#274 Luke Behnke: Why Grammarly Is Going All In on AI Agents

AGNTCY - Unlock agents at scale with an open Internet of Agents. Visit https://agntcy.org/ and add your support. Grammarly is no longer just a writing assistant. It's building an AI productivity platform that could rival Microsoft Copilot. In this episode, Luke Behnke, VP of Enterprise Product at Grammarly, shares how the company is moving beyond grammar correction into intelligent agents, enterprise workflows, and real-time AI tools. We dive into Grammarly's new Authorship feature, why AI fluency is becoming essential at work, how Grammarly is integrating tools like Coda and Superhuman, and what the future of multi-agent systems looks like. If you're curious about where AI at work is really heading, this conversation will give you a clear and powerful glimpse. (00:00) Preview and Intro (03:37) Meet Luke Behnke (05:00) Grammarly's Origin Story and Early Vision (09:11) Grammarly's UX Advantage (13:30) Competing With Microsoft Copilot and Built-In Assistants (17:48) What Is "Authorship" and Why It Matters (20:31) AI Detection vs Authorship Tracking (25:05) The Future of AI Transparency (27:43) Why AI Fluency Will Be a Job Requirement (32:04) Grammarly's Agentic Vision (34:11) The Rise of Context-Aware Enterprise Agents (38:24) Use Cases: Automating Tasks Across Tools with AI (40:21) The Coda Acquisition & Building the Agent Platform (44:48) The Future of Interoperable AI Agents (47:43) Why Agent Oversight Is Crucial in Enterprise AI (55:57) Measuring Grammarly's ROI in the Enterprise

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  We've achieved the success that we have. Forty million people use Grammarly every day. We have been continuing to grow. And I think it's largely because what we've focused on most is how a human and an AI work together and how the AI can really support you. And we just believe AI detection is not going to be enough. Clearly, the answer is not no AI allowed. That is just silly. Institutions that are doing that are not setting their students up for success. What is better is for institutions to be able to say, this is what you can do. And then this is what you can't do for this assignment. Can you use an open-ended chatbot or not? And if so, what are the parameters that you can and can't do?

• 0:41Speaker 00

  And then we can produce a report that says exactly what the student did that both sides can verify. Really, we think the future of how human provenance in the AI era will work.

• 0:52Speaker 01

  Building multi-agent software is hard. Agent to agent and agent to tool communication is still the Wild West. How do you achieve accuracy and consistency in non-deterministic agentic apps? That's where agency comes in. A G N T C Y. The agency is an open source collective, building the Internet of Agents. And what's the Internet of Agents? It's a collaboration layer where AI agents can communicate, discover each other, and work across frameworks.

• 1:33Speaker 01

  For developers, this means standardized agent discovery tools, seamless protocols for inter-agent communication, and modular components to compose and scale multi-agent workflows. Build with other engineers who care about high-quality, multi-agent software.

• 1:56Craig S. Smith?

  Visit agntcy.org and add your support. That's A G N T C Y dot O R G. Hi, this week I talked to Luke Behnke, Vice President of Enterprise Product at Grammarly. I first heard of Grammarly at an AI conference six or seven years ago. And I remember thinking that that product will never fly because Microsoft Word has all of the same features, or so I thought. Since then, I've I've seen the company grow, much to my surprise.

• 2:36Craig S. Smith?

  And finally, talking to Luke, I realized that it does much more than than what I initially envisioned. I've since gotten a Grammarly account and I'm a writer and I'm frankly surprised at how many times Grammarly catches things in my writing that that need improvement. Since our conversation, Grammarly has announced its intent to acquire Superhuman, the AI-native email app, and that takes it a step toward becoming a full-fledged AI productivity suite. Combined with their recent acquisition of Coda, they're knitting together an AI-powered

• 3:20Craig S. Smith?

  workflow across documents, meetings, email. So stick around to hear Luke's vision for making Grammarly a collaborative workspace supercharged by AI. I hope you enjoy the conversation as much as I did.

• 3:37Speaker 00

  So yeah, I'm Luke Behnke. Thanks for having me on today. I lead our enterprise product team at Grammarly. We've been moving the Grammarly product that people know and love now for 15 years into the enterprise over the last few years. We've had a lot of success, you know, selling our Advanced AI capabilities for customizing your style, rules and brand tones and the way that you communicate with customers, helping English as additional language, sort of customers that are working directly, you know, in maybe a customer support use case, helping them all communicate more effectively. And we've had a lot of success moving Grammarly into the enterprise. And I've been lucky enough for the last 14 months or so here at Grammarly

• 4:22Speaker 00

  to lead that effort from the product side. And now we're in the midst of another big set of changes, which we're going to talk about today. That has been really fun to be at the forefront of. Prior to that, I have led product at a number of different companies, particularly a long stretch at Zendesk in the customer support space. And we were, you know, using previous gen AI at that time, previous generation AI, kind of machine learning NLP agents that, you know, maybe if we were lucky back then, they were solving, you know, 10% of customer support issues. And it's just amazing what's happening in that space. And, you know, Grammarly is a part of that, given that we do sell to a lot of customer support teams to help them better communicate with customers. So

• 5:07Speaker 00

  it's been kind of fun to see how all of this tech has evolved over the last few years.

• 5:12Craig S. Smith?

