← 返回

2026.09.30 | TB | Software | E-Discovery: AI Document Processing

原标题：E-discovery Market – AI Document Processing & the Platform Consolidation Squeeze

#software

Third Bridge · 2026-09-30

Analyst-led • 24 Sep 2026

Companies: LAW US

Paul Morse (PM) Director, Enterprise Sales at Everlaw Inc

Saad Maqsood (SM) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

SM [00:00:01]: Welcome to Third Bridge's Interview entitled E-discovery Market – AI Document Processing & the Platform Consolidation Squeeze. I'm Saad, and today, we have with us Paul, who is Director of Enterprise Sales at Everlaw.

Just before we get started on the Interview today, Paul, could you please state I agree or I disagree to the following statement: You understand the definition of material and non-public information and agree not to disclose any such information, or any other information which is confidential, during this Interview.

PM [00:00:40]: I agree.

SM [00:00:41]: Could you begin with an introduction to your background, please?

PM [00:00:47]: Sure. I've been in software sales for now about 1.5 decades, the last four years at Everlaw, which is obviously the industry we will discuss today. At Everlaw, I've been an enterprise account executive for my first three years there, working with law firms, specifically mostly on the new business side. Then, over the last year or a year ago, I was promoted into a sales leadership role as Director of Sales on Everlaw's enterprise legal expansion segment. That's specifically managing a team of AEs, enterprise AEs that are working with customers of Everlaw in some shape or form that had been customers for a year-plus with the goal of growing wallet share among those customers who are mostly defence firms that are passing along the cost of e-discovery to their clients.

Questions

00:01:46] How much market share do you think Relativity holds in the e-discovery market, especially on the software side?

PM [00:02:14]: Quite a bit. Obviously, very few of the customers I'm interfacing with have a direct relationship with Relativity, but I'd say that obviously, the moat for Relativity is through their third-party managed service providers wherever Relativity is, for sure, the default. If I were to quantify Relativity's market share, it's totally speculative. I would say it's probably north of 50%. There's probably quite a bit of long tail, but in the big meaty matters that either Am Law or below law firms are using, I'd say it's probably north of 50%.

00:03:03] Do you think other players in the e-discovery market, such as CS Disco and Reveal, are in the low- or high-teens market share range, or even lower or higher?

PM [00:03:23]: Yes. My speculation, just anecdotally from how many times there's providers outside of Relativity involved in deals would be Disco is probably the only other notable one in the space. Certainly double-digit market share. Reveal, they've made a lot of noise a couple of years ago because they were, I think, at the forefront of AI. From a marketing perspective, I think that allowed market perception to be, okay, this product is more innovative. It's more AI-focused as I think customers have gotten a bit more savvy with regards to AI and Reveal has been forced to acquire the Logikculls and Ipros of the world to keep up because I don't think their core e-discovery product was ever very good. I still am running into them less and less despite the market theoretically tailwinds of acquiring customers of Ipro, acquiring customers of Logikcull. I think that really Disco is the notable other player in the space.

00:04:43] Reveal went through a merger with Brainspace in January 2021 and acquired Ipro and Logikcull in August 2023. Do you think those M&A activities didn't pan out the way the company would have liked or the market was expecting, or were you not a fan of those acquisitions from the get-go?

PM [00:05:10]: I think Logikcull definitely had a good niche of user-friendly, which is certainly not the norm in e-discovery, but I think the product has lacked maybe the enterprise-grade functionality to handle larger matters. I think there were firms that had small cases and they had attorneys that wanted a better solution and Logikcull is a good low-cost solution. I think acquiring those customers made sense, but I think as far as spend potential from those types of customers, there wasn't probably a ton of really strong customers that they inherited from that acquisition. I think Ipro, a lot of those customers had left in the years past. I think Ipro did a good job of locking up customers to multiyear agreements five-plus years ago. Then, over the last few years, those customers have been like, "All right, we know we want to get off Ipro." I'm not sure how much Reveal acquired there as far as inheriting really strong potential customers. I think market feedback from Reveal customers is everyone's been very underwhelmed with how the tools have been cobbled together.

