Redefining Meetings and the Agentic Future With Artem Koren
- Quik! News Team
- 9 hours ago
- 28 min read
Artem Koren is the Co-Founder and CPO at Sembly AI, an AI-powered meeting intelligence platform that helps teams capture, organize, and act on insights from conversations and meetings. He leads the company’s product vision, building AI tools that improve collaboration, productivity, and decision-making for modern organizations. With a background in systems engineering, enterprise consulting, and product leadership, Artem specializes in creating human-centered AI solutions for the workplace. Since co-founding Sembly AI in 2019, he has helped pioneer innovations in meeting intelligence, AI teammates, and agentic workflows.
Here’s a glimpse of what you’ll learn:
[2:24] Artem Koren discusses Sembly AI’s mission to simplify work and reduce information overload
[3:53] How early AI experimentation led to the creation of Sembly AI
[7:14] The internal resistance companies faced adopting AI-first thinking
[9:30] Why some technologists struggled to transition into the AI era
[12:02] Artem talks about AI agents and agentic workflows
[18:16] Reliability, scalability, and the future maturity of large language models
[22:02] Why AI agents are scalable from simple tasks to running entire businesses
[26:02] How AI agents will transform software applications and APIs
In this episode…
Businesses today are overwhelmed by meetings, information overload, and rapidly evolving AI technology. Many leaders struggle to understand how AI agents, automation, and large language models can improve productivity rather than add complexity. As AI transforms the way people work, how can organizations adapt quickly without losing the human element?
Artem Koren, an expert in AI-powered productivity and meeting intelligence, explains how businesses can simplify workflows by using AI to capture, organize, and activate knowledge from conversations. He shares how organizations can embrace AI-first thinking, adopt agentic workflows, and redesign software experiences around collaboration between humans and AI agents. Artem also discusses the importance of building reliable AI systems, helping teams transition into new ways of working, and preparing for a future where agents communicate and solve problems autonomously.
In this episode of The Customer Wins, Richard Walker interviews Artem Koren, Co-Founder and CPO at Sembly AI, about the rise of AI agents and the future of work. Artem discusses overcoming resistance to AI adoption, differences between frontier and open-source AI models, and how agent-to-agent collaboration will reshape enterprise software and productivity.
Resources Mentioned in this episode
Quotable Moments:
“We live in a very stressed and overly information rich world today.”
“I think centrally we make people's lives easier. That's how we help people.”
“It’s no longer functional. It's LLM based, which is a very different way of thinking.”
“An interface is a human API, right? Like it’s just a way for humans to interact.”
“Agents will talk to each other. So that completely uproots the world of APIs.”
Action Steps:
Embrace AI-first thinking within your organization: Teams that actively explore AI capabilities are more likely to uncover productivity gains and innovative workflows, because employees must shift from rigid, rules-based systems to adaptive AI-driven processes.
Build systems that simplify information overload: Reducing noise and organizing knowledge helps employees focus on meaningful work instead of repetitive tasks, because AI tools should improve quality of life by making work easier and more manageable.
Prepare teams for agentic workflows and AI collaboration: Businesses that understand how AI agents make decisions and complete multistep tasks will adapt faster to workplace changes, because agents will increasingly collaborate with each other and reshape how software operates.
Design software experiences around natural interaction: Allowing users to communicate with applications conversationally creates more intuitive and efficient workflows, because future applications will support both humans and AI agents through chat- and voice-driven experiences.
Focus on reliability before scaling AI adoption: Strong infrastructure and dependable AI outputs are essential for building trust in enterprise AI systems because stability, response quality, and scalable architecture are foundational for long-term AI implementation.
Sponsor for this episode...
This is brought to you by Quik!
At Quik!, we provide forms automation and management solutions for companies seeking to maximize their potential productivity.
Using our FormXtract API, you can submit your completed forms and get clean, context-rich data that is 99.9% accurate.
Our vision is to become the leading forms automation company by making paperwork the easiest part of every transaction.
