InfoGov Hot Seat: Global Information Governance Day 2026

Executive Summary

This episode examines how information governance teams are managing increasingly fragmented data across document management systems, cloud storage, collaboration platforms, email, and AI tools. Tony discusses the challenge of supporting multiple generations of legal professionals, each with different technology habits and expectations. The conversation focuses on data classification, defensible disposition, retention workflows, auditability, and meeting attorneys where they work. It also explores how firms can prepare their data for AI while avoiding the creation of even more unmanaged information.

Welcome to the InfoGov Hot Seat

Jim: All right, so today is Global Information Governance Day 2026. We have five identical questions, real answers on time to value, defensibility, integrations, and AI governance.

Welcome to the InfoGov Hot Seat, the show where information governance flips into governed intelligence and data becomes your AI advantage. So I’m Jim Marifield for this Global Information Governance Day special.

I am joined by my co-host for today, Lee Zidwick. Lee, thanks so much for being here.

Lee: Thanks, Jim. I’m looking forward to the conversation and hearing more from our partners on what teams can actually implement this year.

Jim: Yeah, absolutely, and of course looking forward to having the conversation with you and the others.

Let’s welcome Tony.

Hey, Tony.

Tony: Here.

Jim: Here he is.

Tony: Oh, where are you?

Jim: We’re doing well. We’re glad you’re here with us now. You know, we feel a little better at this halfway point. Now Tony Ford’s in the house.

Tony: That sounds like trouble.

Jim: So it’s great to have you, Tony. Tony Ford from Future in Tech, coming again to the hot seat.

Before we get to some questions that we, of course, want to ask you, focused on the solution, do you mind sharing your title, what its focus is on, what you’ve been doing, maybe a personal background experience?

We know you’re a sneakerhead, so you can’t use that.

Tony: I will start there. I’m a retired sneakerhead because my dear friend, Mr. Marifield, here, did a sneaker challenge in ILT, which I lost. And then had the nerve to show up at ARMA with ARMA socks and matching Jordans.

I’m off the game. You can win. It’s a young man’s sport. I’m doing this.

Jim: There’s always an extra, Tony.

Tony: No, no, no. He cleaned my clock this year, so I’m going to just actually go with the sunset.

I’m one of the co-founders of FTI, along with Jim Pickett, my CEO. We’ve been together about 23 years now, through this business and other businesses together. So we’ve been doing this quite a while.

We started a long time ago. We had a managed service company, and we wrote FTI for our clients because we were dealing with a lot of midsized firms. It was back in the day for paper, a lot of crowded filing, and we just built it into something that’s become quite a beast.

So we separated the company a few years ago, and we picked out that pretty good. Today my title is CRO. Tomorrow it might be who knows what. It’s whatever we need to be to do is what I do.

We’re a little bit different than some of our friends who have already had a chance to speak, and then obviously our friend Chris after us. But we kind of meet where information governance and data governance combine, because the spectrum has changed.

It really has moved into, especially in law firms, you’re all managing four different firms just based on vintage. You still have the very structured old litigators who just need a pen and a pad and a box, and they don’t care about anything else, all the way down to the lawyers and the partners under 40 who just grew up in a different world with AI and capabilities that they don’t understand why you need a structure.

They don’t even need a name. It just says that they’ve just blocked storage. It’s all they work for.

So trying to manage and build systems for four different types of users in the same firm gets a little tricky. So that’s what we focus on. The high-level focus.

What Surprises Governance Teams in 2026?

Jim: So, in one sentence, Tony, what surprises most governance teams in 2026? What do you think?

Tony: I think the actual amount of data that’s not in a DMS.

If we go back to the original way we used to structure things, and an R system said that the DMS was a source of truth—and it is—and that’s where we should respond.

Back in the day, and not to make myself that old, but when we used to actually declare records and then put it in the RMS and all those different fun tools, it’s actually inverted now.

Now you have all these different points where people are working, whatever they want to work in, and they forgot to send it. It was probably both, I guess, before we—but the team’s staying in SharePoint, and everyone’s just sharing documents.