I think Brainspace has pretty good product feedback, but Reveal's core e-discovery tool in which certain functionality effectively they've acquired. It's really like a cobbled-together Frankenstein of solutions. I think the product experience hasn't been great for the customers that have moved over there through acquisitions. Then, I've also heard the teams at Reveal have been disbanded, whether just natural attrition that comes after an acquisition to lay-offs to reduced support staff or customer success staff that a lot of those long-term customers of the acquired companies were really favourable on. For various reasons, I think just yes, Reveal, whatever customers they do have, I don't necessarily view as long-term customers just based off some of those experiences. I think the product is still, especially as Relativity and other tools are developing their own AI functionality, I think individual use cases for Brainspace will be fewer and further between. I think that's going to be a headwind for Reveal as well.

00:07:45] How would you break down Disco in terms of its strengths, weaknesses and recent actions, similar to the exercise we just did with Reveal?

PM [00:08:05]: Yes. I think a few years ago, Disco lost some customers in the bad PR of their company. I think also customers of Disco, by and large, are fans of the product. I think where they sometimes churn is due to more commercial reasons. Disco, it charges burst pricing. While they can do competitive deals from a rate perspective and I think they are part of the new class of e-discovery vendors that has more predictable pricing and sometimes that's a good reason to not use Relativity. I think Disco initially had a great way of selling against that with just predictable flat-rate pricing, but that flat rate, once you've exceeded your minimum spend commitment each month, that typically has a burst premium overage multiplier of 50-75%. If a customer really grows in its spend with Disco, naturally, that's from just putting on a lot more volume than they anticipated, then if they were to do the math, they're like, "All right, well, we're getting USD 8 a gigabyte on Disco up to our commitment, everything north of our commitment is USD 14 a gigabyte or USD 15 a gigabyte."

Then, at renewal, Disco, I think they have a sour taste in their mouth because they were told USD 8 a gigabyte. Then, as they do the math, the effective average per gigabyte rate could be USD 13-14 in that regard. Then, I think that puts them up against some pricing pressure at renewal because they look back at how much they spent with Disco and they're like, "Wow, that was a lot more." Disco has some other commercial nuance. They're not offering early case assessment all that frequently, which is a really good way to drive down costs if you're putting a lot of data in ECA environments vs full review environments. Competitors can say, "All right, with Disco, you're not taking advantage of that. You can use that elsewhere and get some cost savings." What else? With Disco, support, I know they came under fire at their IPO for failure to disclose a lot of their services revenue. I do think that that's another way customers can have some aversion to Disco is the unexpected costs that come when you're calling into their support team and that time is monetised by Disco. I think sometimes that can lead to customers just feeling that they're paying Disco too much and are more inclined to change.

I do think where Disco has been successful is, again, a great product. I don't think there's a ton of customers of Disco that are leaving because of the product experience. Again, those commercial reasons are one. I think they've been solid as far as leading with Cecilia and being pretty aggressive promoting AI functionality. I'd say maybe even the best in the space with the exception of Reveal, but I don't know how much adoption Reveal was getting vs that was their legacy marketing narrative of AI-first. I do think Cecilia was a really strong announcement for them and probably created some tailwinds. I think their compressed vs uncompressed pricing is always going to be viewed favourably, which is like, "Hey, Relativity has cost unpredictability by way of services, etc." With Disco since they're billing on compressed data that's being uploaded to the platform, generally, that is predictable before they even have to process it and upload it to any discovery tool. That cost certainty has always been a strength of theirs and certainly a key selling point in competitive deals for them.

00:12:35] Does Disco having a services arm with project managers and consultants prove useful for most customers, or are customers going to well-established LSPs [legal service providers] such as Epiq and Consilio for services anyway? Is there a clear benefit to offering both software and services?