Meanwhile, our mission is to help the top firms in the financial industry raise their bottom line by streamlining the customer experience with automated, convenient solutions.
Go to www.quickforms.com to learn more, or contact us with questions at support@quikforms.com.
Powered by Rise25 Podcast Production Company
Episode Transcript:
Intro: 00:02
Welcome to The Customer Wins podcast, where business leaders discuss their secrets and techniques for helping their customers succeed and in turn, grow their business.
Richard Walker: 00:16
Hi, I'm Rich Walker, the host of The Customer Wins, where I talk to business leaders about how they help their customers win and how their focus on customer experience leads to growth. Some of my past guests have included Liz Fritz of F2 Strategy, Dr.Kara Hartl of Troy Medical and Adrian Johnstone of Practifi. Today I'm speaking with Artem Koren, co-founder and chief Product officer of Sembly AI, and today's episode is brought to you by Quik!, the leader in enterprise forms processing. When your business relies upon processing forms, don't waste your team's valuable time manually reviewing the forms. Instead, get Quik! using Quik!.
You'll be able to generate completed forms and get back clean, context rich data that reduces manual reviews to only one out of 1000 submissions. Visit quickforms.com to get started. Now, before I introduce today's guest, I want to give a big thank you to Dr. Jeremy Weiss of Rise25 and host of the Inspired Insider Podcast. Go check out his website at rise25.com as they specialize in helping B2B companies launch and operate podcasts. In fact, they helped me launch this podcast.
All right. I'm really excited to talk to today's guest. Artem Koren is co-founder and chief product officer of Sembly AI, where he is building meeting intelligence technology for professional services teams. Since co-founding Sembly in 2019, Artemis helped shape AI tools that capture, organize and activate critical information for meetings at scale. His work sits at the intersection of AI, enterprise productivity and the future of knowledge work with a focus on augmenting human potential rather than simply automating tasks.
Welcome to The Customer Wins.
Artem Koren: 02:01
Thank you. Great to be here.
Richard Walker: 02:03
Oh man. I'm excited to talk to you. And we'll talk about that in a minute. So for those who haven't heard my podcast, I just love to talk to leaders about what they're doing to help their customers win, how they built and deliver a great customer experience and the challenges of growing their own company. So let's understand your business a lot better.
How does your company help people?
Artem Koren: 02:24
If I were to summarize all the ways that we help people, I think centrally we make people's lives easier. That's how we help people. And we can get into all the details of all the different ways that that works. But I think the crux of the product is to simplify a person's life at work. We live in a very stressed and overly information rich world today. I think anything we can do to simplify and reduce noise is a big win for everyone, and also an improvement in quality of life, both generally and work wise.
So I think that's the crux of the Sembly product.
Richard Walker: 03:10
Oh, I love it man. It's about the easy button. Let's make this easier. My company name is actually Efficient Technology Inc.. Quik! is what we go by in our brand.
But to me, it's always been about how I create efficiency so I don't repeat the job over and over again. And look, one of the reasons I'm excited to talk to you, and I told you this when we first met, was your product was the very first note taking, AI-driven product I used in 2022. Like, honestly, you're one of the introductions to AI for my experience, and I've taken it so much further in terms of my education and learning about it. But you spark something and you were early to this game, so you started your company in 2019. Can you talk about how you saw the vision for where this was going to go and how this has been evolving?
Artem Koren: 03:53
Sure. So when I got together with my co-founder in 2019, we both were very interested in what AI could do. And we specifically thought about the space of meetings and the fact that what happens inside of meetings never gets carried over into any kind of technology. There's a lot of products out there to support actually having a meeting like the Zooms of the world. But what happens in the meeting was always a black box, and we thought that if technology could understand what that conversation was about, there could be great gains to be had.