Then, if you leave it there and you leave it in OneDrive, you now create a complete collaboration nightmare, and then you’re trying to figure out when to save it into the DMS.

What you have to do now is monitor those areas and try and take—we call them endpoints because it’s easier—because we have to deal with about 15 endpoints right now, going through both the global suite, the Microsoft suite, and the e-discovery tools as well.

Then you pull that back to the DMS as the authoritative source.

The reason you want to do that, and we’ll get a little bit more into it on the AI side, is that in order to teach your AI tools, you have to clean up your data set. In order to get a good data set, before you can get to the learning tools doing their thing, you have to come up with a matching scenario where you’re looking at a couple of different sources: time billing, the directory, DMS.

Then you start to be able to teach the tools how to actually get more than just client matter, but to actually get down into the nitty-gritty, understanding who worked on what and when. Then you can really build a model that your AI tools can handle.

If not, you’re just going to have unstructured data everywhere.

Lee: Yeah, let’s think about that. But I agree.

I’m curious to know, though: do you see any initial or first sign that this surprise is already happening?

Tony: Yeah, I’m starting to see it. We do a lot of classification projects for firms, and we’re starting to see lots of firms with four or five hundred terabytes of OneDrive data, plus SharePoint data, plus Box data. Then they still have a DMS cloud scenario or DMS on-prem, and then another archive behind it.

You’re sitting there saying—and that’s not even for the bad actors, because we all know we all have clients who will still use their Outlook as a way to hide more files. Even if it’s in the deleted folder, they’re still doing it based on the vintage.

Trying to combine all that and come up with something that’s holistic is really, really a challenge. So that’s the surprise.

I think when people see that cloud storage is so inexpensive, they don’t think about it. “Oh, cloud, 50 terabytes in OneDrive.” Then you look at it and you actually go out there and classify it, and you say, “48 of that is actually client-driven. I really need that in DMS.”

Or it’s what’s been touched in 10 years. Why is it still sitting there?

So that’s the surprise that catches a lot of people off guard.

Jim: Yeah, makes sense.

The Top Information Governance Priorities

Jim: So let’s talk about it. Here’s a question: what are the top three IG capabilities you’re prioritizing in 2026? What customer problems are driving those investments in those three capabilities?

Tony: You know what? I would certainly just say ROT and disposition of that type of data, managing unmanageable data, dark data. That’s probably our largest priority.

Something that’s popped up recently, and it was discussed earlier, is having the attorneys work where they work.

We’ve been seeing a lot of requests for this, and we’re doing something for a top 20 firm right now. They like to do the review in Relativity, so the attorneys never leave Relativity, and we give them all of the things out of it.

They keep working in their approval process, and it has nothing to do with the DMS. It just has to do with that’s where the attorneys want to work.

If attorneys work in SharePoint, we deliver them whatever they need to look at in SharePoint. Work in Everlaw, whatever it is, they do as part of their habit, and what we’re trying to do.

Most of all, they want to work in the DMS, giving them the ability to stay in the DMS. It’s a little harder with the DMS because it’s rules, but the other e-discovery tools are really built for that, and they don’t care about the source and what they’re looking at.

Really, what we’re talking about is reviewing that data for transfer, that data for disposition. We’re not actually talking about document review. That would be something that they would do.

But that’s really become a very high priority in a lot of firms: just making the day and the life of the attorney a little bit easier and not teaching them new programs. So we just started that groundwork, and that’s where we’ve really been focusing.

I think the files are not going away. We are still getting requests for label designers and all kinds of fun things, and like everyone else, we integrate with all the big box houses.

The next phase for all of us is to touch rules, because everything has fine-grained keepers now. Everyone has turned. So just having those become more prevalent in the way you handle your disposition workflow, because we’re seeing some real movement around that.

A lot of firms are going to practice group rules that are on the firm, not necessarily outside counsel anymore, because some of them just don’t make sense for the business.

Then you map them, you see them, and the HUC has the right to overrule it. They saw the firm needed the partnership and what they want to do.

So we’ve been kind of flowing around those three areas the most, and there’s a lot of investment for us right there this year.

