From Product Fit to Market Leadership: Gil Makleff on the Journey of Sembly AI

All right, welcome to Hack to the Future with Kyle Roof. I am Kyle Roof and today we have Gil Makleff, who is the CEO of Sembly AI. Hey, Gil, thanks for being here.

Hey Kyle, thanks for inviting me.

Awesome. Now, Gil, Sembly as an AI tool has been around for about five years or even maybe a little bit more than five years. And I really feel that tools like Sembly really hit the public consciousness, I think, in the last year, especially. And for you being around as long as you have, is that a feeling of validation, or is it more of a feeling of frustration, kind of like, get on our level? How does that feel?

I love the question, but so it feels great now to be counting revenue MRR ARR. We're very happy about that. I'm very happy about that because it feels like it's not a validation, but we spent a lot of time on product market fit. It was a so we've been monetizing for about 18 months now. So it took a while to get to the product market.

Yes.

But I think, chat GPT on January 23 was a milestone, yes. The people started moving towards the platform and yeah, the amount of interest today in tools of this type is phenomenal and we're very excited about it. But it allowed, listen, the time allowed us to percolate. What does percolate? We have really deep capabilities in different types of technology, attendance technology, and various others, we built our LLM. And so obviously I like to use an analogy of we use all LLMs, we're surfing on the waves of innovation in the market and we use everything and we integrate it into our environment. And we like to think that we produce the absolute best result in the market place today.

That leads kind of into another question that I was thinking about. When you've got an AI product, I think there's a balance between building a product that has a lot of depth and also an agile one, which can change as technology quickly changes. How do you kind of strike that balance? So you've got something that's impactful now, but is also long-lasting, is a long-lasting solution.

It's a it's a great question for us. We're very comfortable today because we have a platform that can scale. We're running tens of thousands of meetings, so I think over 60-70,000 meetings a month right now. And we can scale to 700000 a month very easily because we've built that depth. We took several years to get the capabilities that we have today. And I think it's critical to do that. I can compare that to, a UI wrapped around an LLM and say, great, I mean, move forward, try and capture market share, but know that you have this whole environment to build, which includes pre-processing, post-processing, cleaning up data, moving information in the right way, and providing value to your customers consistently. And that word consistently is a big deal. So.

Yeah, I imagine you know like as soon as people as soon as it's buggy or it's not fulfilling then it doesn't matter what you add on top of that, huh?

Yeah. Exactly.

When you're building out an AI team for this product, I'm kind of wondering, how you find or identify the right talent? Because I feel like these types of companies are so new. And I don't necessarily mean like the developers or that side of it, but like the other people, like the marketing team or the sales team or the product team even, are those people, are there any traits that you find or things, or is it just finding the talent as you would in just about any other industry?

The world is changing. Peter Thiel this week identified the change and said that people who did well in the past were typically more math-oriented and people in the future will be more communicative-oriented and I think that's generally true, but I truly believe in the old model as well, which means a merit-based approach which is if somebody is capable then they're capable across the board and they need to fit into a team approach. I believe that the corporation, even if it's virtual, 100 % virtual, needs to be team-oriented and people have to believe that they're doing something important for the world and important for the customers of the company. And that's a key point that's never going to change. So I think in AI as in everything else, it's going to be married, it's going to be a process of creativity. Creativity is more and more important because I think the boundaries have changed and there is unlimited creativity going on. Now that's where our background comes in. You know, I have 30 years of management consulting and technology background for companies in the US. We're applying that in our solutions towards, you know, AI into the environments that we know so well. So for us, if you know the difference between functions, and workflows or functions and processes. Workflows and processes related to human beings. Functions relate to the capability of a technology. So we're focused totally on workflows and that's an important distinction as well. So, people who can understand workflows and be creative enough to change, to help change workflows based on new capabilities, that's an amazing combination.

I've always felt that workflows need to be, or SOPs if you will, in a way that kind of a well-trained monkey could probably do a pretty good job. But you need to have somebody that could then identify those situations. And that's where I think the creativity you're talking about comes in, where we need to change this. You don't want people reinventing the wheel necessarily as they're doing their job, but you want them to be able to be creative within it.

Yeah, we are in the process. We're building a catalog of workflows. So not two workflows, but over 40,000 workflows. And some of it's AI-generated, some of it's user-based. So our users give us insights in terms of what workflows they prefer or how they orient those workflows. I can give a few examples, but we can talk about it later.