And I dabbled with AI very early around 2010, I had another startup where we tried to make tech to use AI to automatically recognize cancer and digitize biopsies, and we were a little bit early to the game at the time, but that was really my first inauguration into what I would call the modern era of AI, which is kind of neural net based. And so way back then, I already had a good kind of sense of what AI means from a technological standpoint. And then fast forward, you know, nine, ten years, the space had evolved certainly not yet to where it is today with the GPT and large language models, but enough where you can do a lot of really interesting things that weren't possible before. Yes, it was harder to do. Yes, it was less accurate.
All of that, but still very, very interesting. And so we set out to do that. The state of AI was very, very nascent at the time. And basically every step along the way, we essentially had to invent something new to make the use case workable. And we were even surprised about things like transcription technology that, you know, had a lot of time to evolve.
But yet, you know, even in the from the biggest providers like Google, Amazon. The quality of their transcription was very dodgy, and there were really 1 or 2 providers in the world that could give you a level of accuracy that was acceptable. And so we actually had to solve that problem. We actually built our own transcription engine as well. Then you had to do Diarization, which is who says what.
And then you actually had to extract the analytics from the meeting. And that was all each kind of analytic that you wanted to do was its own invention, right? To get tasks out of a meeting that was its own project. So yeah, it was, it was a, it was a lot of interesting, exciting innovation work. And then 2022, 23 came and everything got flipped upside down.
Richard Walker: 06:36
Well, yeah, I was going to say, you can take a large language model today and start building this type of technology, but you had to invent it, right? And gosh, this is how I felt. I started my company in 2002. There was no Stripe or authorize.net or any way to do payments. It was so, so hard to do.
There was no web or vertical response or SurveyMonkey. There was nothing to send emails and do all this type. We had to invent all that stuff along the way. So I can only imagine what you're going through. But let's talk about that world being upside down.
Did you suddenly go, oh my gosh, we have to switch our models and change things around? Or did it just because a proliferation of tools happened or what?
Artem Koren: 07:14
I mean, pretty much the reality is that we, you know, we saw it coming early. But interestingly, as you saw, the world took a while to turn ships into this new direction, this new kind of technology and user experiences that LMS brings. It was a few years before everyone kind of normalized into it. And today that, you know, that's an accepted truth. But there were a few years where people were like viewing it and saying, oh, but it can't do this and it can't do that, and it hallucinates and that and that.
So it took a while. But the funny thing is, even in our company at Sembly, even though we're an AI frontier company, we had a microcosm of that. We had a good number of our technologists who are very skeptical, like hyper skeptical. And we're not we're not on the ship like they were, you know, they just, they just they were hardcore engineers. They're used to thinking in terms of functional programming, right?
Like input X, output Y, and logic in between. And they couldn't get themselves out of that paradigm. They were in that paradigm so deep that it was very difficult for them to embrace this new way of thinking about what tech can do. And so even internally, it took a while for me to kind of push everyone over that kind of cliff. You know, it's not, it's no longer functional.
It's it's it's, it's about MLM-based, which is a very different way of thinking about implementing experiences.
Richard Walker: 08:53
I have no idea how glad I am to hear this. I'm not alone. I mean, we were not an AI first company to begin with, but we have said we are going to be AI first and what we do, how we think, how we work, what tools we select, what we build. And my gosh, I mean, for six months I was showing people what is possible and nobody would do it. So I was fighting that.
And, and I say it's still a challenge to get every single person on the team really absorbing what's possible and trying to implement it in such a way. Did you find yourself I don't want to bring up any kind of bad feelings, but did you find yourself having to transition people around or out who just couldn't get it?
Artem Koren: 09:30
Yeah, I did. But you know what? It was never adversarial or negative. There were people we had who were just very specialized in a certain way of working. Like that was what they were very expert in.
And it and the door was open for them to step into this different kind of world. But for them, it meant it had to, they had to really jettison a big part of their expertise and switch into a completely different world. It's like someone who, you know, has grown up most of their life coding Sembly, and now you're asking them to do Java. You know, it's like it's for the non techs, maybe that's not as good as analogy, but it's essentially it's a difference between like you're playing with memory address spaces and like specific CPU instructions. And that's all you're doing versus you're talking about business objects, doing methods on business actions.
