Lexful’s AI-Native Documentation: New Accountability and Risk for MSPs – With Pinar Ormeci

[00:00:01] Dave Sobel here, reporting from the floor of ChannelCon. I'm joined by Pinar Ormeci. She is the CEO of Lexful. Now, Lexful went generally available this morning when we're talking. This is an AI-native documentation platform for MSPs, backed by a $7 million seed and chaired by Chris Day, the man who built ITGlue. Now, quick disclosure, Lexful briefed me back in February about their launch, and I covered it on the show, and I was a little bit skeptical at the time.

[00:00:29] So you can consider this a follow-up. Pinar, congratulations on the general availability and joining me on this very busy day. Pinar Ormeci, Yes. Thanks so much, Dave, for having us. And indeed, it has been a very busy couple of days. I figured. So you went GA this morning. For MSPs who haven't been following all this, what is generally available today and what is different from what happened in February? Pinar Ormeci, Sure. So for us, GA really means a couple of things.

[00:00:54] Pinar Ormeci, Number one is now we are integrated into all the major PSA and RMM tools out there. Whatever you have in mind, we are now integrated into that. Pinar Ormeci, Number two is when we launched in February, we really did a phase launch with a limited set of integrations. Pinar Ormeci, And some of the features were, let's say, 0.8 versions of the features that we really needed to harden. But this was a strategy on our side.

[00:01:22] We really wanted to put the finishing touches to the product, the last mile of the product, in the market, under full load, working with all kinds of different MSP data. And now with GA, we have solid MSP grade, really polished versions of all the features that we need. And of course, on top of it, we have many more AI wow factors. That's what we call them. So somebody signs up today on GA. Yes.

[00:01:49] What's the headline feature that they're getting that's impressing the most? Honestly, it's paradigm shifting. So what Lexful does is everything that a legacy documentation solution does, but also a lot of other things these solutions cannot do. And it's easy to say we're just an AI native documentation platform, but there's so much behind that, right?

[00:02:13] So number one is before we even bring any data into our platform, we context engineered a full general LLM, large language model. And the reason why we've done this is LLMs are the context window of a general purpose LLM is huge. And if you use just a generic co-pilot or cloud, it will hallucinate because the context window is simply too large.

[00:02:43] So the number one thing that we have done is we context engineered the LLM to live in the MSP and IT world. And what does that mean? It is much more different than writing a really long prompt and telling it, hey, you're an MSP expert. It's actually building the whole architecture, the walls, the ceiling, and all of that to make sure whatever the LLM does is thinking like an MSP and IT insider and an MSP expert. So that's the number one thing that we have done.

[00:03:13] Number two, look, in all major technology shifts, the operational foundation has to be recreated. Like if you look at cloud, for example, you have to have an orchestration layer. You have to have cloud management. With security, you have SOC centers, right? And with AI, that operational foundation is trusted, structured data, structured knowledge.

[00:03:39] And because of this, for example, while we're migrating from an existing documentation platform, which we replace, right? We are not an addition. We replace those tools. While we're bringing that data, we bring everything as MD files on a knowledge graph. So everything is vectorized and lives in relationship with each other. So now you have a context engineered LLM that is already living in the MSP space.

[00:04:06] And you're putting that as the AI, as the UI on top of that structured knowledge base that comprises of all the MSP's data and their clients' data, assets, credentials, and what have you. So now in plain English, whatever you ask that AI, it will immediately find the right context with minimal hallucinations. And it will give you all the sources. And if it doesn't know something, it will tell you.

[00:04:33] And this sounds like small, but this is a huge thing for MSPs because they are on top of mission critical systems, right? Like client trust and minimizing hallucinations. Being able to trust what the agents are doing is fundamental. So in February, you told me Lexful is a system of record, not a system of action. Correct. Right? And you'd integrate and not replace.

[00:05:00] But since then, you've described the ambition as more like God mode, eventually ahead of the PSA. So that feels like a difference. And what is it and what is it going to be in 18 months? No, so our fundamental approach and the vision really has not changed. It is what, when I say God mode, what I mean is this always up-to-date system of record that can fuel all the agentic work that you are doing, all the agents, all the automation.

[00:05:29] That is the God mode. The more context you give to Lexful, the more context it will give you back. For example, right now we are doing a much deeper integration with ScalePad. And just imagine bringing that customer context now to all the answers Lex is giving, right? The use cases for Lexful is not only solving tickets. It's also around QBR preparation.

[00:05:53] It's also about auditing client documents, finding the gaps as it relates to NIST compliance, for example, or SAC2, Type 2 compliance. And then using Lex to create those documentation based on the templates that you also have. So we are still fundamentally a system of record. We use skills and agents ourselves as it relates to keeping those records up-to-date automatically.

