Watch the YouTube version of this episode HERE

In this episode of the Maximum Lawyer Podcast, Tyson sits down with Wouter IJgosse, founder of DecisionVault, to talk about one of the biggest hidden time drains inside law firms: collecting client information.

Wouter shares how watching his wife manage an estate planning client with 37 beneficiaries exposed just how inefficient traditional intake can be. From information scattered across PDFs, emails, and meeting notes to endless copy-and-paste work, they break down how better workflows can save time, reduce errors, and create a smoother client experience.

They also dig into where AI actually helps, where the hype falls short, why all-in-one legal software may not be the answer, and how law firms can start thinking more like product companies when building their systems.

Listen in to hear how Wouter is rethinking client intake, AI, and legal tech to create better workflows for both law firms and their clients.

What You’ll Learn

• How a client with 37 beneficiaries exposed major inefficiencies in traditional legal intake.
• Why integrations do not solve the problem if your firm cannot collect structured client data in the first place.
• Where AI can actually improve law firm workflows and where it may create more work.
• Why Wouter believes the “all-in-one” software platform is often the wrong goal.
• How better client data can reduce double entry, errors, and time spent chasing information.
• Why AI still needs humans to create the instructions, guardrails, and final review.

Timestamps

00:00 — How 37 beneficiaries led to the idea behind DecisionVault

04:00 — The legal workflows lawyers accept simply because “that’s how it’s done”

06:30 — How AI is speeding up software development and product experimentation

10:45 — The problem with trying to build an all-in-one legal platform

15:30 — What DecisionVault actually does for law firms and their clients

21:40 — Why intake is really a client experience, data, and workflow problem

23:35 — Automating client follow-up and reducing the need to chase information

29:25 — What’s next for DecisionVault, including family law and AI workflows

32:10 — Why AI-powered intake can sometimes be slower than a traditional form

34:20 — Using AI to extract information from documents and handwritten forms

37:00 — Automating estate planning diagrams and reducing manual data entry

39:50 — The future of AI, legal software, and human judgment

Connect with Wouter IJgosse:

Linkedin

Facebook

🎟️ Get your MaxLawCon tickets: maxlawcon.com

🔍 Vet your vendors: beccaslist.co

Maximum Lawyer helps law firm owners build businesses, not jobs.

Resources:

• Join the Guild Membership
• Subscribe to the Maximum Lawyer Youtube Channel
• Follow us on Instagram
• Join the Facebook Group
• Follow the Facebook Page
• Follow us on LinkedIn

Full Transcript: 

Tyson Mutrux (00:00)
Today's guest is Wouter IJgosse, the founder of DecisionVault.

Wouter got into legal tech after watching his wife, who's an estate planning attorney, struggle through a client matter that made him think, there has to be a better way. I know a lot of us have had that same feeling before.

So today we're talking about reducing friction, creating a better client experience, where AI fits into client intake, and what every law firm can learn from thinking more like a product company.

Before we jump in, Wouter and the DecisionVault team will be joining us this October at MaxLawCon as one of our premier sponsors. So if you're coming to Atlanta in October, make sure you stop by and meet with them.

I can tell you, I just had a brief little conversation beforehand, and I can already tell Wouter's awesome. So I can't wait to actually interview him.

Wouter, welcome to the show.

Wouter IJgosse (00:46)
Yeah, thank you.

Tyson Mutrux (00:48)
Okay, so lots and lots of questions for you, but I don't actually want to start with the software just yet. We can talk about it. We have plenty of time. People can ask you about that out at MaxLawCon.

But I want to start with the story. I think the story part's important.

Your wife is an estate planning attorney, and one of her clients wanted to name 37 beneficiaries, which, I'm a personal injury attorney, but that sounds like a ton.

So walk me through that day. What did you see that made you think, man, this is crazy. There's got to be a better way to do this?