PM [00:13:09]: Yes. I think usually, the biggest pushback against services is they're naturally going to drive costs up, not always necessarily so. I think if there's clients that aren't necessarily cost-conscious, law firms are sometimes reluctant to replace services because then they're absolved of the risk that comes from failing to collect a certain document or failing to process a tool. I do think being able to offload to services frees up the lawyers from some of the risk of not necessarily following the ESI protocol. I do think having that bucketed with the software they're using can simplify things for clients and while, sure, it may be more expensive than just the client doing it themselves, being able to put some of the responsibility on Disco for those services is a strong value proposition to a certain type of defence firm, because a lot of times, if you're saying, "Okay, if you need services, but you need to go through an Epiq or Consilio", the e-discovery provider loses some of that ownership of that individual deal.

Then, it's kind of a worse customer experience if they're dealing with three parties instead of two, and if there's confusion about exports, this is the kind of a game of telephone more so. Obviously, it's predicated on those experiences with the services going well for then generally the customer to have a favourable experience with the end technology. A lot of that's less in your control if it's a third party vs someone as another Disco employee, for instance, that's handling a lot of the services component.

00:15:04] Do customers typically pick only one e-discovery tool? Have you seen use cases where it makes sense to have platforms such as Relativity and Disco side by side?

PM [00:15:32]: Yes, definitely. I'd say primarily when a firm is like, "Hey, we've got a list of preferred vendors", generally, those vendors are going to be Relativity service providers that may sell Disco or other tools, but the default is Relativity. The biggest reason, I think, for that having multiple options is any sort of conflict of interest. If the firm is representing one side of a litigation and that individual case team is using Epiq, for instance, Epiq would be conflicted out if the other side of litigation also reached out to Epiq to help with the discovery there. Naturally, since those services are having to deal with the conflict of interest there, a law firm is going to give themselves options so that they aren't out of options if they're conflicted out of that based on who the other side is using. I think that's common, especially with, again, various Rel MSPs, but at the end of the day, the services component is a different thing than the end technology.

I think there are firms that have lawyer preferences and I'd say very rarely. I think the only time there's really a strong top-down mandate of, "Use this provider for all matters", is if that firm has a very ambitious, like a minimum spend commitment and any cases not going to that end technology puts the firm at risk of underutilisation and then covering a lot of the e-discovery costs out of pocket vs being able to recover all of those costs based off just utilisation and all that data is accounted for. I'd say for the most part, yes, these firms have tools they mostly use and maybe they have a couple of minimum commitments with a couple of them. For the most part, those are pretty conservative commitments given the risk of under delivery and that allows firms to give users the end attorneys or support like, "Hey, you've got options, whichever one you prefer." Then, usually, there's someone in administrative capacity that's just monitoring the footprint of each and making sure they're at their spend commitments.

00:18:06] Earlier you mentioned Disco might charge USD 8 per gigabyte with a multiplier on overages. Was that example specific to hosting or another use case? On average, what is the typical per-gigabyte pricing on the hosting side?

PM [00:18:35]: Yes. That example was just primarily for Disco's hosting.

SM [00:18:40]: Industry-wide, what do you think is the average hosting rate most players sit at? Is there a rough range?

PM [00:18:52]: Yes. I think that is so varied, and I'd be very curious to see how that evolves with Relativity's migration off-premise. I think for the most part, the Epiq and Consilios and the KLDiscoveries of the world, these giant LSPs, they've bought a tremendous amount of Relativity data in wholesale. I think that I don't have complete certainty into how this works, but I'd say for the most part, they're buying that data and it depends if it's on server or if it's just Relativity and then the LSP can determine what goes on RelOne vs which goes on Rel Server. I think there's a ton of grey area in that because, obviously, on server, these service providers can go really low. I've seen ECA costs at USD 1, active hosting rates at USD 3. Then some RelOne customers like Rel, it's going to be probably higher single digits for that type of scale.

