The Briq Vision

Episode 2: The Briq Vision

From bots to Otto

How Briq went from building bots that patched gaps in legacy software to betting the company on Otto, and the four principles behind it: generative, orchestrated, autonomous, and conversation first.

The Bet · 22 min

Opens October 27, 2026 at 9:00 AM ET

Read the transcript

At Briq, we've been at it for almost a decade. Don't say it.

It's been almost a I'm gonna I'm also going to shock you with this one. You've been at it for almost three decades.

That's awful to say. Yeah, it's a long time. Time flies by. And in this episode, we're going to talk about that journey and specifically Briq's journey. You know, from day one, we've always, you know, gone after the big existential problems in construction really hard problems to solve.

Talk about what were those some of those early problems that we had with those early customers and what things were we trying to solve in those very early days. You know, it's so funny. When I started Briq, I used to have this pitch that I would say, you know, I'm done selling software to this industry. I just want to make the software you've already bought better. That was my thing. We were like the BASF commercial. We don't make the software. We just make them better.

And really the problem, the first client we had was an ENR top 100, top 50 actually, building a data center for one of the largest social media companies in the world. And they had one version of a PM system, the contractor did, and the social media company had another. And they wanted a bot to speak between those two systems. And they wanted a bot because they tried to use an API or tried to use an integration partner. And obviously it's not an integration problem, it's a process problem.

Yeah. So what gave birth to Briq was the idea that there was burgeoning technology seven, eight years ago that allowed bots, what we call bots at the time to log in, act like a human, navigate, extract or enter data. That was really the birth of Briq. Yeah.

After that, we realized it's hard to be a jack of all trade. So we started to focus on the office of the CFO. Yeah.

We started to focus on workflows that are highly customized by most companies. Again, making other people's software look better for forecasting and WIP. Yeah.

And that was a journey. That was our growth journey. That was how we kind of came to the future of work. If you watched the future of work in 2023, we'd be talking about bots a lot about how bots can operate your accounting system or your CRM system. And that's where we were. And that's not where we are today.

Well, a lot because a lot of those, you know, look, a lot of those workflows, a lot of that, you know, you mentioned broken processes. I mean, you think about the accounting function. Yeah.

There was so much kind of data keying and entry that takes place. It was just kind of a natural fit. That's where most of our, you know, our customers actually led us there. Yeah. The accounting folks wanted to keep on using Vista, you know, Sage, whatever accounting system they owned, and that they wanted to fill the gaps with little computer programs. And in some cases, so many of the processes are so broken and their systems are so archaic, you still had a high failure to launch rate on bots because, you know, even though you created a bot for Vista, the next client did something different different type of business, different line of business.

And so we were always this adjunct, this co-pilot, this thing hung off the side of some very old systems.

Yeah. And that was an interesting crossroads for us. We had to make a decision, you know, right when we started to see that bots were going to get easier to be made, that our competitive moat that the product category we were in was called robotic process automation. When we said, okay, you know what, the competitive moat around possibly one day this LLM thing that's coming out could probably code and could probably code bots. M so we got a little worried and I said well you know if people don't know or the idea that what we used to say the best Briq client doesn't know they're running Briq and that was that was our like motto internally and then what we started to realize well wait a minute if we are that if we're a bot company and nobody knows we're running and it's easy to replace bots eventually using LLMs were at risk.

Yeah. And so we had to make a very big decision and we did. It was a bet good to be lucky to be good but it was a bet on creating an orchestrated model ourselves.

Yes. The other thing I think too and I'll ask is you know was there a moment where you realized that bots had a limit? Yeah. What? And was there a specific kind of customer or use case or a moment where you said, "Okay, there's a limit to this." Oh, yeah. I mean, I won't say the client name. If they're watching, they always do. I'll tell the story. They'll know who they are. We were sitting in a room and we created all these bots and the bot generated a forecast in Excel.

Yeah.

And the bot did that by extracting data out of Sage, took the data, retransformed it, put it into Excel, put into an interactive spreadsheet. So forecasting could take place and then eventually so WIP could take place. And for those that are watching this outside the construction industry, WIP is the way we recognize revenue in construction.