  Yeah. Can you sort of back up and give us a bit of the evolution? I first became aware of Grammarly at an AI conference, must have been in 2017 or 18. And I was surprised. I hadn't heard of the company. I know that you've been around for a bit already at that point. At that point, my understanding it was really targeting students or individual users. And I remember, and I'm sure you've heard this, I said this to you when we spoke earlier, that my initial reaction was.

• 5:55Craig S. Smith?

  You know, doesn't autocomplete and you know, all those things built into Microsoft would already do that. Uh, and here we are, uh, you know, 10 years or nearly 10 years later and you guys are a massive company and um, I, I'd love to know how, how you got started, what the initial vision was. Uh, as you say, you were using AI from the beginning. Uh, and how the product is, is evolved and then we can talk about how it moved into enterprise and what enterprises are using it for.

• 6:35Craig S. Smith?

  And then of course now, uh, your support of, uh, agents. Uh, but, but can you sort of give us a little history of Grammarly?

• 6:48Speaker 00

  Definitely. Yes, we were founded in 2009 and our, our goal, uh, really from day one was helping people build confidence when communicating, when particularly when writing. Um, you know, most people don't actually think of themselves as great writers. We have a lot of data on this. They don't feel confident, um, you know, that they're saying the right things. They don't want to look silly in a, a critical email or, you know, in a, in a, before they hand in homework at school or, you know, before they have a if they're having a conversation with a customer or something like, it's just, um, it's a real challenge and it remains even a challenge in the age of L M's, which we can, which we can talk about, but obviously 2009, long before

• 7:31Speaker 00

  generative AI. Uh, and so we were in many ways, probably a lot of people's first touch with like an AI. Now we call it kind of an agent. I mean, it does it. You install it in your browser or on your desktop. It shows up in every text box where you're working, um, where you give us access. Uh, and it's just helping. It's just providing. A consistent experience. You are right that there are a lot of spell checkers built into OSs or built into different walled gardens. You know, Microsoft has their own system, Google has their own system, et cetera. But Grammarly kind of creates this very familiar, easy to use, really beloved experience that isn't too pushy, but comes in at the right times, helps you correct

• 8:14Speaker 00

  your words. And now, of course, with Gen AI, it helps you generate words, helps you make improvements, like, you know, make it shorter, make it sound more like me, make it friendlier, all these sorts of improvements that you can do. And it just shows up while you're working. You don't have to go to a new destination to write. It's contextually aware of what you're doing. And we've, we really feel like we've kind of, we've achieved the success that we have. Forty million people use Grammarly every day. We have been continuing to grow despite, you know, all the questions you were asking, Craig, of like, is there, is there really a need for

• 8:54Speaker 00

  this in this crowded space? We've been continuing to grow. And I think it's largely because what we focused on most is how a human and an AI work together and how the AI can really support you. You don't have to learn anything to use Grammarly. You don't have to learn how to be a prompt engineer. You don't have to install complex sort of like, you know, customizations or anything. You just sort of start using it. It starts to learn what kinds of mistakes you often make and can help you correct those. It can help learn your tone and the way you speak.

• 9:36Speaker 00

  And it's just this very personal, you know, assistant that helps you write. And that's really what we're known for today. We have, I can talk about our move into the enterprise. You know, mostly we've been for those 15 years since 2009 focused on serving individuals. And you are right that students was a big initial use case, helping students write, you know, papers and work on their homework. Currently, our business is actually

• 10:10Speaker 00

  increasingly, and in fact, now more professionals using our product than students. But both continue to be really, really important, you know, users of the product and personas that on the product team that we build for. And then increasingly since

• 10:32Speaker 00

  2019 or so, when we launched our first Grammarly for Business Grammarly Teams offering. And now, like I said, we serve a pretty good percentage of the Fortune 500 and have had a chance to really, you know, actually launch our product offering at scale into enterprises. And that's been a fun journey, which I'm happy to talk more about, but I'll pause there.

• 10:59Craig S. Smith?

  Yeah, yeah, I'd like to know on moving into the enterprise. Do you sell seats or do you do, is this an enterprise-wide application, I mean, subscription? How do you do that?

• 11:16Speaker 00

  Yeah, we start with seats. We're often, you know, purchased by the marketing department who wants to write better content and, you know, superpower their their group or by the customer support team. You know, who maybe English is not their first language and they are communicating in real time with, you know, their name of the game there is speed and they don't have time often to correct their text via a chat, an L M chatbot or something. So having Grammarly show up right in line and and kind of ensure that they're able to respond to cases, tickets or chats in in real time is a big use case.

• 12:01Speaker 00

  And then also I T often binds us where they see that employees are Employees are bringing Grammarly to work. They they love Grammarly. They use it. They've used it sometimes since they were a student and now they've brought it to work. And so and then IT comes in and says, you know, oh, let's see, we got to decide what this company's doing. We got to see what they're up to and make a decision on whether we want to provide licenses of Grammarly for a subset or in in increasing cases for, you know, giving all employees the chance to to access Grammarly. And of course, they get advanced security controls and and, you know,

• 12:43Speaker 00

  ability to really ensure that it's configured in the way that they want. They can even encrypt any data that we do store

• 12:51Speaker 00

  with with their own key now. Of course, we never train on enterprise data ever. That's just becoming the norm. Hopefully, that's what every vendor is doing these days. And so, you know, they have an opportunity to really understand what we're doing and and control the settings that they want. And so, yeah, we've seen success kind of across those different motions, mostly a seat based seat based model today. Though, of course, as we've had a chance to sell more across an organization, you know, we do. We do have like enterprise license agreement models that become much more flexible as we do larger deployments.