It's a completely different paradigm. It all happens to be tech. It all happens to be programming, but it's very different. Another maybe a more accessible analogy there is like, if you're a doctor and let's say like you spend your whole life like doing, I don't know, like very hands on, like knee surgery, but now there is like a robot that does it, whatever your like nuanced, you know, physical touch that you've developed over decades is now irrelevant. All of that is done by the robot.
Your job is completely different now. Like your job is to make sure the robot does a good job. That's a completely different set of skills. And so we had people who stepped into the new environment and were very successful there and thrived. And then we had some people who just weren't willing or weren't interested.
And those people, those people had to move on to a direction that they're more comfortable with.
Richard Walker: 11:30
Yeah, it's a hard thing, especially as a leader. It's really hard. You see this and you don't know how to really help people make the choice or not. And it's up to them. It's really up to them.
Look, let's transition to another part of this topic, which is the agentic world. There's this promise that agents can replace the people. I don't fully believe that. So I'm kind of curious. You know, your product has to evolve too.
So what are you guys doing with agents and agentic workflow? And like the world of, I guess, the agentic world that we're seeing pop up.
Artem Koren: 12:02
Yeah. So I guess, Rachel, it's a good time to reveal. I'm actually an AI agent. I'm not real. This is not a.
Richard Walker: 12:09
I saw you walk away from your desk before we started recording. I know this is no way.
Artem Koren: 12:13
My agent is that good. No. So I think so. First of all, I think we're still just as an industry. I think we're still trying to nail down terminologies and meanings because agent means so many different things in the world.
There used to be this concept of bots. Now there's this concept of agents and it means different things to different people. And the way, even the way you introduce it, like the agents can replace humans that already suggest to me, like a certain frame around what an agent means, right?
Richard Walker: 12:43
Yeah.
Artem Koren: 12:43
Right. So yeah, so I think, I think we're, we're, I think we're in this space where maybe the internet was in the early 2000, I would say like that and or maybe even like late 90s, but maybe getting into the early 2000. Here's what I mean. Like in the late 80s, early 90s, the foundational methods of the internet came alive. And so we're talking about the TCP IP protocol, we're talking about the HTTP protocol, we're talking about browser rendering and the Dom display object, and so these are the kernels, like the seeds, the kind of the fundamental concepts of internet applications and tech.
And then you fast forward maybe a decade to a decade and a half and you're starting to see like a higher level of objectification of these concepts that are now a lot more like, you don't have to code the right header protocol into your application so the browser can render your page like that's, you had to do that in the 90s. Like if you're writing with like Perl, CGI, you know, not a lot of people do prolapse now, but when you're writing Perl, you had to like hard code the correct HTTP header, and that's part of the output of the application to the to the browser and to the HTTP. Today, you no longer have to do that. That's all abstracted away, and I think so the kind of the core pieces came online in the last few years. And part of that is alums and a few other things.
And I think now we're starting to get into that phase of object objectification, for lack of a better term or abstraction. And agents are the first foray of that abstraction. And there are very powerful abstractions. And so yeah, go ahead. Sure.
Richard Walker: 14:37
Because to follow on your analogy, and I know some of my audience is not following all the detailed tech that you're spouting. And I love it, by the way, because I'm like you, but, but I think what you're, what you're trying to get at is like, if you're building a house, the foundation goes down first and then you can build the, the, the, the walls that go up from that. And then you can add accoutrements to that and, and things like that. So are we now at the foundation layer? Is that what you're really getting at?
Like we have laid the cement and agents are more like building up the sticks of the wall.
Artem Koren: 15:09
I think the foundation is nearly solid. So. I the, the house, the house analogy is good, but I think it's I think a better analogy is like, you know, before we could build houses effectively, like we had to learn the right mix, the right, the right formula for cement. We had to kind of write. Yeah, exactly.