[00:06:20] And if anything, PSA, RMM, or the system of action vendors love working with us because we make them look better. Because now they're AI. They're very like agent food, if you will, right? Agent food that's good for the soul. That's a nice way of putting it. So I want to give a couple of your wins first because they're real ones. So SAC2 was ready in February. Type 2 is now done. Yes. And you've shipped the MCP server, right? Okay. Both great accomplishments there. Now, this is the one that I was tracking that went the other way.

[00:06:50] Your June seed release elisted featured vendor in both PAX8 and SureWeb's marketplaces as a completed milestone next to the SAC2. But today's release says those are targeted before the end of 2026. What happened? No, nothing happened. So featured vendor, I guess, is like maybe it is a bit ambiguous. So we are working with PAX 8 and SureWeb with both the contracts are signed, the API integrations are ongoing. So featured vendor is, if you will, coming soon vendor.

[00:07:20] So we are committed to being in these distribution markets because our goal is really to provide access and to be the partner of choice for that AI journey to the MSPs wherever they are. Some of them want to transact in their own currency, like Australian MSPs. So being in PAX 8 type of platforms gives us that ability. This episode is brought to you by Control Map.

[00:07:47] Growing MSPs are using Control Map to build recurring revenue by expanding their GRC services. Starting now, Control Map is offering a free plan for MSPs looking to get started with providing compliance as a service. Create a free account and run an assessment. Track key items like policies, risks, and evidence in one place. It's a practical way to prove value to a client before deciding to expand your compliance offering. Try Control Map for free today.

[00:08:15] Visit scalepad.com slash Dave to get started. That's scalepad.com slash Dave. Back in February, you gave me a number. Roughly 130 MSPs in active trials. Yes. Today's release has no customer account, no partner account, no AR. Yes. So six months in a year later, I would expect a better story. What's the number?

[00:08:38] So I'm not going to give numbers, but I can tell you we have hundreds of MSPs, either paid. So we have lots of paid MSPs, but we also have hundreds of MSPs in the pipeline that we're waiting for either one integration or another. We have not supported yet. So, for example, with GA, we now have data RMM and central. Like, we didn't have these until now, right?

[00:09:06] So if you're a data RMM MSP, you could not come in. Or, for example, we did not migrate from Hudo MSPs. So all of that changes with the GA, and it really wasn't the, A, the market isn't interested. It was more us opening the doors in a structured manner, like one by one. Okay. So the other math I want to ask about, so $20 per client per month. Yes. $300 minimum, 12-month term. Yes. Right? Okay.

[00:09:34] So a product like Hudo is $27 per user, no minimum, no contract. So the crossover that I did in the math is around one and a third clients per tech. Okay. And MSPs that I was looking at run four or five. This, for most of this, the building is several times like the incumbent price. So tell me what justifies the price differential. Sure. So the reason we charge per client versus per user is we have use cases across the board. Right?

[00:10:05] We are not just a document repository. And if you structure all that you have and then you give the ability to search, create, fix, fine. The use cases are not only for your technicians, but also for your customer success people, for your account executives, for your sales and marketing, for your VCIO people, for the QBR print, for the CEO.

[00:10:30] So for a knowledge operations platform like ours, giving access to all the employees of the MSP is fundamental. Like you cannot gate the access to knowledge via a per user licensing. Right? It's like the gatekeepers. Or worse, I know it from the MSPs. They do license sharing. Like I'm coming from cybersecurity background. Like in this time and age, that is asking for a disaster. Right? So we just say no to those things. That's number one.

[00:10:58] Number two is MSPs very quickly realize when they start using Lexful that we are way beyond documentation. So I think when you're doing a cost analysis, you really need to start looking at, hey, what other SaaS tools does Lexful replace to how much labor am I saving? Like the ROI on our platform due to the order of magnitude operational efficiencies you have.

[00:11:26] Like honestly, like you probably can get it in a couple of months. So we are a knowledge operations platform. And we keep evolving and innovating. So I hear you. That's another thing. Like we have tiered pricing. Right? If you have volume base, it's not on the website. But definitely like come to us. And we really want to show you what's possible. Okay. So I'm going to ask two things that I don't think anybody else will ask.

[00:11:54] And I'm going to try and be super precise because I do not intend these as accusations. I want to get some clarification. Right? So the first one is the product holds credentials. Right? And you also ship an MCP server. Ask Lex does the things like search and actions. Those are your words, right? Yes. From your page. An agent with reach into the password vault acts. It's wrong. And a client goes down. Who owns that? Is that Lexful? Is that the MSP? Is that the end client?

[00:12:22] And like SOC 2 is control. So who owns that kind of failure? Honestly, that's why you need to have all these governments layer before you even give access to that agent. Right? So first of all, we do record level access control to everything. So to the smallest password, you can actually control access at the record level. That's number one. Number one.