So we're sitting in the room and the bots did great. They did it in, you know, 30 minutes. A human used to take eight hours. They built the spreadsheet pixel perfect. I look great. Everybody's happy. Early client. This is amazing. He goes, "Oh no, those numbers are wrong." So why are those numbers wrong? It's like all the contract value is wrong. All those numbers are wrong. Start dates wrong, end date's wrong, can't do S-curves off of this. I said, "What happened? Why?

Why are you know I got developers on the phone yelling at them? What? What did you screw up?" He goes, "No, no, no, no, no. Just all the data in Sage is wrong." I said, "What?" He goes, "Oh, yeah. Just a minor." He goes, "Oh, yeah, no, no. I don't refer to that data." Like he tells us after, you know, three months of development. Oh, I don't use that data. I use the data in my Excel spreadsheet. I said, "Oh." I was like, "Oh my god, I got to find a new business." No, the I was like, "All right, I'll find a new client." So, I was like, "Okay, you know, that's great.

That's fine." Some bots are amazing. We do today to this day, we do millions of bot transactions a week.

Absolutely. Yeah. Yeah. This is not an in an effort to underell the bot.

No, the bot company is a great company. It's internal to Briq. It's its own P&L. It does millions of transactions. Makes lots of people happy. Some of the largest clients on the planet or construction companies on the planet run our bots. We are a vertically specific RPA division within an AI company. RPA being robotic process automation. Works great. The bots are maintained, managed, and run through. And they do what they got to do, but they can't think.

Yes. And I think the other problem is that you know, when you're developing a technology that has to compensate for a less efficient technology, let's say like a legacy accounting system or a legacy PM system. Yeah.

The other big thing is if you do not own that domain of that legacy PR system you are at the whim of old technology and bot some of these bots, we call them digital workers don't work well in that domain in that.

Yeah. It's just you know you get and you know woe is me. I won't say that but you know you get blamed because Sage timed out when on the login because it's crappy old software and everybody knows it right. So, like you, you know, I use a very crass term. When you're a bot company, you're kind of the mistress.

Yeah. The wife is still, they still go home to Sage, right? They still go home to Procore, right? And we were kind of that like side piece.

Yeah. And that's what the history of Briq was. And when we started to realize, wait a minute, that same fear I had about bots being easier to be built, that our competitive mode of having this platform, I said, well, you know what? I bet you they could write a PM system. And we took all of our knowledge of where all the bodies were buried, where every bot filled a gap in Procore, Autodesk, Sage, every bot did. Every bot was filling a gap where humans were frustrated using this older software where they had to do all this extra work.

They're like, "We need to save time." And when you buy a bot, you don't think about which system am I turning off. You think about which tasks am I not going to do anymore?

Which means you're, you know, it's kind of a sad sales process. It's because they want to fire Jenny or Johnny first or the Jenny or Johnny's working 90 hours a week and they want to work 60. That's the difference.

And so all of those learnings, all of it, all of the holes of why OCR fails with drawings or AP and payments are broken or, you know, purchase orders don't have the right stuff. Yeah. And all of the learnings of how did we fill the hole?

Wow. We know where the problems are. Yeah we know what problems these people are trying to solve and we're going to talk about them in this series. But we started to build Otto to solve those problems and Otto started to solve those problems at scale.

Yeah. And it was amazing. Early days of Otto if we're Can we fast forward there now?

Yeah, let's do it. So early releases of Otto we were, you know, we were learning early releases of Otto look like the bot company.

And what is Otto? Talk about what Otto is. Yeah. And I think it's very that's thank you for that. Otto is not an LLM. And I think that's a little bit confusing for folks. Otto uses LLMs. It also uses probably a thousand different components. So think of Otto and applied AI as the use of AI technology in a package to produce an outcome.

So it sounds like a lot of words, but you know, a Lego piece is an LLM. You know, the Starship Enterprise Lego set is a model.

Yeah. And so that's what we are. We're we're kind of that we use the Lego pieces to apply AI to solve practical real world problems. And Otto is that combination.