Like there, you know, you had to kind of standardize how bricks are and like how all the structural pieces work. But today, if you want to put windows in your house, you're not going to like to buy a piece of glass and buy a frame and like, try to, you're buying the whole window product with, with, with that, with the way that it's attached, the way that it insulates. Like it's a whole product, it's a multi-piece product. It has all these parameters that come with it, like sound insulation, whatever. But you're buying the whole window product, right?
You're not like trying to stick frames together on glass. And I think that's where we're at. So we're at a point where the cement formulation is getting to be pretty good. It's not perfect yet, but it's nearly stable. And by cement, I mean things like when you try to like send a request to an LLM, like there's a pretty good chance you're going to get back a response and there's a pretty good chance that response is going to be fast enough, and it's going to be like kind of what you expect.
Now it sounds like, well, isn't that just very simple? You would think so, but it turns out no. And that's why you have companies out there like Open Router and many others that kind of facilitate that, because it's the same thing with the internet. Like when, when the internet first came up, like your, your network was dead half the time because packets are getting lost. Routers are blowing up like things that can't reach each other.
But now we have these like self-healing routings and like all these fancy schmancy things that we just take for granted, like our mobile phone works. Like for anyone who had a mobile phone 20 years ago, you know, like you're like doing this a lot, right? Like where's my network and blah, right? We're not doing that that much anymore. You know, I live in New York City, at least here.
And I think that's the point, right? Because you need that level of stability and robustness to then build applications that really users can use at scale. Like it's not okay if half the time the thing just doesn't work or takes too long, it gives you garbage back. So I think we're reaching that level of robustness now. And now something like agents can wrap that under them and give you a very complete end to end experience using AI tech underneath.
Richard Walker: 17:40
So as you're saying this, I'm imagining part of what you're getting at is the reliability of a large language model. Anthropic's been down so many times in the past week. It's kind of silly to me, but I get it like the demand for their service is so high they can't even buy the computer power. They can't buy the hardware fast enough. They can't install it fast enough.
You can't build data centers overnight. So this has to grow and build up the stability of their model. But do you also think it's the maturity of the models? Do you think we're going to see ten times better outcomes of models still? Or are we kind of getting to that point of stability that the model is serving us well?
Artem Koren: 18:16
So the models will get better, smarter, faster, and they'll continue to do that as long as there's economical value in that. So my prediction in the model space is that I think open models are about a year to a year and a half behind frontiers. So for the audience, when we say frontiers, we mean like an anthropic cloud, we mean like OpenAI, GPT, we mean Gemini from Google. Those kinds of models. And they're running on monstrosity hardware.
That's like billions of dollars. And that's, you know, they're having a hard time growing that hardware. But that's the kind of hardware that these models require and they produce today. They're the most competitive and they're across different benchmarks producing the best results. But there's also a whole community, a huge space of models that are not frontier, that are called open.
And these are in some sense like independent model makers. And these are models like GLM from Xai or Quen, or Deep SEG that most people have heard of. I think a lot of them are Chinese. It just so happens. But there's non-Chinese open models as well.
These models are very good, but they're like year to year and a half. Behind is the kind of the overall quality of the frontiers. And as we move forward, the open models will get better. Differentiable models will get better. But at some point you're going to hit a point where there's a certain level of quality that's good enough for most kinds of applications.
Richard Walker: 19:55
Yeah, that's what I'm thinking.
Artem Koren: 19:57
Right. And then and then, you know, enterprise or large business is going to be okay to pay a premium to get the biggest, smartest, fastest from the frontiers. It's similar to Microsoft Windows versus Linux, like Red hat and things like that. To some extent, it's not a perfect analogy, but to some extent on the server side, it's really true, I think. Or Microsoft.
Apple maybe. Right. Like going back to the history of that, I think we're in a similar kind of dynamic where at some point the frontier models will really be primarily useful because of how much they cost big, big businesses and they'll give them competitive advantage. But I'm also sure that those frontiers will start offering lower tiers that, compared to today's standards, are super strong, but we'll compete well against other open models. So I think it's not a point of like, can it get smarter?
