organisation, so nobody can use Jamie without that consent notice. We also offer transcript redaction.

And then there’s the data itself. We’re a German company, GDPR-compliant, and with all data hosted and processed in the EU. Our customers work under very strict compliance requirements, so we built the controls to match.

Q: MCP has gone from rare to table stakes in under a year. How has that changed what people can do with their meeting context?

A: When we started Jamie, ChatGPT hadn’t launched yet. Today, a huge amount of work happens inside Claude, ChatGPT and other LLMs: drafting, researching, planning, preparing for the next conversation. If your meeting context can’t get into those tools, it’s missing from exactly the place it would be most useful. So for us (and many other vendors…) MCP was an obvious capability.

It’s also essential to the bigger vision. A brain for your organisation can’t live inside one app. It has to be reachable wherever the work happens, and by whatever is doing the work, including your agents.

In practice, that means asking Claude what a client pushed back on last quarter while you’re writing their proposal, or having an agent pull this week’s commitments into a follow-up plan, without leaving the tool you’re in. Alongside our native integrations, webhooks and API, it means Jamie can connect to pretty much anything.

Q: In our first report, the knowledge graph and the live coach sat side by side. One has deep context; the other helps in the moment but has little to draw on. What happens when that barrier collapses?

A: For a long time, you had to choose between speed and depth. A live tool without depth gives you generic prompts. A deep tool that only works after the call gives you great answers too late. Delivering both at the speed of a conversation is a genuinely hard product challenge.

But the demand is obvious, and live, context-rich assistance is what excites me most right now. Imagine: no more making a note to check something later, or scrambling to remember what was said during the call. I’d have more context during the conversation, which makes me better at it and makes meetings far more productive. Ideally, I don’t need to do anything. I just automatically get the right information at the right time, based on whatever preferences I’ve set.

The agentic workflows we’ve recently launched are a step in that direction. They go beyond capture to make sure every piece of information lands in the right place, so you can eliminate more of the busywork around meetings. Briefings, which are already live, are a good example; they put the right context in front of you before the call starts. Live support during meetings is where this goes next.

Xavier Lombard is not a first-time founder. Before Leexi, he built and sold two companies. At the second, an insurtech, he ran a team of around fifty people who spent their days in meetings and on the phone, and he saw how much value was lost after each conversation.

That became the problem Leexi was built to solve.

In 2021, he co-founded Leexi with his wife Carole and their sons Baptiste and Matthieu. Unlike many of its peers, which have relied on large venture rounds, Leexi is entirely self-funded and has now passed €10M in annual recurring revenue.

We talked to Xavier about how ChatGPT reset the market, what staying independent makes possible, and why sovereignty has become a deciding factor for buyers.

Q: How has the market changed since you started, and what has stayed the same?

A: ChatGPT changed expectations overnight: summaries went from a novelty to a given. Since then, three things have moved. Video meeting platforms now ship their own assistants. Free summaries have appeared inside other products. And the market is consolidating, as Superhuman’s acquisition of Fathom in September 2026 shows. What has not changed is the underlying need: capturing decisions without losing them, and trusting the tool with sensitive conversations. That is why we kept investing in accuracy, in security, and in being independent of any one meeting platform.

Q: What makes real-world meetings so hard to get right?

A: When you grow up with French, Dutch, and German on your doorstep, you know that real meetings are rarely in one clean language. People switch languages mid-sentence, bring accents, and use internal jargon. So multilingual support is a design principle for us, not a feature, and Leexi transcribes and summarizes in 100+ languages. Speech recognition quality is the foundation of everything else, which is why we work with specialists such as Gladia.

Then come the problems that sit on top of it. Speech has to be attributed to the right speaker, especially in hybrid rooms. Every answer has to be grounded in what was actually said. Access has to respect who is allowed to see which meeting. And actions have to be pushed reliably into customers’ own tools.

Q: Why did you choose to self-finance, and what does independence let you do?

A: We chose to grow from our own revenue, with no outside investors. Independence gives us three things. First, a long-term view: no exit timeline pushes us toward quick wins at the expense of customers. Second, the freedom to follow customers’ needs, which is how security and governance became part of the product from the start. Third, ownership that stays European, which matters for sovereignty.

The trade-off is real. We do not have the marketing budgets of companies that raise hundreds of millions of dollars. So we compete on product, focus, and the quality of our customer relationships.

Q: How has digital sovereignty gone from a label to a buying criterion?