Wouter IJgosse (01:24)
Yeah, so it was kind of around the pandemic, and she had just shifted her practice to estate planning. One of the first cases that she took on was a single client who wanted to nominate 37 beneficiaries.

I've told the story in front of a live audience of estate planners, and you see a collective shrug go through the audience because everybody's like, man, that's painful.

Obviously, we can joke about it, but we need to build the plan that the client wants. That's a given.

But what this did was expose some of the inefficiencies in the process of an estate planning practice.

Across the board, one portion of practitioners drafts with their own templates, and another portion drafts with tools like Interactive Legal, WealthCounsel, or one of the other drafting platforms.

Those drafting platforms really help draft the trust, the ancillary documents, everything. But what they don't help with is, when a client wants 37 beneficiaries, getting all those names, gathering that information from the client, and then getting it into the software.

What this turned into was some of those folks living in a fillable PDF my wife was using. Some of that information lived in her meeting notes, and then a bunch more lived in back-and-forth emails with the client.

When it was time to draft, all of that had to be manually entered into her drafting platform to produce these documents.

Initially, I looked at that and thought, yeah, there must be a better way. But coming from a software background, I'd rather start with something off the shelf, so I tried all the different platforms at the time.

What I found was that while they have integrations, the real limitation was that the intake forms weren't smart enough to gather all this information to begin with.

The fact that there's an integration doesn't do anything for you if you cannot gather the information. You can't sync it over, so you still end up having to retype it.

That's where I started experimenting, working with a small group of estate planners and doing some proof of concepts. That got me into, okay, there is something here. Let's dive deeper.

Tyson Mutrux (03:55)
I always like when there's someone with an outside perspective who's able to jump in and look at things from a different viewpoint.

You're not a lawyer. You came from product management and sort of an engineering background.

What problems did you immediately see that we lawyers simply accept as, "That's just the way it's done"?

You talked about the data part of it. Was there anything else where you were thinking, why do you do things that way? That seems kind of silly.

Wouter IJgosse (04:26)
Yeah. When I tried to use, let's say, Clio, I tried to use Clio's intake forms and they didn't work. You cannot gather that many contacts. You cannot really do financial intake with them.

But when I talked to attorneys, a significant portion had the view of, "Well, I couldn't figure it out, but all the others must have figured it out. I'm just not tech savvy enough."

That's a funny take.

Other than that, some of it is enabling technology. As an example from a very different angle, once AI can talk, now AI can pick up your phone.

Before there's technology that can do those things, it's hard to even imagine it. Who's going to pick up the phone? We need a human for that, right?

There was a little bit of the same thing as I started experimenting and building some of the initial versions of this software. More became possible.

Then you layer on the next pieces. Let's gather financial information.

I built an initial version doing contact intake, uploading files, and asking other questions of the client. Then all the pilot users were like, "Yeah, but what about financial intake? Because that's really painful too."

Tyson Mutrux (05:56)
Sure. Intake in general can be pretty painful. Automating it is a step above what we're so used to.

With your background, how are you using AI now to help advance the tools you're creating?

I'm imagining that someone like you, who knows your way around code, is probably using AI quite a bit. It's allowing you to run where before you were probably just walking.

Wouter IJgosse (06:29)
Yeah. We're a software company, so we use it heavily to help with research into the tool. How is it built? Make a plan. Let's see what it's doing.

We're incrementally improving the software much faster than we were able to before because AI can help speed-run through our planning.

If I want to fix a small bug, I can dive in, but it's going to take me a minute to figure out how it worked again and what thing is calling what other thing.

AI can run through that, give me the three lines it wants to change, and with the background of knowing the tool and how it's built, I can see, yeah, that's fine. Go.

So it's speeding that up.

I've also done some experimentation with what kinds of AI features make sense to plug into this product.

There's a little bit of a debate between, do you use the AI that's outside to talk to the tool, or are we building an AI thing inside of the tool?