I think the user. No, I think user feedback is so critical for any company in the space to continuously listen to the feedback. And if a customer comes to you and says, you know, we need this, some companies think, it's a nuisance, another request. I can't believe it. But we treat that as a golden nugget where we listen and we apply it because we know that we can generalize it to other customers as well, in most cases.

Love that. I love that a lot. I want to slightly different direction. Leaning on your time at UMT Consulting and also when you're at Ernst & Young, when in this field with these AI companies when you're thinking about acquisitions and mergers for AI companies, would you think now is the time to buy, sell, or hold?

I think being careful is important. And I referred before to UI wrapped around an LLM as a light solution. I think that there are a few important components. One is the fit with the existing company, and how that works. Number two is how the new acquisition is going to help you reach your vision and whether will it happen more quickly with the acquisition. I know this seems trivial, but it's so critical. And three, the cultural synergies between the company. If the culture is so different, it'll cause so many issues that nobody will want to deal with. It'll be distracting and people will leave and it'll be very difficult. So M.N.A.S. are a delicate thing and depending on the specific size group and the type of the group, there are a lot of considerations. So I would say careful consideration as usual and with AI and more.

You mentioned a kind of LLM that just has a user interface wrapped around it. Would you say that IP, like if the company has its IP rather than just a use case, would be an important thing?

Of course, if you're able to identify a unique capability, today in technology it's much more difficult than in the past. So there's a way to differentiate your company by having this focus on quality, delivery, customer use cases, and consistency that helps you grow the company that you're focused on. If you have a unique IP, the trick is to grow the company to a point where you're maximizing the value of the company. That's kind of the two dimensions. So yes, definitely a unique IP. But don't rush to sell if you're truly unique and you can keep that moat in place, don't rush to sell, grow the company, and then sell when it's most appealing to you.

So now we're going to play the game Surviving the Apocalypse. After the singularity happens and AI becomes sentient, what is one skill you have that our robot overlords will find useful enough so that they keep you alive?

I love the question. I've been thinking about it so much. So when we reach singularity, Neuralink will have developed a chip that will allow us to compete in several domains. And overutilized. So we have, I don't know if you know this, but we have about 100 trillion neurons in our brain. And, you know, so, our chip will enable us to use those in a much better way than we do today. And we'll be able to integrate AI into our capabilities. So that's my answer for the future. And it's a, it's a sliver of light in a lot of dystopian thinking. But I think that Sliver could develop into something awesome. So that's my, that's my.

So we actually won't have to worry about the robot overlords in the sense we will be the robot overlords.

I believe we will live in tandem with the very capable AIs, yes.

Let's talk a little bit about Sembly AI. I've had a lot of fun clicking around through the website. For people that aren't familiar, can you give us the 30-second elevator on it?

Of course, so Sembly AI is meeting driven flawless execution, and flawless execution comes in three ways. I'm going to take 45 seconds, not 30.

That's fine. We'll allow it. We'll allow it.

It gives efficiencies. Efficiencies because it's a great note-taker. It does, on average, better than humans do. It's not influenced, et cetera. The second dimension is proxy attendance. It's a big deal. It means that you can send your agent to a meeting. You don't have to attend. But you can rely on your agent to take great notes. And you can review the meeting in minutes to the hour afterward thing that we do well is effectiveness, which means we get your data, think of a fabric in the organization, it'll get your data from a meeting magically into your target systems, Microsoft systems, Notion, Slack, and any systems that you work with will get the data into there in an organized manner. And the third dimension is we're giving you next-step actions in your meeting, do not so think about this, most people have one summary of a meeting for everybody. No, we're changing the game. We're revolutionizing. We're changing. We're delivering something awesome very soon, which will allow Kyle to have his notes and next steps from a meeting. So if there are eight participants in a meeting, each one, based on their role, based on the meeting type, based on the industry, will be provided with a next recommendation recommended next-step actions, which will make you very powerful. So that's what we do. We're enabling working teams to produce better results, which will not impact not only the bottom-line inefficiencies but also the top line because people will have more time to focus on product delivery to their customers.

So when you talk about the next steps part of it, what's a use case there? Would it kind of tell me these are the things, the tasks that you need to create? Can it create tasks for me in something? How does that go?