A: In 2022, buyers mostly asked about features and price. Today, security and legal teams are in the room from the first meeting. They want to know where the data lives, who the sub-processors are, whether it is used to train models, how long it is kept, and how deletion works. Regulated customers, banks in particular, now ask about AI governance as well.

That shift has given sovereignty a concrete meaning: the customer stays in control. That means knowing where data is hosted (for us, in France, with no customer data used for training), which companies handle it and under which laws, which AI model is used, and who can access it. It also means the vendor’s own independence, both in its ownership and in not depending on the rules of a single meeting platform. Sovereignty is not a label. It is something a buyer should be able to verify.

Q: Where are AI meeting assistants heading?

A: We see three shifts. First, from notes to action. People no longer want a transcript, they want the next step done, so assistants will prepare meetings, follow up, and update other tools. Second, from stand-alone tools to suites, as the market consolidates. Third, from after the meeting to during it, with real-time transcription, subtitles, and translation. Across all three, trust and governance will weigh more and more in buying decisions.

03

Not every vendor is playing the same game. These four maps reveal where the market is crowded, where there’s white space, and where the most interesting strategic bets are being made.

Disclaimer: placement reflects the primary value proposition customers buy, not every capability a product offers.

Moat #1 · Capture

Where products listen, and who decides what gets recorded (and how).

Walled Gardens

IT sets the policies, enforces consent, and decides where recordings live. What they capture mostly happens inside their own meeting platforms, with some in-person notes too. Fellow, Microsoft Copilot, Google Meet AI, Zoom (ZoomMate)

Compliant Ears

The hardest quadrant to build in. These tools capture more ambiently by extending beyond meeting platforms, without sacrificing the governance and compliance controls enterprises require. Dragon Copilot, Abridge, Nabla, Leexi, Freed

Personal Recorders

Easy to adopt and built for the individual. These tools operate on the user’s terms, not the organization’s, and they’re extending beyond online meetings into in-person conversations, phone calls, and even smartwatches. Fireflies, Fathom (acq. Superhuman), Read AI, Wispr Flow, Granola, BlueDot

Open Mics

The broadest reach of any quadrant. These tools push furthest toward ambient capture, following users beyond scheduled, online meetings into the rest of their day. They’re built for the individual, though a few carry the certifications and admin controls enterprises look for. Plaud, Notta, Voicenotes, HeyPocket

Strategic Insight

Capturing online meetings is table stakes. The new race is to capture everything else, as quietly as possible. Bot-free is the norm, and recorders now fit on a lapel or the back of a phone. Governance is catching up, with horizontal tools starting to adopt the playbook healthcare wrote. But consent is still largely unsolved.

“A tool that captures 80% but misses the critical 20% isn’t 80% as valuable. It’s a fundamentally weaker product.”

Julien Bek, Partner at Sequoia

Moat #2 · Intelligence

What products do with what they’ve captured, and when those insights are delivered.

Knowledge Graphs

Most of these tools wouldn’t call themselves notetakers. They’re intelligence platforms that win through accumulation, connecting conversations, documents, and systems across people and teams to map the tribal knowledge that lives inside an organization. Glean, Tana, Sana (acq. Workday), Otter.ai, Piper (acq. Aircall), Read AI, Noota

Live Agents

Where other tools produce outputs, these take action, answering questions, updating systems, and moving work forward before the call ends. Live action is getting more capable, but its value depends on depth. The more context an agent can draw on, the better its actions get. Jamie, Attention, Timeless, Fireflies

Structured Summarizers

Reliable, easy to love, and often built into the tools teams already use. These tools deliver clean post-meeting summaries and structured notes from a single conversation. What they don’t do is connect the dots across meetings. MeetGeek, Microsoft Copilot, Attio

Verbatim Recorders

These tools work in the moment, but only on the conversation in front of them. Think live transcripts, real-time translation, and AI features that can catch you up and fill in the gaps. Notta, Tactiq, Krisp, HappyScribe

Strategic Insight

We used to think the choice was depth or speed. It isn’t anymore. The most interesting tools now act during the call, drawing on everything an organization already knows. This is the moat where the agentic workplace assistant is being built.

“The most interesting company would surface information when you need it. Actually, when you don’t even know that you need it. We’re not there yet. It needs to proactively surface things. Today it’s very, very reactive.”

Naseem Moumene, VP at Northzone

Read the full Q&A

Moat #3 · Integrations

How deeply a product embeds itself into the tools a business runs on, and how well it connects the people who use it.