Again, if these LSPs can blend that and then package it to a customer, it's like, "Hey, you get Relativity for USD 5 a gig." Then, sometimes Epiq can determine which goes on RelOne, Rel Server, Epiq's own in-house tool. A lot of times, I think they'll find the right way to create margin arbitrage on hosting. Obviously, they're making a ton more margin on services. To answer your question, I'd say the average for really big matters or big customers that have six-figure spend potential a year, I'd say they're going to be looking at mid-single digits to low teens on the high end, but a lot of that can vary. If it's only Rel Server or if it's only RelOne, that can create an additional nuance.

00:20:53] If, as you said, Relativity's market share is north of the 50% mark, do you foresee any competitors materially chipping away at the company to the extent that it risks falling below 50%?

PM [00:21:40]: Yes. I think the biggest potential disruption to Relativity's market share would be the Harvey, Legoras, some of these frontier models, if Claude for Legal, their newly verticalised arm, if they were able to create a serviceable e-discovery platform, I think that could be a huge threat to Relativity's business. I think where Relativity's moat is really is in its distribution network. If a lot of these service providers are comfortable selling Relativity and its price, at least on paper for hosting is the cheapest, I think that naturally, firms are just going to stay on that course and inertia will win. I do think a forward-thinking firm who is getting a ton of value from adopting AI functionality, and I think the leaders of legal AI are obviously well documented. If Harvey or Legora acquired a cloud e-discovery tool, if they acquired Disco, I think that could be a huge threat to Relativity given where Disco is trading at.

I don't think that's out of the realm of possibilities. I just think, okay, these attorneys, the more and more they use certain AI features, I think then if they see a serviceable e-discovery tool that embeds those e-discovery or the AI functionality really well, I think then they could be stronger advocates to move off Relativity. I think that that, for sure, can happen because I think what I found in market is the autonomy for e-discovery solutions is generally pretty high. It's usually a partner who is the one that's communicating to the client like, "Hey, these are going to be your projected costs. This is the e-discovery component. We recommend you use provider X." Usually, the client will sign off on that, especially if provider X for e-discovery is the lowest in cost. I think these partners can be swayed. I think that's where you'll see like unless there's a minimum spend commitment that requires every case to go there, partners are willing to test out a new tool.

If Claude or Harvey or Legora were to say, "Hey, we've got a new tool, it's going to be really cheap and you can use all of the AI functionality that you've got an intended value from", I think that attorney could have pretty strong conviction to be like, "Hey, I'm going to start recommending that." Relativity's stronghold on that next case, you'll have some legacy data always with Relativity and it could be a slower sea change. I do think that no one necessarily loves Relativity. It's more of just, "Hey, this is what we've used and it's lower cost and our service providers who we rely upon are using it." I do think Relativity, if they fail to innovate very quickly for AI, whether that's through third-party partnerships or building out more robust AI functionality themselves, I think that puts them certainly at a product risk if attorneys and end users are getting a materially more valuable experience using a more AI-driven approach.

00:25:22] How would you rate Relativity's approach to embracing AI so far across its Relativity Air platform and suite of modules?

PM [00:25:53]: Yes. It's so varied because I think a lot of Relativity customers are not fully aware of what exists with Relativity's AI functionality because maybe, one, they're using it on-premise, and there's really no AI functionality for on-prem. I think some of this usage of Relativity happens in silos. I think a user can be like, "Oh, yes, we're using AI. We're on Relativity and we use AI", but that AI functionality can vary widely. It could be like, all right, this is Epiq-like AI add-on and they can bolt it on to a Relativity on-prem instance. I think since those customers aren't necessarily working with Relativity directly, they may just be unaware of what they can do with Relativity or what they can't do with Relativity. I think it's like, "Hey, it's better than nothing. Relativity is telling us that we're using AI or our MSP is telling us now we're using AI using Relativity." I just don't think there's a very widely known full scope of what Relativity's Air can do for them.