Of course, it depends on the integrations that you have, but if you have an integration with a tool like Monday or a tool like Microsoft Planner or any other tool that you use, it will depend on the type of integration. If it's a native integration or other integration, it will have the capability to automatically set up a template after the meeting and give you a head start on what needs to be done. But also we're very strong believer in human and AI synergies. So think of the world, you talked about the future dystopian future maybe or not a dystopian future, but think of the world in three phases. The current phase, the intermediate phase, and the future phase. We believe that in the intermediate phase, we need to focus on synergies between machines and humans and emphasizing those synergies to produce great results for companies. So we're going to help working teams produce better results. That's the idea.

Do you say that it's a product that is mostly team-based or could an individual use it as well?

So, of course, individuals could use it, but the powerful capabilities are for working teams. The real powerful capabilities are for corporate working teams in companies. They don't have to be big companies. There can be small companies or very large companies, but the teams will benefit tremendously from Sembly AI.

And I don't want to say something incorrect, but I felt like the price was unbelievable. It was something like plants are like $10, $20 a month, or something like that.

That's right. And we have a great pricing approach, we'd like to provide the best pricing possible. We have even better deals than on our website for large implementations. And we have a philosophy of win-win. We like to win with our customers. We'd like to create an environment that's as good as possible for our customers. That's right. It's unlimited meetings.

I usually don't talk about the price of these types of things, but it shocked me at how inexpensive I think I might buy it. This isn't a paid sponsorship either. Gil did not pay to be here. This was just one of those things, you know, as you're going and looking through the stuff, I really couldn't believe the capabilities at the press point. I think I'm sold. I'm sold.

Please do. I love it, I love it, please do. It's so, Kyle, it's so easy. Yeah, because there's no downloads, there's nothing to do. You add an email and register. That's it.

Yeah, I was impressed. I'm not impressed by a lot of tools. I was actually, I was quite impressed. When speaking of teams, as you said, the power is in the teams. What's a stumbling block that you've seen for companies not implementing tools like yours? What's something that holds, is it like fear? Is it just like unfamiliarity? What might stop them from implementing it?

Yeah, so it depends on the size of the companies. So some companies go onto the Sembly AI website and register and don't talk to us even. There are smaller teams, 10, 20 people will join and would like to do it by themselves. And it's not a problem. We prefer if companies call us, we have a sales approach, and we have a customer success approach, which is how to build a center of excellence how-to, we can educate people, we do webinars, we shorten the cycle for education, typically. So, you know, at times it's, the problem, I'll just share with you a key concept that I think is important. Technology adoption is instantaneous today. So if we come out with a new feature, et cetera, theoretically, it will take 10 seconds for somebody in Europe or Asia or anywhere else in the world to utilize that technology because it exists. Human adaptability is not that quick.It takes people time. They know their current workflow and most people hate change. So how do you adjust between the changes? And so what we do is we adhere to workflows in companies, but we provide suggestions on how to make modifications. And when people come up to speed, I can do that. It happens to us a lot. This thing, I didn't know you could do that. Then the modifications are gradual and it's, it's, more palatable to team members and working teams, they can actually accept the change, understand the value, and move forward to the next level. So it's gradual change versus abrupt, difficult change. That's what we try to do. And I think change, is probably the most difficult part of adopting AI technology because you can rethink the way you do things. And it takes time. But I think this idea of adoption versus adaptability of humans is a big gap today. And will continue to be because you see people in the world, I mean, not in, you know, not near us, but that haven't used chatGPT yet, you know, they exist. So.

Yeah, very good. Final question. Where do you see Sembly growing and becoming something else other than what it is right now as it kind of would expand? Or would you see that you're going to stay focused kind of in where you're at now and develop these features that you currently have in more depth?

No, we're moving. We're shifting, we're moving. Every day is a change for us in terms of learning. We're a learning organization. I think Peter Senji coined that term a long time ago at MIT. We are a learning organization, but we are also building capabilities based on our background. So we know corporate America. We know what the corporation needs, and we're continuously building capabilities not only for the bottom line but also for the top line of the company. And that's what's exciting to us about the future. And that's where we want to because the competitive advantage is going to be obvious when companies start implementing, this technology. It's a clear competitive advantage. And we are the source of unstructured information, right? Think about it. We it's like, it's a gateway into the, applications into the company. So we're a gateway. Knowing where to move data and also how to provide insights is a critical thing for firms going forward. Companies that will do it more quickly will gain a competitive advantage in the marketplace. So I'm excited about that part.

That's awesome. Well, Gil, thanks so much for being here. I appreciate your time. This has been fantastic.

Thank you, thank you, Kyle.

All right, and thanks to everyone else for listening to Hack to the Future. We'll talk to you next time. Bye.