• 13:30Craig S. Smith?

  Yeah, and I guess the because the the natural question is, how do you compete with Copilot, for example, which is alive on the browser

• 13:45Craig S. Smith?

  or in Microsoft Word? I mean, first of all, I can see the value, particularly for non-native speakers of a language

• 14:00Craig S. Smith?

  that or people who are not

• 14:04Craig S. Smith?

  just

• 14:07Craig S. Smith?

  Cannot handle language well. For whatever reason, it's uh it's wonderful to be able to to send something in in absolutely correct language and have confidence in it.

• 14:24Craig S. Smith?

  But again, the the various uh assistants uh that are integrated into many platforms uh do something similar. Is it that Grammarly is is following the keystrokes so you're not uh leaving your sentence

• 14:46Craig S. Smith?

  to click on an assist? I mean, what what do you see as the reason Grammarly has been able to survive?

• 14:53Speaker 00

  Yeah, I think a couple of reasons. Um definitely the fact that we are working where you're working in any application. It's not just gonna work in one suite of tools. You know, our goal is not to just um sell you our our core uh productivity uh tools that we make and you know, AI is an add-on to that. And then once you leave that uh tool, uh it doesn't work or it, you know, isn't supported, right? Our uh we're just there to get to know you as an individual user or um you know, organizations can customize us to their needs and we show up everywhere where you're working. And you know, we've even found that in in so-called sort of like Microsoft shops that we sell to. Um you know,

• 15:33Speaker 00

  we we uh have data on where, you know, which apps employees are using. And in in many cases, still sixty, seventy percent of the time, employees are not working within a certain walled garden, right? And so having a consistent experience everywhere uh you're writing um is is the biggest reason why um people use Grammarly. They know that they're gonna get that everywhere. They're gonna get that support regardless of which apps um they are using. And so we've really managed to kind of carve out that neutral position on any website, any application where you've allowed us to run will be there to help your employees or your, you know, if in an educational institution will help students do work.

• 16:15Speaker 00

  So that's probably the biggest one. I also think, you know, there's something that's that's been true since day one for Grammarly, which, you know, L L M's obviously solve so many different interesting use cases and we're so excited about the future there. But this idea of just a consistent final check on the work that you're doing and have that again, consistent everywhere you're working, but that last sort of confidence that I don't have any of these like red underlines that I just made a silly mistake. And it goes even further than just basic spelling and punctuation. It really says like here's, you know, this, these sentences are a little awkward. Let's fix them or, you know, really gets into some advanced suggestions and kind of having that final check before

• 16:59Speaker 00

  you hit send, I think is still unique for us. There isn't really any provider who's doing that and doing it across every system. I mean, some of these tools have built in spell checkers, but you know, we think, like I said, we're a familiar experience. So the way that the AI and the human work together in Grammarly, I think is very unique. We've, we've managed to find this very fine line between popping up and telling you, hey, you should really fix this with, without being annoying. And, you know, that's kind of like the, this very unique, very Grammarly thing. And as, as we transition and look forward to what we are doing and can do with L L M's, that we don't want to lose that history because it is such an important part of our product. And I think it is why

• 17:44Speaker 00

  people really do fall in love with, with Grammarly.

• 17:48Craig S. Smith?

  Yeah. And we'll get to the agentic stuff in a minute, but I'm really interested right now in this feature you came out with. I I don't know if I think last month called author or authorship. Authorship, yep. Yeah, that that will scan a document and and give you a and I haven't used it. So I don't know if it's visually by color coding sentences or something with sentences. Were written by a human, which were written by AI, which were AI edited.

• 18:31Craig S. Smith?

  And and that's particularly relevant to me because, you know, at this point I work with. And I haven't tried grammally, I'm sorry to say, I I will, I've got a. You should try it, it's free, you can try it for free and see if you feel like. Yeah, and I've got it on my list. But but I do use a couple of different models. Chat or GPT four O is the one I use the most. And, you know, I'm back and forth all the time as I'm writing, doing research or.

• 19:10Craig S. Smith?

  Or sort of brainstorming with the model on language and on, you know, doing outlines. And on occasion, writing a draft. Or asking for help in smoothing out a sentence or creating a transition. It speeds up the writing process enormously. And now a lot of places are are sending these very draconian. Messages saying if you, you know, no AI allowed. And if you're caught using AI, you're out and that's for. Yeah.

• 19:54Craig S. Smith?

  But it's much more complex than that because, you know, it's at this point, it's.

• 20:01Craig S. Smith?

  It's part of your workflow. It's not like you're writing and telling a model to write something and then copying it and using it directly. I mean, certainly I understand the concern about that, but I don't know what parts of my stuff is AI written because it's it's just such a spaghetti process. So how does authorship work?

• 20:31Speaker 00

  Yeah, great, great question. So you're exactly right that, you know, as these LLM tools become more prevalent, it is becoming less clear what is mine, what was written with or by and then improved by me. And we have seen obviously a rise of a lot of these AI detection tools, which basically are trained on patterns that typically come out of an LLM tool and they can detect, you know, whether it's likely that this text was generated by AI. And we believe those tools are definitely useful. We offer an AI detection tool for students to pre-check their work before they turn it in.