I haven't watched the replay yet, but Lawmatics recently did a demonstration of their new Merlin AI. They went down the path of putting AI into the tool and helping with getting more leads and selling to them.

I haven't quite figured out yet where we're going to engage with that, but that's more about using AI inside the product features.

Tyson Mutrux (08:08)
That is such a big decision.

Do I put it inside the AI, or do I put it through something like MCP and put it inside Claude? How do you do this?

You're almost predicting the way people are going to work in the next five years. That's what you're having to do.

That's a really, really tough decision. I don't know the answer to that.

Wouter IJgosse (08:32)
I think, at this point, it's probably both.

Tyson Mutrux (08:35)
Yeah.

Wouter IJgosse (08:35)
Because the builders and the people who are running experiments want to hook it up to their other AI workflows, and that probably makes sense.

You have things like MCP and APIs and all the connectivity to the outside world that an AI can hook into.

A small portion of the lawyer landscape will be able to do that and want to engage with that.

But what I think will happen is software vendors like ourselves will look at what people are building with this and say, great, let's take that and now do something internally.

For the tooling that Lawmatics, for instance, was building and releasing, it makes a ton of sense to put that inside the product.

For a tool like ours, it's more about connectivity. We're sitting between the other tools. We have about 25 different integrations, so we sit between all the different tools people use.

Some people use one of the big drafting platforms and Clio, which is a very common, straightforward approach, and we sync into that.

Others draft with their own templates, but they're happy to do the data gathering in a nice, structured way through our tools.

There are many different flavors.

The tooling is becoming very capable. Then it's really about, what kinds of workflows do you want to build?

The overtime aspect is interesting too because there's all this hype around the "SaaS apocalypse" and, "I can just ask Claude to build me HubSpot."

Yeah, sure. Maybe. I don't think so, but maybe.

But next week, one month from now, or two months from now, who's going to support that thing and make sure it keeps running?

It's probably worth it to just pay a couple hundred bucks and be done with that part, and only build the piece that's your IP around how your firm runs and how you build those workflows.

Tyson Mutrux (10:29)
Mm-hmm.

Wouter IJgosse (10:29)
You don't want to have to support all the different pieces of software.

Tyson Mutrux (10:40)
Sure.

I'm curious, how tempted are you?

You're in this in-between space. You see this thing built over here by Lawmatics and this thing built over here by another company.

How tempting is it, because you do have that expertise, to just go out and build this universal thing that fits everything?

Is that ever tempting to you?

Wouter IJgosse (11:05)
It's the siren song of the all-in-one.

It doesn't exist.

Well, it exists in some legal verticals.

To some extent, bankruptcy and immigration are examples where the information you need from a potential new client directly maps onto forms you need to file. You can have some mini case management system around it.

There are a couple of tools that do that entire chain, and that makes sense. You might still need something for marketing and sales, but it covers it pretty well.

In general, though, it's not a good idea to try to be the be-all, end-all, at least not from a software vendor perspective. It makes more sense to have good integrations.

Now, with some of the AI tooling that's coming out, it's becoming possible for people to skip a couple of tools.

But that's where it becomes important to say, okay, what do we focus on?

With DecisionVault, we're really good at gathering information from clients. Specifically, who are the contacts in this person's life? What are their assets and financials? What files do you need? What questions do they need to answer?

We're heavily focused on estate planning, but we're doing some experimentation now with family law.

The types of law that have to do with people and money are where we're targeting.

Then the question is, how do we have the tool do more?

Instead of saying, "We have smart forms and we integrate," it's, "Yeah, that's great. What can you do for me so I don't have to log in every day to figure out whether this client filled something in? What's the next step?"

That's where we're focusing with some AI and some things that are AI-inspired.

Some of those functionalities are much more rule-based. Did they fill it in? Let's follow up again.

You don't need AI for that. It just makes it expensive.

Tyson Mutrux (13:20)
Whenever you have a customer who's like, "Hey, I want this feature," and I bet you get this all the time, how do you filter those out?