I think the on-prem is the biggest variable there, which is just on-prem provides a ton of restrictions on what AI they can use at all in Relativity's e-discovery space. Secondly, I think the notion of AI first pass review is pretty appealing because I think there's a ton of perhaps contradictions in how I think a lot of e-discovery AI functionality is being built vs how firms have historically done e-discovery. I think a tool like Cecilia, which is like, "Hey, you can query with Disco." Cecilia is Disco's AI product, you could query the entire database and really uncover the highest relevance or highest important documents, but the person who would be querying that is going to be like a partner or a senior attorney on that case, who really understands the case and is motivated to uncover the smoking gun evidence as quickly as possible. However, a lot of times, that attorney is not touching that e-discovery environment in traditionally how e-discovery workflows have gone until multiple months after that case has been put on Disco or any other tool.

Building AI functionality that really transforms how an attorney could use e-discovery requires a behavioural change that doesn't always take place. I think Relativity has done maybe a better job of while not having the sexiest of AI functionality embedded within Air and their other AI instances, something like, "Okay, we can immediately do AI for responsiveness", and then that across all their documents can uncover AI, kind of that first pass review, which is more meeting users, "Okay, this is how you've always done. You've spent this first month in a large document-intensive corpus usually just marking responsive and non-responsive, in that with a click of a button, Relativity can do it for you." While it's maybe a little less ambitious from the full functionality like a Disco Cecilia, it is, I think, a good job of matching maybe how they're already using Relativity and it's probably a more commonly used tool by paralegals or lit support or junior associates that may not ever query an entire document set because it's really not what they're tasked to do.

00:30:11] Are you seeing players such as Harvey and Legora crossing into the e-discovery realm yet, or is it more of a potential threat only if they decide to enter the market?

PM [00:30:31]: Yes. I think it's more of an if they decide to go into it because I think there's table stakes things that Harvey and Legora can't deliver on. If attorney uses Harvey and Legora for some of these core e-discovery use cases, I do think that's risky to the attorney of not necessarily following ESI protocol and defensibility of like, "All right, if they use one of those tools and fail to produce something, I think they could for sure be in violation or at risk of probation, etc, if they were to use those tools because I think those tools haven't necessarily been court-approved to be e-discovery tools." I think the more common use case is people are continuing to use the legacy e-discovery players. Then, you'll talk to an attorney on that case and that attorney is going to say, "Yes, and I exported out the 1,000 hot documents that we found, and then, I had Claude summarise it or I used Harvey to draft up something." I think they're toggling to Harvey and Legora in e-discovery and probably more like the later stages of e-discovery and doc review in a way that probably complies with, "Hey, we're using a court-approved ESI certified tool like Relativity, but we don't really love Relativity's AI functionality, so we are using Claude and Legora in conjunction with that, but it's two different product experiences."

I think back to my highlighted risk of Harvey or Legora moving into that space, I do think if they can control that entire environment and go from day one of litigation and being a place where, okay, documents are now being put into Harvey or Legora and if Harvey and Legora can do things like spreadsheet redactions, some of these core e-discovery features that do take a while to build, but I also think there's probably some scale limitations if Harvey and Legora are unable to ingest and do a lot of their AI functionality on millions of documents, that will always be a barrier to them getting substantial e-discovery cases because that sort of scale can only be handled by the big e-discovery players. I think that's something to note as well. I don't see Harvey and Legora in the short term, as their product suite stands today, being able to deliver on things like redactions and handling certain document types that, okay, firms will always need an e-discovery tool to do that.

00:33:41] Beyond the AI-native legal tools and frontier models we just discussed, out of the existing e-discovery tools available today, would you say any player has a good shot at materially chipping away at Relativity's market share, or will the market remain similar to how it is today for the foreseeable future, with just a few percentage points gained here and there?

PM [00:34:19]: Yes. I think probably in short-term increments, not a substantial opportunity. Maybe with the transition to on-prem to RelOne that could provide more of a significant sea change in market share. Beyond Harvey or Legora and net new players entering the space and chipping away at market share, I do think the legacy players that take a substantially different approach to AI partnerships provides an opportunity. What I've heard in the market is that Relativity's announced partnerships with MCP integrations with Claude, etc. That's more of like administrative. I don't necessarily think while it provides walled gardens to somewhat shared data, my interpretation of Rel's AI partnerships is less around, "Okay, you can actually start doing some of your Claude workflows or Harvey or Legora workflows in Relativity." I think Relativity is trying to develop something that's a bit more of like a frontier model itself and to provide some of that functionality, which I think raises their ceiling if they can start stealing market share in workflows from Harvey or Legora, good for them.