• 21:17Speaker 00

  But to your point, this is very problematic for universities, particularly where students have been, you know, accused of cheating, even maybe perhaps been expelled from from the school over allegations of cheating. And there's a lot of gray area there on I didn't know what was allowed, what wasn't allowed. Actually, I didn't use AI and this was falsely flagged, which definitely can happen with AI detection. And this is serious business. People, you know, these students kind of lives and reputation are at stake. And it's also true increasingly in the workforce as

• 21:58Speaker 00

  AI detection is being used on resumes and cover letters coming in, for instance. So it's making a first round of decisions around, you know, a hiring situation, or if you're writing a report, a research report or something, you know, you want to be able to understand how much of this was just generated with AI versus generated by a human. And we just believe AI detection is not going to be enough. It's certainly one of many tools, but it's not going to be enough. And so it actually all started, there was a student named Marley, who was accused of, she used Grammarly to write her paper. She believed her school told her it was okay to use Grammarly, which most schools do.

• 22:40Speaker 00

  But then when she turned it in, she was accused of using AI. And she posted this TikTok about it, which went viral. And we saw it and it gave us an idea to really rethink this whole space. And the way we did is with our authorship capability, which allows a student, it can be pre-configured by their university or the student themselves can enable this feature. It's a little fingerprint icon that they can click. Starts tracking from the start of their document. Works now in Google Docs and in Microsoft Word. Starts tracking your, how you created that document. Did you use AI? Did you paste in from external sources? If you paste it in, it brings in some metadata.

• 23:24Speaker 00

  We understand it came from Wikipedia. We can help you actually properly cite it, automatically cite it for you so that you're being clear that you pulled in something from another source. Or if you copied and pasted it from an LLM, you know, ChatGPT, et cetera. We can see that, we can tag that this was generated by AI. You can also cite that. Increasingly, AI citations is becoming one, you know, in some cases, an acceptable form of using AI. But then as you change it and make it your own, we're actually watching you do that. And then it's not just inferring based on language patterns. It's really understanding what did this human type versus what did they use an AI tool for?

• 24:10Speaker 00

  And at the end, it produces a report. That says percent of text that was generated by a human that was copied and pasted that was generated with AI or improved, you know, with AI. And then how was that text that was generated by AI actually then enhanced by a human? After the fact, you can even watch a playback of how the student or the individual created that document. And then you can turn it in almost like the certificate alongside of the the piece of work that says this was really generated by me. We just came back from a large educational conference, ASU plus GSV a couple of weeks ago.

• 24:52Speaker 00

  This was sort of, we launched this generally available, we're really like leading the way here and it was really the bell of the ball there. Institutions really believe this is the future. And like I said, I actually think there's a lot of business use cases for this as well. And so we're really excited to kind of be leading the way on that. And I think it's it's just one of many examples where I think our thoughtful grammarly thoughtful approach to how humans and AI work together in this future. Clearly, the answer is not no AI allowed. That is just silly. Institutions that are doing that are not setting their students up for success in their world. And

• 25:35Speaker 00

  you know, what what is better is for institutions to be able to say, this is what you can do. And then this is what you can't do for this this assignment. You know, can you use an open ended chatbot or not? And if so, what are the parameters that you can and can't do? And then we can produce a report that says exactly what the student did, that the student, you know, can verify. Both sides can verify. And it's just really, we think the future of how human providence in the AI era will work.

• 26:05Craig S. Smith?

  Yeah, it's funny. For from my point of view, I think schools should teach

• 26:14Craig S. Smith?

  how to use L L Ms or other.

• 26:20Craig S. Smith?

  Other AI systems of because they're not going away and they're only going to become more integrated. And I also don't, I mean, I understand in education because he wants students to be doing the work. But in other settings, in an enterprise setting, I mean, as long as it the document is factual and and not plagiarized. Um, I don't understand why it would matter if if it's written by AI or not, unless. I mean, there's been a lot of research about AI generated content, filling up the internet and then

• 27:06Craig S. Smith?

  getting into the training data and and there being some, uh, you you know, some collapse to the mean uh in uh in generation because the training data is so uh so similar, but uh but in the enterprise, do you do you find the enterprises concerned about this as much as they might be in uh in education as again, as long as the output is factually correct and and not plagiarized?

• 27:43Speaker 00

  Yeah, I think the incentives are better aligned in the enterprise where, you know, whatever tool gets the job done. Productivity is the goal, right? And we're going to talk about agents and and and all of that. But there is a really critical thing that you said there, which actually I I loved this. We were at a we had a customer advisory board with a number of our biggest customers and we had uh so someone from a a a large uh higher ed institution who said that they actually get a lot of uh businesses, you know, that they do a lot of their feeder school for many of the big employers in the area. And they have increasingly had those employers say the most important thing that they want when they're hiring uh someone uh out of this uh institute, out of this university is who are your best users of AI?

• 28:28Speaker 00

  Um, and he was laughing and he said, we call those cheaters, uh, like, you know, this is who you want to hire. Our job is to produce. People who are AI fluent, theoretically AI fluent, so they can get these jobs. And yet, we're actually labeling a lot of these people as cheaters. And I think this is a huge, huge problem and one where I think Grammarly has a really critical role to play in, in bridging that gap from, from students to, to professionals. But then, once you're in the professional landscape, you're, you know, increasingly being performance reviewed on your use of AI, right? Toby from Shopify and a number of other companies have followed suit on this is going to be a critical part of how you're evaluated as an employee.