How do you decide which ones should be applied and which ones should be skipped?

I'm sure that's got to be tough because you might have a customer who really wants this feature, but it may only apply to five percent of all users.

So it may not make sense to do it, but that might also mean you lose that customer.

How do you balance all that?

Wouter IJgosse (13:51)
Yeah, that's the art and science of product management.

Tyson Mutrux (13:55)
Sure.

Wouter IJgosse (13:55)
Who do you want to be?

Pretty early on, I drew the line that we're not going to be a case management system.

I don't want tasks. I don't want billing. I don't want time tracking. All of that belongs in a case management system.

In a way, some people are using DecisionVault as a mini CRM because it shows your list of clients and your open matters.

Okay, fine. They use it that way.

But I don't really want to compete and enter the arena where you have MyCase and Clio and all these big guys fighting it out. We don't need to go there.

We do something very unique that none of their intake forms can compete with. They're not at the same level.

I also don't really want to get into document automation.

We do a little bit now, but it's really a mail merge. If you want to do anything with conditional logic, formatting, or whatever, go buy something like Gavel or HotDocs or one of those tools we'll have integrations with.

But if you just want to fold some names into a Colorado statutory power of attorney form, great, we do that. You don't have to buy another piece of software.

There are a couple of these edges to the product.

Now we're focused on workflows and supporting those workflows, making sure the tool can help drive work forward rather than the paralegal having to keep a list of, "What do I still need to get from this client?"

Tyson Mutrux (15:28)
All right, I've worked around the edges of this enough now. Let's jump into what DecisionVault actually is.

Tell people what it is.

We've danced around it a little bit, but make your pitch. What is DecisionVault for estate planning attorneys and, eventually, family law?

Wouter IJgosse (15:48)
Yeah, we'll get into family law. We're not quite there yet, but it's coming.

DecisionVault helps with gathering information from clients.

The term "intake" is really broad. It goes from the first interaction with the client up to signing, or wherever you draw the line.

I read your book in the last couple of days. You were talking about intake and then needing to close, so I guess that whole piece is intake.

Tyson Mutrux (16:10)
Nice. Thanks.

Wouter IJgosse (16:11)
So I guess that's all part of intake.

Tyson Mutrux (16:16)
Well, it could be.

It's funny, for personal injury, sometimes our intake isn't just intake. We're still collecting things all the way through litigation too.

It's interesting how it can extend depending on what they give you and when you need it.

But most people consider intake to be from the moment they contact your office up to the point where they sign.

Wouter IJgosse (16:26)
Bingo.

Tyson Mutrux (16:26)
That's generally the way people view it.

Wouter IJgosse (16:40)
We don't really get involved in that initial piece.

If you want to run ads and get a lead, great. You can do that somewhere else.

What DecisionVault is really good at is the client data gathering.

As you said, it doesn't stop right before you meet with the client.

We have an online portal and questionnaire-type tool. You can build out multiple questionnaires for your different practice areas.

The tool is really good at gathering all the contacts. It does comprehensive intake of the client's financials and lets them answer other questions.

Then you can go and have a meeting with the client.

But what happens, just like you said with personal injury, is that you still need more stuff.

Guess what? In estate planning, you still need more stuff too.

Across the board, generally every intake form, from Clio Grow to MyCase or whatever, stops before you meet with the client for the first time.

It doesn't really work because some clients fill in a lot and other clients fill in a little. There doesn't seem to be any way in the world to move one closer to the other.

The attorney, especially in estate planning, sits down with the client, talks through their situation and what the plan design is going to be, and then you need homework.

I need this document. I need this deed. I need these other three relatives you talked about. I need their names, dates of birth, and emails.

There isn't really a good way when you're using off-the-shelf questionnaires and intake forms to send it back out and get more information.

DecisionVault helps with client data gathering across the lifecycle of the case.