If Harvey and Legora continue to take off and firms have really strong uses for it and people are mandating, "Okay, we want to use Harvey and Legora as much as we can", if Relativity is not giving them that functionality in Relativity to continue scaling and doing those Harvey and Legora or Claude workflows more frequently, I think that could index more usage towards an e-discovery player that does have a more open partnership from a product perspective with the AI players because I think the real value of AI adoption and things like that, especially if firms are mandating, "Hey, use AI as much as you can in e-discovery", if that AI adoption and familiarity is really concentrated around those tools, then I think that could force more people not to use Relativity because of their reluctance to, I think, more deeply embed third parties into their product workflows.

00:37:05] Beyond the risk of AI-native start-ups breaking into the market, do you see any disintermediation risk to software platforms such as Relativity and Disco due to AI, such as functionalities being in-housed by law firms and corporate legal departments or document review volumes dropping because processing is done faster?

PM [00:37:56]: Yes. I think what AI is great is obviously, distilling large amounts of data into small amounts of data. I think as AI thinking about, all right, what drives up e-discovery costs, it's generally large amounts of data that you know very little about. You have to continue keeping that data in an e-discovery environment so that you are only removing it from that environment once you know it's irrelevant to that case. I think as AI can accelerate findings of which documents matter and which don't and the ones that matter that are on the platform for a long period of time are much smaller in quantity from a proportion perspective than the ones that don't, I think the AI could, okay, if you only need 10% to 20% of this data on the platform long term because that data is responsive or relevant to this case, the faster you can call out that 80-90%, that dramatically changes the amount of spend the customer is going to have with the platform, because once that data is removed, the cost of that hosting footprint goes down 80-90%.

I think the e-discovery players are motivated to uncover what matters so that it's a competitive advantage for the law firm and they have a really strong product experience. They're like, "Wow, I want to use that tool next time because it found the smoking gun evidence on day one, otherwise, that would have taken me months." I think the e-discovery tools aren't as motivated to highlight which data is truly junk because that would have a meaningful revenue impact, if not only on day one, you can uncover the hot documents, but on day one, if you could identify the 90% junk. Obviously, then that 90% junk could be removed that day and then you only have one month of hosting at that volume. I do think that the e-discovery players, they recognise, "Okay, we're highlighting the key pieces of evidence in e-discovery, but simultaneously, we don't want to do too much to probably identify all the bloat that drives up hosting costs."

As AI improves, and I think right now, the limitations of Harvey or Legora can be, okay, just the scope of the number of documents they're able to analyse, but as those frontier models and these e-discovery tools as well improve, "Okay, how quickly we can make sense of wider ranges in volumes of documents", I think that could potentially just reduce the amount of time a large amount of data sits on these tools because these firms are under client pressure of reducing e-discovery costs. I do think these firms, they're under more pressure to use technology to improve efficiency. If that's reduced e-discovery costs, I think that's a quick win for these clients that, especially for these big cases, have a ton of scrutiny on the e-discovery bill. I think one last piece that hasn't been mentioned, and it's a unique moat of Relativity's is their stronghold on insurance carriers. Oftentimes, whether it's with the long tail of slip-and-fall cases, but 75%-plus of litigation, I'd say, is covered by a carrier, at least in the defence litigation space. If that carrier is footing the bill for e-discovery, they're ultimately approving what gets used. These approved vendors have all been carrier-approved. A lot of times these carriers themselves have approved vendors.