Lee: That’s really beautiful. Out of this, real quick, Tony, I was a priority listed. Any that you would say is most driven by the AI adoption pressure? We talked earlier about how quickly things are moving.

Tony: Yeah, the classification side is driven by that a lot, because a lot of people assume that you can auto-classify very easily.

It really is a two-part mechanism as you look at it. I mean, you have to match it first. There needs to be human interaction. You have to teach the tools, or to your point earlier, you will end up with a bunch of just bad data.

You will just have it all classed. The thing about machines is, when they don’t have the answer, they make up the answer. So you have to really, really be very careful on how you leave the machine.

That’s really the pressure.

We’ve been focused a lot on AI research because I think that helps a lot more. That’s a very good, easy step.

I’ve been trying to map into some of the AI tools that are out there. A lot of our tools are really fun out of things. But these AI tools generate a ton of data.

So we’re trying to make sure that we don’t create another big problem by talking too much, calling too many times, downloading too much. We’re just trying to figure out how to minimize our footprint within this whole undertaking.

What Does Success Look Like in the First 90 Days?

Jim: So, some of these complexities, though, for a typical law firm or other regulated organization, what would you say a successful 90 days looks like with your platform? The first 90 days?

Tony: I think the great thing about, I think, all four of us as platforms—I can speak for myself, speaking for the industry—I think we’re all very, very good at adding outcomes in the first 90 days.

I think we’ve all really come a long way. We used to focus on data, not tapping. If you were live or if you had a first pass in 90 days, you were a woo-woo.

But now we’re actually going out there and looking at data, coming up with ways to report, and showing you what your ROT looks like.

Let’s really take a look at what kind of data you have.

Then truly trying to hone in on, like everyone said, just take a practice group and start working on retention. You know, we’re a workflow engine and everything. We allow you to create all kinds of different retention and disposition workflows by practice group and area of law.

Creating those forms so you can get a feel for what’s out there. The attorneys and secretaries can get a feel for collections and see it in your dashboard, so you can see, “Oh, none of this has been worked on in eight years. We could probably address some of this.”

Some of the problem needs to be archived. Some of it just needs to be destroyed. Some of it needs to go back to clients.

But having that as an outcome, because we get it so fast now—the way we integrate, it used to be box-driven, and you’re wondering what was in a box. Now it’s just API calls.

So you get a ton of data very quickly.

In the first three weeks, you’re running reports. Once you run those audits, you now have, in some ways, it’s great because you have more information than you’ve ever had. Otherwise, it’s a little scary because now you see what you have.

You can figure out all about actors, and then you have to figure out what to do with them.

Audit Logs, Reporting, and Defensibility

Jim: Yeah, that segues nicely into our next question, about audit logs and things. Either in an audit, or you’ve got to produce some audit logs for an audit, especially to continue to comply with those ISO certifications and SOC 2 regulations.

What kind of reports or audit logs does your tool have?

Tony: Every page in our system gives you an ability to create a report. So you can email or create an Excel report that’s for management from any page in the software.

You have instantaneous reports that you can actually do. You can do an Excel or PDF from anywhere.

Then, because it’s a workflow tool, your whole approval workflow is becoming a part of the audit. Your collection, which is the audit, we have all of that saved.

We keep all that together, and you can actually see it show up in your dashboard as either initiated, completed, or available to be completed every day.

You can see and search by date range and go back. We store forever, because once you start moving or destroying things in other areas, it may disappear. But we need to know where it was and what it was.

So we keep that in our log, and we can push that back to the DMS. We can push whatever you want. We can push the full data around, as you want to design it based on your firm.

But that’s really—we have a three-product approach on that.

People forget to capture until it’s too late. The approval process, people don’t focus on that. They don’t realize that when you start the approval process, especially on transfers, that you didn’t go through.

Our workflow is both simultaneous and sequential. A lot of people can be working on the transfer or a workflow at the same time.

If you’re not able to keep up with what you’re doing, just make sure that you keep that log and circulate that log to whoever’s going to be checking the logs.

You could get caught in an audit. “Jim did this to prove this, and it went out with just my approval,” and perhaps that’s probably not the best-case scenario.

It should have gone to Lee, and then Lee should have sent it to Jim.