It can. And I'm sure some more specialized players will make them as smart as they can possibly be on their way to AGI. But for business purposes, there's, there's kind of a point of saturation where you don't need like, maybe you don't need GPT ten, maybe you're like, GPT nine works amazing. And, it doesn't really make mistakes.
Richard Walker: 21:18
I think this is where architecture comes into play too, because you know what I've been building privately? The architecture I've built. It works better on Sonnet Claude, sonnet, which is a lower model than opus. Opus doesn't work as well. So you can architect your solutions to actually be optimized for a model type or a model setting, if you will.
And I imagine it getting better and better, but that doesn't necessarily mean my product gets better and better because of the way I've architected the product. So going back to your analogy of building the foundation, let's now talk about the definition of agent again, because you're right, it's a spectrum. And I've, I've laughed about this because there's so many different variables that go into this. So I really want to hear your kind of prescription of what an agent is.
Artem Koren: 22:02
So the way and again, there is no universal definition, but I can tell you how I look at it. So an agent is a, is a kind of application that's AI, AI-enabled or AI powered that. Take multiple steps to accomplish some goal. And in those steps, it can make decisions. And once you.
I think once you have those factors together. So it's AI-enabled multiple steps to make decisions. I think you got yourself an agent. And in those steps and in those decisions, it can use tools to help it make decisions. It can use external information.
It can use internal information. So it can do things in each of those steps to make a decision. And then the decision could be to take the next step to do something else. Sometimes. And a lot of modern agents work this way.
The decision could be to do multiple things at once as a next step, like launch multiple subagents and get back information. So I think as long as you have that it's AI powered, multi-step, and makes independent decisions in those steps for itself to reach some kind of goal. Now you have yourself an agent and this construct is very scalable. It can do everything from generating a particular kind of document. Like you have an agent that specialized in that.
So it's like maybe a two, three, four, whatever step agent that, okay, like it's very good at generating statements of work. And so for example, like you ask them for statement work and it already knows, okay, I need to pull like this kind of information together. I know what the sections in the statement of work are. I'm going to fill them out. If I don't have enough information, I'll have to make a decision about asking you to give me more information.
And then the next step could be to actually render the document and fill in the sections, all of that stuff, right? But so it takes a few steps and it makes a document, but you can also have an agent that can build an application, or maybe you have an agent that can run a store, right? Or maybe you have an agent that's a CEO of a company, right? Like that. That construct is an extremely scalable construct because in the sense and this is where we're getting super interesting.
In the most simplified way. It's not that different from what people do.
Richard Walker: 24:18
Right? Right. So I had Jeff Woods, who is an author of a book called AI Driven Leadership, and he talked about it as this a job is a set of skills applied to a process, pure and simple. So jobs can change, different skills are needed, different processes are needed. And so you don't have to feel like you're losing your job.
You can take your skills and apply it to a different process. You can up your skills, etc. but when we talk about agents, that's how I think of it as well. What's the job? The job is to read this document and give me some insight from it. Read this document and extract invoice number from it, you know, whatever it is, whatever.
But you have to distill it down. So I love your definition of agent, by the way. It's probably the best I've heard, and I think of it as infinitely small or infinitely big. And like you said, scalable. But I think the job can be really tiny?
Yeah, it could be really tiny. It could do one simple thing that requires discipline.
Artem Koren: 25:12
You can have an atomic agent. Maybe it needs to make one decision along the way, and that one decision is important because if you don't have a decision, then you have a chatbot. And that's. Yeah. So this is where we can draw a line.
That line is blurring a little bit, but classically, ChatGPT versus agents, right? Because ChatGPT, you ask a question, you get back an answer. It's not really multi-step. Okay. People are going to argue with me.
Now. There's thinking there's research now. Yes. Okay. ChatGPT is evolving, and it's kind of a little bit agentic today, to be honest, but it's a very simplified kind of agentic.
But yeah, you could be very atomic in its agentic work.