• 29:10Speaker 00

  You know, this is, you're going to be managing agents, you're going to be, you know, using this technology day in and day out. And it's going to be expected of you that you are fluent and we are not doing a good enough job training people, even within organizations. I think we have a bigger burden to think about this like a learning and development, you know, situation. And, and to put the same amount of care that we would into onboarding new employees or teaching them how to use certain systems or helping them understand our security and privacy policies, you know, helping them become AI fluent, especially in this transition period that we're in is, is just so important.

• 29:51Speaker 00

  And I think Grammarly has a really important role to play, again, with that thoughtful AI human interaction, how we can bring some of this AI to the masses. Instead of it being this hard to use tool that you, you know, we're in this like early stage where we've all been given all of these raw materials, really, you know, with AI. And some people are emerging as super users, some people are afraid and, and self-admit that they're just not touching this stuff. And they don't, you know, really know what it means for them. And I think Grammarly can do this great just because of this thoughtful way that we've always approached bringing AI to a wide group of, of people. I think we can be an interesting bridge there in, in this kind of exciting moment in, in AI.

• 30:37Speaker 01

  Building multi-agent software is hard. Agent to agent and agent to tool communication is still the Wild West. How do you achieve accuracy and consistency in non-deterministic agentic apps? That's where AGENCY comes in. A-G-N-T-C-Y. The AGENCY is an open source collective, building the Internet of Agents. And what's the Internet of Agents? It's a collaboration layer where AI agents can communicate, discover each other, and work across frameworks.

• 31:19Speaker 01

  For developers, this means standardized agent discovery tools, seamless protocols for inter-agent communication, and modular components to compose and scale multi-agent workflows. Build with other engineers who care about high-quality multi-agent software. Visit agntcy.org and add your support. That's A-G-N-T-C-Y dot O-R-G. Tell me how agents are being integrated in the Grammarly, or agent management, or agent creation, or what are you guys doing there?

• 32:04Speaker 00

  Yeah, so I think it really starts with this understanding that while Grammarly has been known as this kind of ubiquitous AI proofreader, as I was talking about, something that we've built behind the scenes that just works when you download Grammarly, either as a browser extension or on your Mac or Windows PC.

• 32:24Speaker 00

  Like, we've built this incredible, what we call the AI superhighway. You know, Grammarly, our proofreader agent, we think of as like the OG, the original agent, AI agent for many people. And we had to do a lot to make that agent work in every single situation. You know, the way that Google Docs renders text is very different from the way that Microsoft renders text. And for Grammarly to work seamlessly across those two surfaces, to make it work in Excel and Google Sheets, to make it work in almost every text box on the web. Every desktop application, we put a lot of effort into this AI delivery mechanism essentially

• 33:04Speaker 00

  that brings these really thoughtful, helpful AI helpers right to your fingertips while you're working. It can, I know, understand what's happening on your screen. It can understand you personally. It can understand the kinds of things you're trying to do and it brings it right there. And so we call that kind of our superhighway, this ability to have someone configure something in Grammarly and then have it go right to their employees' fingertips as they're working. And we kind of laugh that we've been driving one car down that highway so far. This is our little spell checker, proofreader agent who comes down and tells you, "Hey, you know, you should fix this sentence. It doesn't make sense," or "this word spelled wrong."

• 33:43Speaker 00

  And we actually believe that there's an opportunity for us to have a lot more agents drive on that superhighway. Those agents will be configurable by the end user or by the organization. I can give you some examples of what we're thinking about for these agents. But to really allow the end user to bring in more of these agents, they're powered by the knowledge they've given it access to. Of course, knowledge becomes such an important part, just as it does with any human doing work. You have to understand, you know, what's the knowledge that this agent has access to? What are the skills that this agent can do? And then have I been clear about the instructions for this agent?

• 34:27Speaker 00

  And so when we really put together kind of knowledge, skills, instructions, and then bring those agents directly into where employees are working, I think there's some really interesting opportunities, you know, to... One, for instance, that we've been thinking of is, let's say, you have your CRM system of choice. Something that we do often is...

• 34:51Speaker 00

  You know, you want to understand, I'm about to go meet with a customer. What's, what's the latest, what's happening here? And we have this, this really cool tech that we've built at Grammarly that we call Knowledge Share, which actually highlights a, an acronym on your screen and defines that acronym for your organization. Historically, you had to set that up. You had to give us a list of all your acronyms. You had to give us definitions of all those acronyms and keep those up to date. But with knowledge, you know, and ingestions and integrations and with L L Ms, we can actually be much smarter about saying, you know, when you hover over a customer name, you're hovering over Acme Co. that you're just reading in a document. You're like, oh, are they a customer? You can hover over it. And let's say you use Salesforce. We built you a little Salesforce agent.

• 35:36Speaker 00

  It can pop up automatically right where you're working and say Salesforce, or sorry, Acme Co. is a customer. Here's all the information. There's a current open deal, you know, an open opportunity. Here's the current spend. Here's the current contacts. Here's the latest communication. You can just bring that context directly, you know, to your employees. Of course, it could go much further than that in terms of the ability to actually orchestrate all of the things we do every day around managing our meetings and agendas, managing our calendars and setting up time with other people, you know, taking notes, writing documents, sharing documents. You know, obviously there's a million things that we do where we're, we're tab switching, we're copy and pasting, we're bringing things from one to another.