You can even lock the client out when you go draft an estate plan because you don't want them messing around with it then.

Tyson Mutrux (18:35)
Sure. Yeah.

Wouter IJgosse (18:37)
That's the client perspective: gathering the information.

On the back end, you can push a button and sync it into your other legal tools.

That includes drafting platforms like WealthCounsel, Interactive Legal, and other platforms, as well as the well-known case management systems and a bunch of document automation tools.

Then there are things like APIs, Zapier, and potentially soon things like MCP. We're experimenting there.

On the client side, it's a much nicer experience than the alternatives: a fillable PDF, a case management form that feels clunky and can't gather all the information, or just having the paralegal do it over the phone or on paper.

It looks better for the client, and it makes the practice seem more tech-forward.

On the law firm side, it saves a ton of time.

We have a nice asset sheet that comes out in Excel. That's always a huge selling point with the paralegals.

The contacts will sync to the drafting platform, and there are a bunch of different ways we slice and dice the data.

We also have some tooling that helps the attorney prepare for the meeting. You can build a meeting template, pull in information from the questionnaire, and design it so you can either pull it up on your screen or print it before you walk into a client meeting.

Everything is there to document how we're going to draft this person's plan.

We build these stepping stones and follow-up steps that build on the questionnaire and the information.

Tyson Mutrux (20:19)
There's this diagram on your website that I really like.

It has DecisionVault at the beginning, and it's almost like the place where all that information flows into.

From there, it flows into the CRM, the case management system, and then into the drafting platform.

I think that's a good way of visualizing it because this is one of those things that's hard for people to visualize.

That's why the diagram is perfect.

It's sort of a filter, or maybe a funnel, where everything goes through it and then gets sorted.

Okay, this information goes here. Maybe this information goes here, here, and here.

You don't necessarily need this information in the CRM, but you need it in the case management system. You need this information in the drafting platform, but not necessarily in the case management system.

It's a really cool sorting system.

Where do you think the whole intake issue lands?

Is collecting information really an intake problem, or is it more of a user experience problem?

Does that make sense?

Wouter IJgosse (21:42)
It's both of those things and a bit of a workflow problem.

When I'm reading through your book, it talks about getting the decision maker out of all these things and above it so they don't have this founder gravity and can enable folks to do their work.

Great. How are they going to do their work? What does that work entail?

If you want to add more automation to the practice, what you'll find in practices such as estate planning, probate, and probably also family law, is that information about the client is pivotal.

If that information isn't structured in a way where it can easily flow from one system to another, it'll inhibit a lot of the automations you might want to build.

If I automatically email a fillable PDF to the client, that's great. It saves 30 seconds.

But then it comes back and I have to read it or copy and paste it over. That can take two hours to put into one system and then another system.

Client data is really a core piece of any kind of automation you're going to build in several of these practice areas.

It makes it possible to do more work with the same workforce and make the paralegals' work more enjoyable because they're focused on things that touch the client instead of copy-pasting things into another tool.

Tyson Mutrux (23:13)
Completely.

Double data entry is one of those things that just kills you.

You get things wrong, it increases errors, and it kills productivity.

What does DecisionVault do to help with chasing down information or the follow-up process?

Wouter IJgosse (23:37)
That's a new thing we've been investing in and started rolling out with some beta firms.

We realized that we're trying to stay away from CRM features, but having some kind of step-by-step set of stages that a matter goes through probably makes sense.

We've added something like that. It's not fully visible yet for everyone, but there are a bunch of firms that have access to it.

Now that we know what stages a matter goes through, we know whether we're in the invite stage and need to follow up with a client to get started, whether we're waiting for the meeting to happen, or whether we're in the homework stage.

We've been building functionality to say, okay, I need these two files and this data point.

Then the software will periodically email the client and follow up with them.

It can also give them a button if the client is annoyed and doesn't want to do it that way, and then report back to the firm that this client doesn't want to do it this way and you should follow up manually.