Richard Walker: 25:52
All right. So how has this formed your view now of your product and how you innovate and keep moving forward? What are you doing with Sembly now with all of this information unfolding?
Artem Koren: 26:02
Well, with this new abstraction, with this new concept of agent. This is really going to have huge reverberations across all tech. All I would say is user facing tech that we have out there. It's I, you know, my, my opinion is that it's, it's really going to reformat all applications. And there's, there's many, there's many reasons for this, but one, one, one.
So one thing to think about is that applicant applications up to recently, like, let's say Microsoft, right? Like Microsoft applications, they were made for people to use, like they were, you know, like you had to learn to use the app and then you use the app and you get your PowerPoint done, or you get your Excel done, whatever it is that you're trying to do. But in the future, all applications will need to also or only support agents because there's really now an extra layer of indirection between you and the technology, and there's no reason why an agent couldn't learn to use an app really well and do the thing for you. And if you're going to have a mix of people and agents in your company, it makes sense to have applications that support those agents. So that's one big change is that all applications in the future, I think will be agentic applications, either in terms of being able to be used by agents or have agency built into themselves.
And then the second bit, and that kind of works in tandem with the first, is that agents will talk to each other. So that completely uproots the world of APIs. Yeah. So if you think about user interfaces as an API for humans, right? Like the, all the buttons that I click.
That's a human API. Like nobody tells you that, but that like, trust me.
Richard Walker: 28:07
That's what it is.
Artem Koren: 28:08
You heard it for the first time today. It's just like, you know, like this is it, is it a wave? Is it a, is it a particle? An interface is a human API, right? Like it's just a way for humans to very specifically interact with the screen to get something done.
It's the API that we can use because we're not computers and we don't send code to each other. But then there's an API for apps like Zapier is a great example of an aggregator of these APIs. There are many more that define very specific ways that an app can send a request to another app and get back a response. And that request response cycle is extremely well defined. So everything works.
And if you make a mistake like you send the wrong question, like the wrong field, your thing will flop, right? It'll say, what are you talking about? I don't understand you because computers are really good at being exact. But now you have agents who couldn't care less about being exact. You've talked to ChatGPT.
You make all the spelling mistakes. It still knows what you're talking about. So now agents can just talk to each other and play in text. And so if I have an agent, let's say there's an agent that I have that's really good at building case studies for your business, right? So historically you have a professional services company or any sales team, and you have a customer and they want a project.
And they're like, okay, like, you know, have you done things like this before? You say, yeah, we have. And let me send you some case studies. Okay. Those case studies are not going to be exactly about what this customer is asking.
They're going to be some templated case studies that you've generated, because case studies are hard and they take a long time to do, and they require a lot of effort. But what if, like you could send a bespoke case study each time you have a customer, like the customer says, you know, I want to build this thing. I want to build it in France. And you know, I wanted to take four months instead of the usual six. And this thing has to be left to right instead of right to left.
And, and you say, okay, like, here's a case study. And your case study will show them as close as possible to what they're trying to accomplish, that you have done something very similar to this before and that you can do the work. And so you can build a complete case study, rich media, beautiful images, visuals, your brand, their brand, whatever you want in a, in a, in a click, because there's an agent that knows how to do that. Okay. But in the process of building that case study, that agent is going to, at some point of decision making, it's going to say, okay, I need some customer quotes for this study, right?
Like I want to quote from a customer. So that agent should be able to talk to another agent that knows about your customer conversations. And it should say, hey, John, I don't know. Right. Whatever the agency is going to be.
I'm building a case study literally, literally in text. If you had a chat screen, you could see this and say, hey, John, I'm building this case study and it would be really great if you can give me quotes to reflect this and this from, from these kinds of customers. Do you have any quotes like that? And John goes, you know, thanks, Amy. One second.
Let me take a look. And then John, the agent, comes back and says, here's some quotes from our customers. You can include them in this case study. And Amy says, great. And she puts it in the case and delivers it to you.
So agent to agent collaboration is, is part, it's just part like, it's, it's, it's going to be table stakes. Yeah, very, very soon.