• 36:19Speaker 00

  We're trying to get all of these different systems, the 400 plus apps that an average I T organization rolls out today. We're trying to get these things to all work together. But with Grammarly just being right at your fingertips, being able to bring this knowledge, be able to interact with these agents directly where you're working, ask them to go do something, check in on their work, see what they're doing and come back to you. Again, right where you're working. We just think it's a really interesting future. And again, everybody's talking about agents these days. We don't think we're the only ones, but it's a, we think it's a really unique take on agents. And what I love about it is kind of bringing this concept down to earth for a lot of employees who are like, when I'm supposed to be doing agents? Like, what, what does that even mean?

• 37:04Speaker 00

  But, you know, we, we really think of this agent store where you'll be able to go and pick all the ones that are relevant to you. They'll be. Contextually aware, you're working in a CRM, you have specific agents that show up for you when you're in your CRM versus you're in your employee or, you know, applicant tracking system and you're writing a a set of feedback. Well, we can bring in a lot of context and do a bunch for you automatically around, you know, the interview process, for instance, and, you know, this kind of context, knowledge and, and skills that you can then really just build these, build or use out of the box, these really easy to use agents, we think is really interesting.

• 37:44Craig S. Smith?

  And will that be a separate platform? I mean, uh, uh, if, if, if, if you're writing in a document and, uh, you remember, oh yeah, I have to email Joe that reminder or the, this information, uh, can, would, is it available right there that, oh, you know, you, you block a couple of sentences and, and click, send this to Joe? Or do you build your agents on a side panel? And then, I mean, how does it work? Yeah.

• 38:24Speaker 00

  Yeah, yeah. The, the way we see it working is like that, that Grammarly, the, the proofreader would become one of many of these agents. And in that case, you, you might have your email system connected. Uh, you can ask questions of it like, hey, do I have a bunch of emails I need to respond to? And you can do that anywhere. You can start to type them. Or if you're working, like you said, in a document and you're like, oh, I really need to send this over to Joe. You can select it and say, you know, send this over to, to Joe, for instance. And here's, you know, I can actually compose the text and hit send. And then the agent goes off and actually communicates with your email system and sends it. So that's exactly the idea is that, um, you know, these, these little agents for different systems that you use can kind of be invoked from, from anywhere, just like today with Grammarly, you know, you can, you can bring up

• 39:08Speaker 00

  uh, kind of your writing assistant anywhere you're working. You'll just be able to do a lot more than just a writing assistant.

• 39:16Craig S. Smith?

  Yeah, uh, and, and again, um, you, you mentioned an agent store.

• 39:24Craig S. Smith?

  Is, is that on a different website or, or does Grammarly generate a panel on whatever application you're working in to give you a list of relevant agents or how, how do you, uh, do that?

• 39:42Speaker 00

  Yeah, but we're working on it all as we speak, so it's still a bit in, in flight. But yeah, the idea would be that you'd just be able to, um, you know, sort of, um, click a little button that says add some agents here and go and search for them. Um, I'm sure we'll also have, uh, you know, a, a website, a store where you can actually go and install as well. Uh, but the idea would be that you'd be able to discover these things directly where you are working. Um, and you wouldn't really have to leave the experience that you're in to kind of discover what might be relevant for you. And then of course, there's a whole nother take on it, which is that IT can also configure, you know, what their employees have access to and who has access to what. And, you know, we, we, actually a lot of this platform is being built

• 40:25Speaker 00

  um, off of our acquisition of a company called Coda, uh, which we actually completed in, in January of this year. And Coda has built 800 integrations, uh, to other systems. Those integrations have very clear skills about what they can do, what they can read and write in those other systems. Their permissions can be set up by individuals or can be managed kind of globally by IT. And so they really built this incredible foundation for, um, you know, Coda's core product is a, is a document, kind of team hub and document product. And they built this amazing integration platform for how you bring data into documents. Um, but that same technology is what we are using uh, to build out this, uh, this agent platform based on

• 41:10Speaker 00

  these 800 plus, uh, knowledge integrations.

• 41:13Craig S. Smith?

  I see. Is this, um, are you talking about stuff that is going to be launched? How much of this is available to enterprise today? And how much of that is going to be available to the individual Grammarly user?

• 41:30Speaker 00

  Yeah, I am. I am definitely talking about stuff that's coming in the in the near future. And you know, today we do have, we have our proofreader agent, we have AI detection agent that you can invoke. We have a plagiarism detection agent. So you know, those are examples. We have the ability for you to configure what we call like a writing expert, somebody who's really good at the thing that you're doing, who comes in and gives you more advanced feedback on what you're doing. So we have the shoots of what I'm talking about here, and I'm not talking about totally theoretical things today. Individuals and enterprises can kind of bring in these, these additional capabilities beyond proofreading and generating text. And so we really have the, the, the foundation there, but bringing in these 800

• 42:15Speaker 00

  knowledge integrations and creating the agent store is very much what we're working on as we speak. So you're getting a little sneak sneak preview. We've been going out to talk about this actually a lot. Shashir, our CEO has been talking about this a lot in a number of of places. Just this is, this is where we're going and we're, we're pretty excited about it. And like I said, we have, we have the, the basics and it will be available, you know, both for individual users who, who want to connect their calendar and their Gmail and, you know, other systems and be able to use these tools together as well as available for enterprises to set up centrally, you know, enable kind of enterprise search, results augmented generation across their data.