There's so many different terms that people have used for what Otto is.

The best term that I've heard is applied AI. It takes the best of LLMs, image generation, data analysis, design analysis, bots. In some cases, we now call it computer use. So there's Otto Computer, there's Otto Translate. All of these components create Otto. And think of Otto as a platform for work. That's the key here. It just uses AI first. When you think of the platform for work you use today, it's database first. It's I have an Oracle database and I have an accounting system on top of it or a SQL database.

Otto looks at it in reverse. Otto is corpus first. What am I trying to solve first? That's what a that's how AI works. It's much more ambiguous because you don't have tables and rows necessarily.

We it eventually gets there, but the idea here is that Otto is a place you work instead of your ERP or your PM system.

Yeah. So, exactly. So you say, okay, so we have these digital workers, these bots that are trying to use this old software and it's there's all these gaps happening.

Yeah. We say, hey, why not instead of trying to break our necks building these bots that use this software, why don't we just reimagine the software?

Yeah. Reimagine reimagine the domain and the and you know, you can you can think of the bots as kind of like the system of action of the things that are doing the work and the systems of record of where that work is happening.

Yeah, I understand. So would you say that you know maybe Otto is the combination of the system of record and the system of action in one place?

Yeah. And this is a big statement. I think this will shock a lot of people.

This is big. Yeah. I think this will Yeah. Yeah. We're no longer saying let's go make your accounting system look better. Yeah. We're saying today and in the future our Genius application layer, the Genius layer sits on top of the Otto platform. Yeah. So the Genius layer is are apps built by Otto to replace existing software. So we are no longer saying we're going to be an AI agent or chatbot or bot on top of these systems. We're we're stating now that these systems are inefficient.

These systems are not serving you well and these in these systems are not structured to be AI first. Therefore, AI implementations of these systems are going to fail.

And we're, you know, we're saying it. I don't want to be the negative, but those projects that you have on top of your legacy software that was developed 20 years ago in whether it was, you know, client server or Ruby on Rails or, you know, some other technology that sticking an agent on that ain't going to work.

Yeah. That's the bet we're making. And I'm pretty good at making bets.

Yeah. Yeah. In this industry. Yeah. And we'll talk about some of the more details on applied AI, autonomous work, and agents. So you stated the vision there quite clearly. I mean, what is what does what's some of this early like we'll talk about some specific use cases, but like what does this look like in practice? Yeah. For companies?

Yeah. 100%. The here's the reality. You have a limited IT budget.

Yeah. I mean, let's just talk brass tacks. You have maybe, you know, we sell, we spend about 1.5% of our total revenue in it, in construction. It's one of the lowest spends in any industry.

Yeah. We have higher revenue topline, lower profit, but ultimately we do spend one and a half percent. M when you spend 1 and a half% and 90% of that 1.5% are spent on systems that a human need to administer or enter data into.

The word accounting literally means entering data into a system where things that happened elsewhere. That's literally I'm accounting for something that happened elsewhere.

Yeah. So, not only are you paying for the software, you're paying a lot of people to operate that software, good, bad, or indifferent, a lot of these people did not enter this industry to do data entry. And that's a huge portion of the cost. So, we know that you have a limited budget. Coming in and saying, "Hey, how much do you have for the agent budget?" That's a net new ad.

Yeah that's a net new ad. How do I, you know, somebody comes to us as real world example and we'll talk about lead to award, but you know, classic bot company story. I have leads from Dodge. I need them put into my CRM. The integration is weak. Doesn't work. Something's wrong with it. Whatever. And then we'll build we'll build a bot to bring it in. Oh, by the way, can you score the lead? Yeah, of course we can. Oh, can you look up the plans and specs?

Yes. Can you analyze the plans and specs? Can Otto? Yes, of course it can. Can it do a sample price proposal? Yes, of course it can. Can it build the proposal? Yeah, of course it can. Otto can do all these things. So then you pay $300,000 to Salesforce and you have $10,000 for that budget. All of that work that Otto did, reading the lead, extracting the data, looking at the plans and specs, tracking what the owner built into the past, does the owner self-perform?