The Utilities

Wide reach, shallow roots. These tools are often adopted by whole teams and connect to plenty of apps, but mostly in one direction. Otter.ai, MeetGeek, Circleback, Timeless, Fyxer

The Fortresses

The hardest quadrant to displace. These tools combine deep, bi-directional integrations with wide team adoption. The product becomes the place where outputs live and where teams build shared context. Hard for product teams to build, hard for buyers to replace. Attio, Gong, Fellow, Jamie, Momentum (acq. Salesforce)

Focused Transcribers

Built for the individual, not the organization. These tools do one thing well - capturing and exporting meeting content - but they don’t attempt to deeply embed themselves in broader workflows or connect the people around them. Tactiq, Voicenotes, BlueDot, HappyScribe, HeyPocket

Specialized Assistants

Narrow by design, but deeply embedded within their niche. These tools integrate tightly with the systems of record that matter most in their domain - CRM, EHR, ATS - but the data they handle is sensitive by nature. Used by many, shared by few. Dragon Copilot, Suki, Claap (acq. lemlist), Sybill, Jiminny, Praiz

Strategic Insight

Integrations still decide whether a product stays a tool or becomes a platform, and the whole market is moving toward the platform end. Tools are being adopted by wider teams and wired more deeply into the systems they run on. MCP made connecting easy. Depth is now what separates a convenient tool from one a business can’t live without.

“For these meeting assistant companies to become really big, they need to crack collaboration. That’s the next stage where they can really become systems of work. For that same reason, the top player will be significantly larger than number two and number three. Network effects from collaboration features are strong.”

Julien Bek, Partner at Sequoia

Moat #4 · Market Strategy

How a product reaches its users, and what that relationship is ultimately worth.

Vertical Giants

Specialized tools sold to enterprise buyers within a specific vertical. The technical complexity is high, user expectations are exacting, and that’s reflected in the price tag. These products grow by going deeper. More workflows, more integrations, more of the stack owned within the niche they already dominate. Abridge, Suki, Attention, Nabla

Enterprise Giants

The biggest contracts and the broadest deployments. These tools typically arrive through existing platform relationships or dominant distribution. Lower price per seat than their vertical counterparts, but what they lack in depth they make up for in scale. Gong, Zoom (ZoomMate), Google Meet AI, Momentum (acq. Salesforce)

B2B Boutique

These tools are built for specific, smaller audiences (financial advisers, recruiters, software teams), and note-taking is often just one feature of the overall product. The total addressable market is limited by definition, which raises a real strategic question: do you go deeper into the niche like a Vertical Giant, or broader like a PLG Mover? Metaview, Auctor, Saturn, Noota, Supernormal, Syslea, Fysio Scout, Easier Labs

PLG Movers

Where the energy is right now. Low friction and often free to start, but increasingly moving upmarket: expanding into team deployments, adding governance, and chasing enterprise contracts. Fathom (acq. Superhuman), Granola, tl;dv, Wispr Flow, Plaud

Strategic Insight

Last time, we said vertical had an edge. It’s showing. The horizontal market is crowded, squeezed between bundled platforms and PLG tools racing upmarket. Meanwhile, the quadrant built for niche audiences went from the thinnest on the map to the most crowded. Some of the most exciting products in this market solve one specific problem, for one specific profession, very well.

“Many of [these products] have captured that core behavioral loop: join a meeting, get notes, retain users. But now it’s becoming a question of who’s going to build out the rest of the features. And that’s still all to play for.”

Naseem Moumene, VP at Northzone

Read the full Q&A

06

This report was informed by a survey of 2,100 meeting assistant users across the US, UK, France, Germany, Italy, and Spain, conducted online (CAWI) by OnePoll from March 20 to 27, 2026, and by interviews with investors and founders.

Our research started with 100+ vendors, selected based on market relevance and product differentiation across the category.

Once the axes and criteria for each moat were established, vendors were shortlisted to ~50 against rubrics defined per moat. Placements are based on product testing and publicly available information gathered between February and April 2026. Not every vendor appears in every moat; inclusion reflects standout positioning rather than exhaustive coverage.

We’re not analysts. Placements reflect a point-in-time view based solely on what was publicly available. There were no vendor briefings, and this space moves fast. That means some of this will date quickly.

hand-drawn with care - May 2026
thanks for reading all the way down here
Gladia Research

Executive summary

Q&A with Jamie’s CEO and Co-Founder

Q&A with Leexi

The Magique Quadrants

Back to the Market Map

Gladia Research · Meeting Assistant User Survey

Low loyalty, high stakes: what 2,100 meeting assistant users really want

We surveyed 2,100 people who use AI note-takers at work across the US, UK, Germany, France, Spain and Italy. What we found surprised us: the tools users know best aren’t the ones they’d pick, and most are ready to move. Here’s what’s driving it, and what they’ll switch for.