• 42:58Speaker 00

  And then, and then invoke these individual agents that they've built or that have been built in, you know, in a, in a store.

• 43:06Craig S. Smith?

  Yeah. And the age, I mean, it's interesting. I just got back from an IBM event and they're talking about their agent platform. And then I had a call yesterday with a startup talking about its agent platform.

• 43:23Craig S. Smith?

  You, and I'm, and probably it will be the case that in a year or two years, we're all going to be managing, you know, a dozen agents individually. And then in enterprise, there'll be, who knows, thousands of agents.

• 43:42Craig S. Smith?

  You you use the term wild garden. And there are these ecosystems within different companies. But for example, with Grammarly's agents, will they be able to speak to agents in other ecosystems outside of Grammarly? And and how do you see this just conceptually developing?

• 44:11Speaker 00

  Yeah.

• 44:12Craig S. Smith?

  That everyone's going to have all these agents, they'll be able to talk to other agents from other ecosystems. I mean, is a few years ago, I talked to somebody, not a few years ago, more within last year. At Microsoft, they use the term society of agents that we're going to be living alongside the society of agents. How do you see that developing as each company builds up this agent workforce around their core product?

• 44:48Speaker 00

  Yeah, actually, that's funny. I listened to that society of agents podcast of yours, actually, just the other day. So that was super fascinating to to call it that way or to call it that society. I think there's some interesting things that that evokes that I that I agree with. But yeah, I think a couple of things are becoming true. You know, number one, it is clear of agents.

• 45:12Speaker 00

  As if, you know, I think that the command line interface for AI was in a is and was kind of an amazing first view of what these truly like magical tools could do. I think we all remember the first time we, you know, in November 2022, we went to a text box and we started asking questions and the things that came out were were pretty amazing. And that was an amazing way to kind of get this technology into people's hands. But what I but I think we all see is that agents being.

• 45:45Speaker 00

  More practical use cases for AI that are more contextual, that don't require to be prompted. You know, like I said, they have knowledge, they have skills, they have instructions. And they can then kind of go off and do a bunch of things instead of being, you know, here I am. I'm waiting for you to ask me to do something and then you better give me, you know, sort of more and more instructions as I, as I go. Clearly, we're getting to a place where we're seeing that agents are the future of the way that we're going to interact with a lot of these AI models. I don't think I'm saying anything too new there, but what I will say, I don't believe there'll be one winner take all in agents, just as there wasn't one winner take all in mobile or in SaaS or in cloud.

• 46:27Speaker 00

  This is, there will be different agents for different jobs, no doubt. Grammarly's place, we really feel, is these very practical productivity use cases and writing assistance use cases. But that, you know, doesn't mean we're going to get into the every single agentic use case in the world. Maybe, yeah, I guess never say never, but I think there'll be a lot of these systems. And so we've been watching closely what's happening, you know, with MCP, Model Context Protocol and Google's A2A, Agent-to-Age Protocols. There's a lot emerging there. It's very early days. You have to set up these servers.

• 47:08Speaker 00

  There, you have to provide it a URL that sort of has godlike access to whatever data's on the other side of that there. We're very much in the early days there. I think Grammarly could provide either an interface to those protocols or maybe an improvement if we can help you get authentication and permissions more clearly defined between all of these systems. Perhaps we could be even an improvement over some of these protocols. We may also be jumping on some of these protocols as they emerge. It's very exciting just to sort of see what's happening there. But I, but I think, you know, to answer your question, I think you are right that these Grammarly agents will be able to talk to other systems. They'll be able to interact in certain ways.

• 47:53Speaker 00

  You know, eventually, I definitely believe we will be kind of managing many of these independent agents. It's going to be very, very important. I think something that is not being talked about enough is Um. How you inspect what these agents are really going to be doing? Um. You know, just like an employee, you check in regularly. You understand, are you blocked? Did something happen? Did you make a mistake? Maybe you went in the wrong direction. Uh. And these agents, they'll be able to work very fast. And so you really have to be uh careful of what they could go off and do, right? Hallucination is still a very real problem, even though it's gotten much better over the last few years. Uh. And and just it's hard for these things to take a lot of instructions all at once uh and take actions.

• 48:39Speaker 00

  And um. So it'll be very interesting how, and I think this is another place where Grammarly in the future could could play a sort of. How are you able to see what's really happening across the various things that you're doing? Is anybody stuck? Did something go wrong? Has something gone in the wrong direction? Do you need to intervene? Um. And so that that'll be another kind of interesting thing to watch is uh as as we get to more of these workflows sort of happening behind the scenes. How are you still? You're still in control of this AI. It is still something that we humans have built, uh, that we have uh created a set of at least initial instructions for.

• 49:15Speaker 00

  Uh. And therefore, it's really is important that we're able to kind of understand uh what these AIs are doing and to truly manage them like like you would a human on on your team. And I think Grammarly has a unique role to play, seeing across a bunch of the places you're working and a bunch of the tools you're working in. Uh. And to help give you that layer uh as well.

• 49:37Craig S. Smith?

  Do you work with uh companies like Microsoft? Uh, who have their own um agent, uh, you know, co-pilot built into their products? Uh, and and do you see uh there there being kind of a convergence among a lot of these?

• 50:02Craig S. Smith?

  Companies that are building agent agentic platforms. And and that there'll be.

• 50:11Craig S. Smith?