Is this a is this a you know a bake off or are you getting are you really just being taken to the cleaners looking for a price? Do they have a preferred contractor? Where's that evidence? Does this work for you? Is this an ideal project? All of that analysis is not cheap.

No. And it is a lot of GPUs. It's a lot of energy. And you have $10,000 for that, but you have $300,000 spent on the empty filing cabinet that Joe or Jenny doesn't want to even use or ever keep up to date or frantically keeps up to date that one time a month when you print the report.

That's different. And it's really cheap for us to add the filing cabinet. And it's really expensive for the filing cabinet to add the intelligence layer. Yeah. And they didn't come from that.

Yeah. But I did come from the filing cabinet world. Yeah. So I think that's the big difference.

Yeah. No, I mean that's it's it's it's and that's what we say, you know, in this first episode that we did is to think big.

Yeah. I mean rewiring your brain around the idea that you can, you know, I know people are kind of afraid of changing an accounting system or PM system, but I mean the opportunity cost of not doing that is incredible right now. Yeah.

And I think it's under I think it's underestimated at this point. But really faster.

Cheaper, faster, smarter. That's the goal. Okay. So, what is the vision for Briq today?

So, the vision for Briq today is to take everything we've learned and all my domain expertise and the great clients we have and to make Otto the single source of truth of both action and data at every construction company. Ultimately looking at Otto, we want Otto to be your working operating system.

Everybody touches it. Everybody asks questions of it and people get their work done in it and have Otto do their work on their behalf that Otto can do. It can't do everything. That's the that's I want to be frank. Otto can't autonomously do everything for you, but it can do 60%. And if I can take 60% of the work off your desk while making you more productive, win more business, watch your fee, freeze your overhead, all the stories that we've been talking about.

If I can do that, then Otto is a success. And I think you know the four key pillars of Otto that really differentiate it as a platform for work is the first one is it's fully generative. So if you want Otto to look different at your company than another company does, we support that. So unlike the old systems, I remember in some of the companies I worked for, we wouldn't allow customization. We wouldn't even allowed user-defined fields because that would break the product because products old software products need to be a lowest common denominator.

That world has shifted. That world has shifted. Otto writes different modules for different clients. That is a big factor. We want it to be orchestrated. I'm tired of coming back to best of breed.

I think it's full circle. The power of one is Otto is the one.

Think of it like Neo in the matrix. Otto is the one. It's that one platform. So we want it to be orchestrated across departments. We want to help break that, you know, concept that everybody blames every other department for their failures in construction, right? So it's generative and it's orchestrated. It's autonomous. And this is the big one. When I talked about that story of the lead coming into Dodge, the Dodge lead being research, downloads of plans and specs, the extraction of that and the comparison of that to other projects.

That's a human's job today. And if I can make that happen in seconds as far as the human is concerned and have that data waiting for them rather than them having to go do it, that's a massive savings, massive savings of operation. You know, that's the that's the you know, three of the four pillars. The fourth pillar is and I'm going to say it and for everybody at home knows it. It's an unnatural world to operate construction projects with a tablet and a finger or with a mouse.

We've I've sold a lot of software in this industry have made you conditioned to expect to open up your day on the job site looking at a screen.

That's what we've conditioned. Briq is 100% dedicated to making every work stream conversation first. And that's I think when you say think big, think of a superintendent doing a daily log, not at the end of the week or the end of the day at a screen with buttons. Think of it as a ongoing conversation with his best friend and getting that information out in real time on the daily log. Having that information in there without anybody typing on the keyboard, it sounds aspirational, but so did mobile devices at one point.

Conversation first is what's going to change this industry. Conversation first, generative, orchestrated, autonomous. Remember those four things. The next time you look at AI, ask him how it orchestrates, how it eliminates work on your behalf, how it changes based on what you need within the confines of an enterprise company, the rules and regulations of what a corporation needs to be and think about how it changes how people work. That's what's going to change AI, applied AI in your business.