The carriers, as you can imagine, are going to be buying on cost. They do not care which tool is faster and they see a Relativity on-premise option for USD 3 a gig and then Disco or someone else for USD 10 a gig, the carrier is never going to change from using that USD 3 a gig every single time. These e-discovery sales teams and technology providers, it's such a steep mountain to climb to get the carriers to change their philosophy on what e-discovery tools will be used that that is where the inertia becomes even more difficult to disrupt because there's just never going to be a time where these carriers are going to zoom out and say, "Hey, what's the best tool for us? How could we better leverage AI?" It's more so of like, "Hey, we've used these solutions, where this is approved. We're not going to change things up." I think that the insurance stronghold on Relativity service providers specifically and those service provider costs buying in wholesale, whether it's through the carrier or just their own house contract with Relativity, it makes the cost disparity extremely hard to overcome.

00:43:36] Given the current talks on the adoption of AI among law firms to automate tasks previously handled by junior associates, could a reduced intake of these associates result in fewer user licences? Ultimately, could this create a net-negative or knock-on impact on subscription-based legal tech tools? Is this something you have taken into consideration or you're seeing happen?

PM [00:44:34]: I think probably historically speaking, where Relativity had user licences for some of their on-prem instances, it would have been a bigger impact. I think, two years from now, the probably majority of Relativity will be on RelOne, especially as all net new cases move to RelOne and I haven't heard really anything as far as user licence costs there. Even if they've got smaller case teams using Relativity, I think the ultimate driver of cost or spend with Relativity will be the amount of data. I think the bigger risk in how Relativity's revenue could be disrupted is if AI, I think, can pull out the junk either more quickly when the data is on Relativity or another tool or even before it's ingested, if perhaps Claude builds some product that can analyse a bunch of documents and maybe that wouldn't be as defensible in the eyes of e-discovery and ESI law. If the amount of data that the end clients that are in litigation are providing to the law firm to put on to Relativity, if functionality emerges that can whittle down that size of that data, I think that's a bigger risk to Relativity's revenue than the smaller case teams based on user licences.

00:46:09] Looking from the outside in, do you have a rough sense of what portion of Rel Server's customer base has completely migrated to RelOne vs those who are still primarily on Rel Server?

PM [00:46:36]: Yes. It runs the gamut, I think, for sure. Rel, I think there are some legacy cases that will still be on Rel Server for 5-10 years, the lifetime of these cases, but I think those types of cases are more like on ice in their tens of terabytes. I think there's just low cost to keep them on Rel Server, but also there's just not a ton of day-to-day work. These cases are mass tort matters that will probably get judgement in 2040. I think some of those, as a percentage of a customer's footprint with Relativity, it's like, yes, these legacy matters, I think, will stay on there forever. I do think that the managed service providers are probably not putting any new cases on Rel's Server because of RelOne. What I've heard in the last year or so is that RelOne's pricing is getting a lot more competitive than what it was.

I think sometimes these firms are getting options of like, "Hey, RelOne and Rel prem pricing for the first year of this contract are going to be the same." I think that motivates customers to put everything under RelOne, and it obviously reduces any sort of administrative workload of needing to migrate those down the road. I would be sceptical that they would maintain those prices at renewal. I think Relativity is rightfully acknowledging risk of, "Hey, if this data needs to come off Rel Server, it needs to go somewhere, and it's not certain that it goes to RelOne." This could open up the door to some churn risk. I think they're motivated to get these cases in the RelOne doors as easily as possible through lower pricing, but my speculation would be at renewal, these RelOne rates are going to go up pretty considerably.

00:48:55] Would you say RelOne pricing is still at a premium relative to the rest of the market? Do customers already consider it quite expensive despite price reductions?

PM [00:49:17]: I think RelOne pricing is probably slightly higher than what they get from some of the other large players in the e-discovery space. I think on-prem was always way lower, but the service providers would make up for that with services. I think Rel on-prem required a lot more services. Maybe it was preferred too by these managed service providers because RelOne, there is more user level abilities to create search term reports, which historically would have to be farmed out to the LSP. If a firm can do that, it lowers obviously the services cost for that matter. I think the exception would be where RelOne pricing is still competitive to anyone is for some of these really large Relativity customers that even through an LSP or through Relativity directly are spending USD 500,000 plus a year. I think that that economy of scale is going to get good rates regardless of who they're going against and regardless if it's RelOne or Rel on-prem.