[00:12:51] Like for example, an L1 tech cannot access finance passwords. Right? So even within the same technicians, you can actually control who accesses what. Number two is all of the answers Ask Lex gives or actually all the reach even Ask Lex has is dependent on who's asking it. So based on who's asking, there are some contexts Ask Lex can even not reach. Right? So that is fundamental.

[00:13:21] And ultimately with AI, you have to have human in the loop. Like we do everything to minimize risk, to minimize hallucinations. That's why we have the source everywhere, the links and all of that. But when it comes to mission critical stuff like passwords and credentials, we always say that there has to be a human in the loop as the final arbiter. And then in the beginning, I don't think any agent should just go in and get credentials and so on.

[00:13:50] In a backup market that keeps consolidating, MSPs are asking a smart question. Who's actually going to be there for the long haul? Comet Backup has stayed focused. Same flexibility, same MSP first pricing, same commitment to letting you run your business your way. No surprises. If you want a backup platform built for how MSPs actually operate, go to CometBackup.com.

[00:14:21] I want to sort of offer you a little bit of my argument against Lexful, and I want you to push back on me right here. If you ask MSPs why documentation fails and they don't blame the tool, it's because of the billable, it's senior tech hoarding their knowledge, it's leadership not enforcing it. If the problem is incentives, giving them the best platform ever doesn't change the human action. So why aren't you just an expensive piece of software on the shelf?

[00:14:49] What makes this different than relying on the humans to drive that? I mean, how is that working, right? I mean, I told the hundreds of MSPs, like they tell me, they give all kinds of incentives. Like it's the human, like even in our own company, it's just the human nature. People hate going back and writing things once a problem is solved, right? Like people have tons of stuff to do.

[00:15:15] So this has to be fixed at the automation level. Like you cannot rely on humans to do manual work. And this cannot be the best practice. Like your best practice cannot be forcing people to do manual work. Like that's what kind of a best practice, especially in the AI era. There are better ways of doing things. And that's why exactly why Lex will exist, right? So yeah, I'll push back heavily on that. Manual labor is not the answer to keeping documentation up to date.

[00:15:44] That's a fair position. So as we wrap up our time here, like what's the thing that you think an MSP should not do in the next 12 months? Should not ignore us. Okay. And I'm being honest because guys, like I'm a tech executive. I've been in the tech space all my life. I was very involved with digital transformation. And these are transformational times.

[00:16:12] Like the urgency, like any MSP who says, oh, we don't need AI or we're going to do something else first. Like you're already falling behind. Like this is something that we've never seen before, right? And it's only accelerating. And for AI to work properly, your data has to be a part for the agents to do their job.

[00:16:35] So if you have data out of like all over stale or actually lives in people's heads, and now you bring all these agents, you're so excited. Like what will that agent do? It's just bullshit in, bullshit out, right? I'm sorry, but so it has to start with the knowledge layer. It has to start with structured knowledge. And that's really Lexful.

[00:17:00] So we are not a documentation platform per se in the sense we did not first create a documentation platform and then start try like doing AI on top of it. No, we are an AI native knowledge infrastructure platform. And then we bring other people documents in a way that works for the agents, works for the AI. So like joking aside, like MSP is like the shift is here. The time was yesterday. Okay. Like you have to be the customer zero.

[00:17:31] And the other thing is, so this is another thing. Operational efficiency is one thing, but monetization is where the big next thing is, right? So all of your clients, like there are 200 million small, medium businesses. I don't know, 100 million of them have an MSP, right? So they are already using AI and they're screwing it up. They're uploading their sensitive documents to whatever. And there's a whole this shadow AI already going on.

[00:17:59] And MSPs are responsible if there's a security incident. If, you know, they are responsible to their customers. So they have to figure this out. They have to become super comfortable using AI as a customer zero. And then really help with the governance, help with security, help with deployment, like help with this forward deployed model.

[00:18:23] So, yeah, like it's, there is an urgency that I cannot like emphasize enough. Well, I think you've made the case. I really appreciate you joining me today. Thank you today. They always ask the hard questions, but I hope I did a good job. Well, you've had much said. I appreciate you joining me. Thank you so much. What would you fix in your business with the right playbook? Small Biz Thoughts members get a library of templates and operational resources,

[00:18:50] recorded member calls, and classes through IT Service Provider University. Operational education built specifically for independent MSPs. Start at smallbizthoughts.org. Interested in advertising? Head to mspradio.com slash engage. The Business of Tech is written and produced by me, Dave Sobel, under ethics guidelines posted at businessof.tech. Thanks for listening.

[00:19:18] I'll see you on the next episode. Produced by Picture This Video, part of the MSP Radio Network.