What we realized was that it was hard for firms to really follow whether someone had uploaded a particular file or entered the information.

First, make that visible on the matter.

Then we're going a step further and automatically following up once the firm selects what information they still need.

That starts driving everything forward.

It's like, when does the power company know your power is out? When you call them and tell them your power is out.

We're making it more automated.

Once we hook things like that into APIs and MCP, it becomes a very interesting mix.

We go through certain stages in DecisionVault, and that can flow out to other tools.

For instance, it could trigger that a matter is ready for drafting because we have all the information now.

That can go out to other tools.

Then you can start doing really interesting things with the information and hook the tool into even more workflows.

Tyson Mutrux (25:43)
It's interesting that you said you aren't a CRM because it reminds me a lot of Calendly.

Calendly isn't a CRM either. It's got a very specific thing that it does, but it has to have some CRM-like features, where you have the follow-up about the upcoming appointment and then the follow-up after the appointment if you add those things.

It's cool that you're starting to build those things in because I think that's something people will want, even though you're not a CRM platform.

The best place for those activities to come from would be DecisionVault as opposed to a dedicated CRM because that's the thing doing the work. That's the thing collecting the information.

I think it's a brilliant move.

I'm curious, how does it look from the client side?

The client gets the form or whatever communication they receive. What does it look like on the attorney side or the law firm side?

Is it like a client portal?

How does the client side look compared to the law firm side?

Wouter IJgosse (26:43)
The law firm owner side, or really the firm staff side, looks like your regular dashboard-type tool.

I have matters. I have client users. I can open a matter and get to all these different functionalities that I need.

The client side is much more focused.

You're in your questionnaire. Once you've finished it, you might get to a menu where it lists multiple questionnaires if you have access to more than one.

Maybe they've been doing estate planning with this attorney, and a couple of months later somebody passes away in the family, so they can do a probate case through the system as well. They have access to both.

For the firm staff, it looks like your regular software-as-a-service tool.

The client side is much more focused on, "Hey, this is the information we need from you."

One thing we plugged in a year or so ago was moving away from passwords.

Now it defaults to the client putting in their email. They get a link emailed to them, click the link, and it logs them in.

Clients would get confused: "Where's my password? I have to create another password."

Then, like everything around this tool, we have the opposite case where clients complain, "I need a password and I need two-factor authentication."

That's fine. We support those things too.

But the baseline is just put in an email and it logs you in.

We've had law firms complain, "Yeah, but I also use the MyCase portal, and now the client has two logins."

No, they don't. They just put in their email address and they're good to go.

Tyson Mutrux (28:27)
I actually like that.

I don't know if you're familiar with OnePager. They've since closed down, I think, but OnePager was really cool.

It was the same type of thing where I could send this one-page site to a client, and it would save whatever activity they were doing on it.

Let's say you had a checklist you wanted to send to a client for things to do.

Wouter IJgosse (28:47)
Yeah.

Tyson Mutrux (28:47)
They could do this thing, check it off, do this thing, check it off, watch videos on it.

The way they accessed it was through their email. There was no password.

I really like that.

You don't see a lot of platforms where you just use your email and then get a link emailed to you.

It saves a lot of headaches with passwords, so I think that's a really smart idea.

What plans do you have for DecisionVault that you can talk about?

Maybe there are some things you can't talk about yet, but I'm sure you probably have a roadmap.

What are some things on that roadmap that are coming up?

Wouter IJgosse (29:26)
We're interviewing a number of firms.

Some of our existing subscribers, like I mentioned, also do family law.

We have existing subscribers who do some estate planning and some family law, and we've gotten to the point where we have several mockups we're putting in front of them.

That's getting more concrete toward having a family law MVP.

I know some firms have been using DecisionVault for family law, but the honest look at it right now is that we're missing some of the information.

We're pretty close, especially on the financials.