Richard Walker: 31:43
Yeah. And I, and I see it playing out in a couple of ways. Personally, I see it playing out in software that we build where the agents are talking to each other for various purposes. And when you see the conversation, that's magical. When you don't see the conversation, you think, Is it working?
Are they doing it because they don't have to show it to any interface at all? They can just have the conversation and when it is working, it's amazing, right? You get the outcomes. But there's another one. I mean, you brought up the idea that everything's gonna be driven by AI.
And I'm actually designing my first site to be driven by AI first meaning in the sense you can get to a dashboard, you could do administrative things, you can click the links, but you can also just tell the agent, hey, I want to add a user and it will do it for you. Hey, I want to start a project. It'll start the project for you. It'll navigate to the place in the site you need to be so you can just have a chat conversation with the site. In other words, and I think that'll easily turn into voice.
So you can just talk to it and say, hey, add my user, send out ten invitations to users, blah, blah, blah. I think that is going to be really prevalent, like personally because I'm designing it, but I think that's going to be a prevalent way that we see things, especially in the context when you need user input because you're asking the user to provide intelligence, information, insight, make decisions that the AI is not going to do for you. Man, we're running out of time and I want to geek out with you so much more. But I do have to wrap this up. So thank you for all this.
Before I get to my very last question, what is the best way for people to find and connect with you?
Artem Koren: 33:12
Sure. So personally, you can find me on LinkedIn at KOREN. Happy to connect there. Much better LinkedIn than email. Also, you can find our website, it's www.sembly.ai.
There's a free trial. You're welcome to register and try the product. No, no obligation and see if it works for you. We haven't talked a lot about what Sembly does, but it's partly agentic. It makes your life easier.
It really helps you in your work. We have thousands of organizations and teams that use the product. And I know Rich is a long time user as well. So yeah, so definitely give it a, give it a try, especially if you're in sales or you're in project delivery. I think this, this, a project can be a big help.
And I'll add that if you put in the coupon podcast 2026 all caps, you'll get an additional discount as well.
Richard Walker: 34:07
Nice. Good job, I love it. Yeah. I look for everybody listening. Sembly is the OG. They were here first doing this note taking stuff.
Maybe not absolutely first, but they were at the front lines. They've weathered the storms. They get it. So go check them out. All right.
Here's my last question. Artem, who has had the biggest impact on your leadership style and how you approach your role today?
Artem Koren: 34:30
I mean, a lot of impacts to my leadership style. But I think for this one, I want to take it all the way back. And it's something I, as someone I don't think about very often because of how early this was. So this was one of my very, very first bosses when I was still in high school. And when he hired me, he didn't realize I was in high school.
And so at the end of the summer, when I told him, okay, I have to go back to senior year in high school now, he was like, what are you talking about? So his name is Scott Lackey. And he was a tech manager. He was the manager of the development team. Where I was working was a company called alloy.
They had an e-commerce catalog for teens and things. And I think a lot of my technical team management style actually comes from him. And I'm very. And I realize that now in reflection, like I didn't realize that then, but I think he was a really great example of how to get specifically technology teams to perform really, really well and to engender a very high-performance culture on those teams. and a lot of his methods and methods I still use today.
Richard Walker: 35:52
Nice man. I love hearing about the old stories that really had this impact. So thank you for sharing that. All right, I gotta wrap this up. I want to give a big thank you to Artem Koren, co-founder and chief product officer of Sembly AI, for being on this episode of The Customer Wins.
Go check out Artem's website at sembly.ai. And don't forget to check out Quik! at quickforms.com, where we make processing forms easier. I hope you enjoyed this discussion. We'll click the like button, share this with someone, and subscribe to our channels for future episodes of The Customer Wins. Artem, thank you so much for joining me today.
Artem Koren: 36:26
Thank you for having me.
Outro: 36:28
Thanks for listening to The Customer Wins podcast. We'll see you again next time. And be sure to click subscribe to get future episodes.
Comments