What you’ll find

  • Which meeting assistants users have actually heard of, country by country
  • Who plans to switch in the next 12 months, by market, age, role and sector
  • How users rate their tool on 9 attributes, from speed to domain terminology
  • Where accuracy breaks: names and numbers, jargon, multilingual meetings
  • Country cheat sheets for the US, UK, Germany, France, Spain and Italy

01 Brand awareness

Three platforms own awareness. Twenty-three tools fight over the rest.

Teams, Google Meet and Zoom are known by half or more of users. Every independent note-taker is fighting for what’s left. That’s not because the products are bad. It’s because most users never picked their tool: it came from IT, from leadership, or bundled with software the company already paid for.

Who users have heard of

Awareness values for the 23 independent tools and the country splits are locked. Unlock with your professional email.

  1. Microsoft Teams Copilot / Teams notes61%
  2. Google Meet / Gemini notes55%
  3. Zoom AI Companion50%
  4. Attention
  5. Avoma
  6. Circleback
  7. Claap
  8. Commure
  9. Fathom
  10. Fireflies.ai
  11. Freed
  12. Gong
  13. Granola
  14. Jamie
  15. Jiminny
  16. Krisp
  17. Leexi
  18. MeetGeek.ai
  19. Nabla
  20. Otter.ai
  21. Read AI
  22. Supernormal
  23. Sybill
  24. Tactiq
  25. tl;dv
  26. WisprFlow

Only one independent note-taker breaks 20% awareness. Is it yours? Answer locked. Unlock with your professional email.

Q1. Which AI meeting assistants or note-takers have you heard of? Multi-select, n = 2,100.

Most users can’t name an alternative. Most plan to switch anyway.

02 Switching intent

59% plan to switch in the next 12 months

Users didn’t pick their tool, and they aren’t attached to it. Nearly 6 in 10 say they’re likely to switch within a year. In one market, it’s almost 9 in 10.

59% likely to switch meeting assistants in the next 12 months, all markets

Likely to switch, by market

Market values are locked. Unlock with your professional email.

  1. France
  2. Germany
  3. Italy
  4. Spain
  5. United Kingdom
  6. United States

Q9. How likely are you to switch AI meeting assistants in the next 12 months? n = 2,100.

So what would make them switch? The features they actually rely on.

03 Why they switch

They trust their notes. They’ll switch for the ones that get the details right.

97% of users trust their AI meeting notes at least to some extent. But trust depends on what they ask the tool to do. The features people rely on every day, like names, numbers, jargon and multiple languages, are the ones that decide what they pick next.

97% trust their AI notes, at least partly. Only a locked share trust them completely.

48% are in multilingual meetings frequently or in almost every meeting. Yet only a locked share list multilingual as a top priority.

For vertical tools, domain language is the whole product. In legal, clinical and finance settings, a misheard term isn’t a typo, it’s a liability. Domain terminology is a locked ranking.

Which features decide the switch, and where today’s tools fall short, market by market, is in the full data.

The full dataset

59% of meeting assistant users plan to switch. See what they’ll switch for.

Every question, cut by country, sector, role and brand. Instant access.

What you unlock

  1. What users prioritize when choosing
  2. The performance scorecard, with the weakest markets
  3. The multilingual blind spot and trust, market by market
  4. Awareness vs. real usage, and who actually picks the tool
  5. Who’s about to switch, and where
  6. Country cheat sheets for the US, UK, DE, FR, ES and IT

By unlocking, you agree to our Privacy Policy.

04 What users prioritize

The #1 reason users pick a meeting assistant isn’t speed

05 The performance scorecard

Only a locked share of UK users rate their tool well on names and numbers

06 The multilingual blind spot

In Spain, only a locked share never have multilingual meetings

Plus trust, adoption, who picks the tool, who’s switching, high-stakes meetings and six country cheat sheets.

Methodology

Survey conducted online (CAWI) by OnePoll for Gladia, 20–27 March 2026. 2,100 adults who use or have used an AI note-taker or meeting assistant at work. Subgroups under 80 respondents are excluded from reporting. Awareness-to-use conversion is a Gladia analysis of Q1 and Q2.

Read the Market Map

Methodology