But there are fields missing like, "Do you consider this a marital asset or not? Why? When did you acquire this?" Or, for each child, what's the five-year address history?

We have a lot of it. We're just missing that final 20 percent that would really make it helpful for family law.

Family law is definitely something we'll start going into during the rest of the year.

The second piece is thinking more toward workflows and how we can support the firm.

Instead of being a smart database, it's a tool that has information and does stuff for you.

That's the step we're taking.

That involves things like automatically following up with clients on missing data points, and there are some more ideas we have there.

I wrote all the code up until about a year and a half ago.

Over the last year, I've really been building out the product team, delivery team, and engineering team.

Now I have more free time because I'm not spending all of my time in the code, so I can explore.

Where do we fit in with these AI tools? What can I make happen?

I've done a bunch of technical proofs of concept.

That's where the technical angle is helpful. Okay, we'll spin this up and see where it ends up.

Some of the stuff you see, like the ABA had an article about the AI-native estate planning practice.

It was a great story, but it didn't hold any water because I tried it.

We have a questionnaire. What if I could just talk to ChatGPT and answer it all?

What I found was, man, that takes even longer than clicking through the form.

It's really boring and annoying.

Tyson Mutrux (32:08)
Really? Really?

That's interesting. Wow.

Wouter IJgosse (32:12)
Think about it.

If AI is going to ask you one question at a time for radio buttons, for instance, that's a lot of waiting on tokens and whatever.

I can just click, click, click, click, click, click, click, and I'm on to the next screen.

The thing that does help is writing out names and addresses and all of that. That's maybe okay.

But it just feels endless when you turn it into a chat conversation.

I also tried things with voice.

I have the perfect name to test this out. It will butcher my name seven ways on Sunday.

That's the other piece.

You hear things like, AI will interview them over the phone and figure out what question comes next.

Great.

If you let AI figure out what question comes next, you end up with some amorphous blob of information that's not structured and will not go into the tools you need next to draft these people's plans.

A lot of the AI stuff is, "Let's do AI drafting."

Great. If I let AI draft an 80-page estate plan, now I have to go review an 80-page estate plan to see what it's done.

Tyson Mutrux (33:21)
Sure, yeah.

Wouter IJgosse (33:22)
Right.

Some of these things have good sound bites, and I'm sure there will be applications that work out.

But they don't quite make it just yet when you try to actually apply some of them.

Tyson Mutrux (33:36)
For people listening, I'll embarrass myself a little bit.

Beforehand, I was asking you how to pronounce your name because I wanted to take a shot at it before you told me.

My guess was "Vuter Ishgase," I think is what I said.

I got the V part right, which I was proud of, but I got the rest of it completely wrong.

So you're completely right.

I actually don't want to shift gears yet because I want to ask you something about documents.

I have an idea for you, and you may not like it, but I'll give you the idea.

Whenever you're collecting the data, people are uploading documents at some point. Is that fair?

Wouter IJgosse (34:18)
Yes. Yes.

Tyson Mutrux (34:19)
Okay.

What I thought would be kind of cool is if you front-load the document upload part of this as an option, where it then uses AI to extract the information and autofill some of the things.

The person would still have to click a check mark to accept it or an X to not accept it and then type in their own answer.

It's a very basic idea, and you may have already thought through it, but I figured I'd throw that out there.

Wouter IJgosse (34:44)
Yeah, it's on our list of AI functionality to look into.

If I think through the platform as jobs to be done, the thing that it does really well is data entry. That's the job.

Either the paralegal sits there and writes it out, or the client sits there and types it in.

Obviously, the law firm has then outsourced it to the client, so we want to make sure the client wants to fill it in.

That's on our list of AI applications to explore.

Can we make it easier on the client?

Can we let them upload a financial statement and parse that into our structure already?

Or upload a prior trust and pull the names of the people out of there?

That's definitely something in our awareness.

Another one that we're closer on, because we have a technical proof of concept, is dealing with people who are unwilling or unable to do this on a computer.