Things like that, I think people haven’t heard yet, because I think a lot of it was done by email for the last 10 years. Now we have all of our systems doing it, and you’re just not used to the checklists and all that stuff being in your governance system.

It used to just be in something you built yourself, and then you would look at an Excel report.

Now we’re all starting to see your dashboard with the whole approval process: when it was approved, how it was approved, and what was changed.

You could substitute. Jim can say, “I’m going to send this to Lee because I’m not available.”

All of those different things need to be followed the way you set it up. So that’s something that we all have to be very cognizant of.

Jim: So you’re saying people could actually take vacations?

Tony: That’s a terrible word around here. Don’t say vacations.

Jim: Yeah, not to the CRO. You’re the revenue guy, right? So you don’t want anyone taking vacations.

Meeting Attorneys Where They Work

Lee: One of the questions that is on my mind, and I think on many people’s minds, is that we talked earlier about the DMS being the primary source of truth, but we’re needing to meet the attorneys where they are.

That means all of these other new solutions—collaboration tools, AI platforms, things of that nature—that are integral to their practice.

So, from a governance standpoint, what repositories and collaboration tools do you best cover today? DMS, Microsoft 365, Teams—

Integrations and Industry Collaboration

Host: File shares, iManage, however many others—and where do you feel you still have gaps?

Tony: We are pretty much all over all those that you mentioned. The only one you didn’t mention was part of you. But all those we integrate with—Box is a big one for us, those file shares and ShareFile. I mean, we still have the center to these for a lot of firms, and we also work with the Google side as well.

So, for our non-legal folks, that’s a big one. We have those integrations as well.

Where I would like to get better, I think, priority-wise, would be e-discovery tools. When you look at how great they all are at legal holds and walls—and I don’t know where that’s at—we’re good with that. But I would love to be in a lot of these different places, because you see some really, really fantastic custodian work in these products. I think we can do a little bit better on that.

What I’d love to see, just as an industry, is for us to talk to each other more. Instead of competing, we all do a lot of the same thing. Why don’t we just integrate with each other?

Stop switching everything out. You have to have 100% from me, and then if you don’t have me, you call them. And if you don’t have them, then you call another company. Just come up with some things that we could meet up with.

There are a lot of great tools out there. You both are active in the space. There are a lot of things that we can all do that don’t really become competitive. We should all be talking to each other.

Now, that’s my soapbox. I’m going to see that happen one day.

Host: Yeah, no, I would like to see that happen one day too. One of the things that takes me—I’m doing a little sidestep just for about 30 seconds—but what’s been coming up in a lot of the roundtable conversations is the idea of practitioners coming up with a framework: What do we need? What is it that we’re looking for? Kind of an acquired requirements document of sorts.

It’s a little bit too formal, but in a scenario like everybody’s working together, there’s transparency across what firms need and are looking for and what you guys are able to do.

Tony: Yeah, I’d be all for it. I think it’s something that we should really talk about.

There are some mega-tools out there. There’s one that we get all the time in this part of you, and every firm has that. We all need to talk to it. We’re all going to have to talk to Harvey at some point, or some other AI tool.

There are a lot of tools out there. It shouldn’t be completely specific. It should be worthless for specific. So we could figure out a meeting ground where we could all still sell our products, but also still have a way for you all, as end users, to work better. The whole industry gets pulled forward.

Host: Yeah, seems like we’ve got to have a follow-up to this conversation.

Tony: Yeah, that’s a whole workshop. I think they did make it.

People think I’m insane, and I’m sure it’s going to be so hated by me and all that. But that’s fine. I’ve got no loss.

Trusted AI and Automation

Host: So let’s talk about AI for a moment, Tony. What specific controls do you provide to enable trusted AI use—things like data minimization, permissioning, retention, even training models and things like that?

Tony: All those. But what we are really trying to do is limit it to very specific things. Like I said earlier, we’re in search mode for AI, like all of us. We start with the lead, and we’re going to link to vector.

So now we’re really coming up with very aggressive ways to introduce that. We have, in testing right now, a machine-learning tool for auto-classification.