  Sort of specializations that, oh, you go to Grammarly for this, you go to so and so for that. Or or do you think it'll be?

• 50:21Craig S. Smith?

  Just what people are comfortable with. I mean, it's a massive market.

• 50:27Craig S. Smith?

  Horizontal market. And there's plenty of room of for a lot of different players. So, you know, I grew up with Grammarly, I'm I'm gonna use Grammarly, I grew up with. Whatever it is, Microsoft, Copilot, because we are talking about the next generation. I mean, I'm. I'm kind of an obsolete. User at this point. So how how do you see that developing?

• 50:57Speaker 00

  Yeah. You know, I think our. Stance on being this kind of, we don't have any.

• 51:06Speaker 00

  Other motivation other than making employees or students extremely productive and capable in their day to day. We have no affiliation to any one, you know, platform. I think. Indicates to me that we should be partnering with Microsoft or Google to find places where. We we can offer our capabilities on their platform as well. And bring, you know, if it's the best way to get these solutions directly into the hands of a of a user. Like, as you said, clearly, there are, there may be organizations who decide to go all in on on one ecosystem.

• 51:50Speaker 00

  But I actually think that that hasn't been how technology has played out. That's definitely not how SaaS has played out. And I think. AI will look a lot like SaaS in terms of the. You know, there was a time where it was like you will have. One software provider for all of these sort of suite of productivity tools. And now I think we all struggle with it, to be honest. We use so many different tools, even within our, within our Grammarly organization. We use so many different, you know, document surfaces. We use so many different, um, you know, calendaring tools. Uh, you know, we, we have, we've had to take a little bit of, uh, of control of it, but yet it's just how it works. People choose the best tool for the job.

• 52:34Speaker 00

  SaaS and kind of the, the software revolution of the last 15 years has really made kind of best tool for the job. Uh, will, will win and will become part of a stack and interoperability is extremely important. And we think that's something really unique for, for us where we have an advantage. And so from that perspective, you know, even though we, I guess ostensibly compete in many deals with Copilot. And we often sit right alongside and they see different use cases for different tools. Um, and so, you know, being involved in these other ecosystems is likely also, you know, in our, in our future. I would love to get to a place where we have, um, this really incredible extensive platform of AI agents and apps.

• 53:18Speaker 00

  Where we could see an organization going all in on, on the Grammarly platform for how they run a lot of their productivity tools. And we may get there. I, I, I'm excited about that day and that vision. Uh, but certainly for now, I think really being a good citizen of everywhere our customers are working is our goal. And so from that perspective, partnering, you know, may, may very well make sense as well.

• 53:42Craig S. Smith?

  Yeah. Uh, and when do you see the, the, the full agentic, um, product line, uh, hitting the market? Is this something that's weeks away or months away?

• 53:57Speaker 00

  Yeah. Um, like I said, I think we have a number of these additional use cases, agents beyond proofreading today that are already live, um, with Grammarly. And I think, uh, over the next weeks and months, you'll see more and more of this. It may result in kind of a larger launch moment, uh, in sometime this year. Um, but we're, we're really just. Trying to iterate as quickly as possible. Speed is the name of the game in this, at this time, as you know, it seems like every day you wake up and there's some new, you know, AI paradigm shift out there. And so for us, we're not, we're not waiting. We're really just going to move forward on incrementally releasing a lot of this capability to customers.

• 54:41Speaker 00

  One thing we've done, 15 years of serving a pretty wide group of consumers is we've gotten very good at testing experimentation and learning. And I think it's such an advantage to our enterprise business that we can do a lot of that learning with our self-service kind of individual customers. You can see what, and of course, these are just people that work in organizations or that, you know, students that go to a school. So they're still the same end users, but they're generally willing to sort of try, adopt, learn. You don't have to go train 1,000, 10,000 people on how to use the, it's just one person where you can go and get a lot of feedback across a pretty large group.

• 55:23Speaker 00

  And then we can take that innovation and what we've learned directly to the enterprise. And I, it's something that I've loved. Everywhere I've worked, I've been fortunate that we've always built for the end user, you know, the employee first. Zendesk was like that in customer service. It was all about making the agents who's sitting there 10 hours, eight hours, 10 hours a day in Zendesk, how do we make their life amazing? And then, of course, we have to care about the buyer, the administrator, make sure that it meets all of their needs as well. But if you start with that end user, you know, we've, we've done a lot of proving the value

• 56:06Speaker 00

  of Grammarly to our customers. Today, we spent a lot of work over the last year quantifying the ROI, including actually allowing some employees to use Grammarly, some to install a version of Grammarly that isn't actually helping you in any way, but it's working behind the scenes. And then being able to actually show the employer, look, your customer satisfaction scores for the agents using Grammarly, the, the customer support agents using Grammarly was actually much higher than for those not using Grammarly. We've had to. Really, you know, produce these ROI metrics, but it starts with building these amazing experiences that the employees love, they want to use.

• 56:50Speaker 00

  And then it becomes infectious within the organization. More, you know, they say, please don't ever take this tool away from me. You know, we have seen that where they say, oh, maybe we'll consolidate on one of these tools. And employees say, I can't, I can't give up Grammarly. And so they end up, you know, using us alongside other tools. And so that kind of focus on the end user, delivering value to the end user, making sure that what we have is is really working. It's such an advantage for us because we're able to to really see how how individuals are using it and then bring those innovations directly into into the enterprise.