00:50:34] If Relativity lowers RelOne prices to incentivise customers to migrate from Rel Server and then hikes prices at renewal, do you see considerable churn risk, or are customers effectively locked in due to high switching costs once they are on RelOne?

PM [00:51:15]: Yes. I think that's for sure a risk to Rel, and I think they probably view getting through these next two years of migrations as a huge milestone. If I'm Relativity, I'm making sure that every single large on-premise customer is motivated and supported to migrate to RelOne because I do think once they're in RelOne, it's harder to churn. I think because of a better product experience, because they're in the cloud, because they now have AI functionality, I think the time for competitive conquesting of Relativity customers is really right now because on-prem lacks a lot of modern product infrastructure that's becoming more commonplace. I think RelOne customers are having a better product experience. Unless their costs are exorbitant, they're less motivated to move off RelOne, whereas Rel on-prem customers they have had a worse product experience and maybe they've had more acid reflux as far as unexpected services costs. I think it's easier to create wedges against Rel on-prem customers than it is RelOne customers. Back to my original statement of if RelOne can migrate all these on-prem customers to RelOne, I think they will be in a strong position to keep those long term.

00:52:48] Given Relativity's 2028 mandate requiring all new matters to be hosted on RelOne, is 2026-28 the best time for competitors such as Disco and Reveal to go after Rel Server customers before they start putting new matters into RelOne?

PM [00:53:26]: Definitely, definitely. I think this is where probably some of these services providers are responsible for the legwork of migration. Since Relativity lacks a strong in-house support team for that sort of stuff, if the service providers were motivated to move those elsewhere, I think they hold the keys. I think for Relativity competitors, if I were to find the best way, I think it would be to also work with some of these MSPs to motivate them in some way to rather than migrate and do all this work from Rel on-prem to RelOne, where if things go wrong, Relativity is not going to be able to really support you all that much, is trying to craft an easier story of migration elsewhere because my understanding is like a Rel export. It doesn't always play nice even with RelOne, whereas a Rel export can be ingested into another tool and preserve a lot of work product. It's not like RelOne has any advantages of inheriting live matters than a non-Relativity provider.

00:54:58] Is the onus on competitors to offer monetary or other incentives to LSPs so they start pushing alternatives to Relativity, or are LSPs already motivated to push alternatives, given that the high margins they enjoyed on Rel Server will decrease with RelOne?

PM [00:55:37]: I think they are. I just don't think they are able to overcome the pricing hurdle, if these law firms are like, "Oh, wow, we can get RelOne pricing for the same price as our Rel on-prem pricing, why would we make a switch, especially if all the other options' costs are way higher?" I think that could come back to bite them at renewal where we mentioned Relativity could raise their prices for RelOne but I think these service providers, it's like, "Hey, we can't really support the competitors of Relativity's costs when they're 50%, 25% higher." With certain volumes, even those 25% differences can make hundreds of thousands of dollars in cost disparity. I think these LSPs, they're like, "Our customers don't necessarily love Relativity and they're not necessarily attached to Relativity nor are we", but when their other options' costs are way higher, we're probably just going to stick to what we've been doing.

SM [00:56:41]: I'll leave the questions there in mind, Paul, I appreciate we're already over time. For the final minute or two, if there is anything you feel like we didn't discuss and is important to highlight, then please feel free to do so now. Otherwise, we can wrap things up there.

PM [00:57:07]: All right. Yes, that sounds good to me. I think we covered it all, but I appreciate the chat and always happy to expand at a later date.

SM [00:57:16]: Perfect. No, I appreciate it. I appreciate your time. Yes, hopefully speak again soon.

PM [00:57:21]: All right. Thank you.

SM [00:57:21]: Thanks again. Have a good one.

PM [00:57:24]: Bye.

Transcription ends at 00:57:25 of the recorded material