You know, it’s one of those things—as I said, we find that our classification model works better with batching. We don’t have as many false positives as the machine-learning tool at this point. So we’re hoping to release that.

Right now, because of the way we match—we might have five sources that we’re looking at, as different than the workday of sources, to build that matching model—it really gets pretty tight. We’re not finding a lot of false negatives or false positives.

The machine, on its own, is not learning as fast. You just don’t want to have a 10% variance. So we’re still working through some of those areas.

It’s great for PII. It’s great for a lot of that stuff because it’s so easy to program. You could find a Social Security number, find a credit card number, dates, names, phone numbers. Those are very easy to get around.

But when you get to the content, you really have to understand what you’re reading. Comparative source contract law is very difficult. IP law is very difficult.

So if you let the machine learn alone, it could really just leave off a bunch of gaps because it all looks so similar until the machine’s really trained.

I’m looking forward to seeing where we go with that. Right now, our release model is matching. I know Kathleen doesn’t love that. She would love it to be all machine, but it’s also so expensive to spin up these engines. We can’t pass the loans. We’re trying to let the technology keep improving, and then we’ll release something.

But we’re plug-ready. That’s the thing that we can plug into your tools. We’ve had some firm tests that we would work with their companies.

Host: Yeah, that’s awesome.

Prompt Libraries and Cost Controls

Host: To follow up on the AI conversation—and prompts—let me get your thoughts. We asked Nick the same question, because there’s this thought about prompt libraries and things like that. This isn’t specific to our industry. Do you have to preserve prompts and outputs and all that stuff?

Tony: I think you should. I don’t know if you’d have to, but I think it should be controlled by you. I don’t think it’s something that should be outside of your walls.

It should be a different AI sandbox, and I think the infrastructure is such a big spend when you have it controlled by a big company.

We were doing a project for somebody a couple of weeks ago. We had to list about 60 terabytes on a weekend, and they would see the spikes in the performance that was required to get through that in a place like Tokyo. It’s a price.

If you’re not careful, you’ll start to see bills for usage that are really outrageous. Whereas if you control it yourself, you actually know how to monitor it. You can govern it. You can actually say you can keep your prompts, and you can do a lot of different things.

I think it’s a collaboration we have to get better at, because no one wants to see a variable data bill—a bill coming in after the fact for $15,000 for what you spent trying to onboard someone.

These are all things that the industry is just adapting to as we go.

Host: Yeah, I agree. I think the closed models are where it’s at. You get to see your Azure usage, your OpenAI usage, your tool usage, instead of having somebody else host the hosting and all that stuff.

We’re definitely figuring out what the best practices are there. Some great companies out there do real work, so they’re going to have a playbook that we all should follow.

Tony: One hundred percent.

Start Small and Dig In

Host: So here’s our final question for today for you, Tony. One sentence or two: What should listeners do in the next 30 days to improve IG outcomes?

Tony: Sign up.

Host: What else are they doing?

Tony: No, no. Seriously.

Host: I was expecting that. I like it.

Tony: You know, I’m going to actually allude to something like that. I think everyone said the same thing: It’s not possible to do the whole firm at once.

It starts small and big. It’s all about user adoption. It’s all about success and proven success. Owners in the firm will be your champions, and they’ll take it from there.

One matter is not too small, and one practice group is not too big. So whatever works for your culture, and you can get results—all these tools, anything you use, will tell you what’s in what drive. Anything you use will tell you what’s a DM. You can actually come up with real things.

So it’s about attention and disposition, and then coming up with a plan that you can pilot or execute.

Host: I like it. You’re coming up with it.

Host: Yeah, listen, I like that. I like that tagline: “Start small and dig.” I think that’s good.

Tony: Yeah, that’s good.

Host: Did you use AI to come up with that, Tony?

Tony: Tony. You know, Tony.

Closing

Host: Tony, thanks for joining us today. Great hearing everything you had to share.

Tony: Yeah, absolutely.

Host: Why don’t you make sure that’s rough, too?

Host: Thanks for being such a good sport. We love it. We’re busy, for sure.

Host: Well, you are. All right. Good luck. Have a great rest of your day, all right?

Tony: Yeah. Bye.