For some law firms, 15 or 20 percent of folks want to do something on a piece of paper.

Great. Let's let them fill it out on paper, scan it in, and then have AI read it and fill in a DecisionVault questionnaire.

That's something we have working as a proof of concept, but we want to make sure it actually works at scale.

Tyson Mutrux (36:08)
Are you going to use something like a QR code or barcode when it scans in?

There's an AI case management system, and the name is escaping me, but when they request medical records and those records come back, there's a cover sheet.

They scan it in, it hits the barcode, and it automatically puts it in the file.

Are you going to do anything like that, or is that one of those things you're not going to mess around with?

Wouter IJgosse (36:36)
Yeah, we're still in the testing phase.

Something like that, yeah.

Tyson Mutrux (36:41)
Gotcha. Very cool.

I do want to shift gears here because I don't think you've mentioned this yet.

I've got it on my screen now for people who are just listening. This is the PowerPoint part of it, or diagrams, I'm sorry.

It says, "DecisionVault can help populate your PowerPoint diagrams with information gathered on the questionnaire."

What is this? This is kind of cool.

Wouter IJgosse (37:05)
About two years ago, at one of the conferences, the diagramming piece was all the rage.

You still see this in some of the tools that are out there.

Some tools are focused on estate planners and maybe a little bit on financial planners as well, letting them do estate planning.

Where that lands in terms of what's allowed, I'll leave that up to the group to decide.

A lot of those tools can load in estate plans and then visualize them.

People came to us and said, "Can you make me a diagram?"

The problem with those diagrams is that it's pretty easy to let AI figure it out and make you an image.

But if you have an image, you cannot edit an image.

Some estate planners have PowerPoints they use to explain to the client how the plan is supposed to work.

When this person dies, when that person dies, the assets go into this structure or this trust, et cetera.

What we did was say, hey, we can take your PowerPoint template with some placeholder values in there and do the kind of mail merge I was talking about earlier.

We do that in PowerPoint, and we do that in Word.

You can put the names of the client into that PowerPoint.

Again, it saves 20 minutes of the paralegal copy-pasting the client names and trust names, hopefully into the right spots.

It's very embarrassing if you have a diagram that has a prior client's information or a typo.

That's what the diagramming feature is.

There's no fancy AI with that.

It's really just taking your current templates.

This is relevant for some firms and not all firms, but there are plenty of firms that have their PowerPoint format.

You can build one template with two children, three children, four children, the different slides, and then fill that in with information.

Then you edit from there because it's PowerPoint.

You can drag things around and add to it because every estate plan will be a little different for each client.

Tyson Mutrux (39:21)
Very cool.

That's one of those things you hadn't mentioned, but I wanted to make sure we mentioned it because it seemed pretty cool.

DecisionVault is one of the premier sponsors at MaxLawCon this year, and I know you and your team are going to be there.

If someone stops by the booth at MaxLawCon, what kinds of conversations do you hope to have with people?

I think some people are a little hesitant to stop by some booths.

What sorts of conversations do you want to have with people out there?

Wouter IJgosse (39:52)
Obviously, I'd love to see the estate planners.

By then it's October, so we probably will have more to say about family law and how we might be able to help.

But I'd love to learn more about how people are incorporating AI into that flow.

What are the limitations? What things are working well? What limitations are they running into?

Piecing the tooling together is becoming a very interesting part.

I'm the perpetual product guy, so every solution will inevitably find its way back to product.

It's a little unclear where this will go, but one plausible direction is that most of the tools are becoming tools that can be used for a specific person.

People may spend part of their day chatting away with their tools or just having agents run stuff.

I don't think it gets to, "It just runs itself and we don't need the humans anymore."

That story was having a moment six months ago. I don't think that's working out that well.

Even when we run benchmarks and look at how good AI is, when you run that benchmark, they spend so much time building up all the instructions.

And guess who does that?

The human.

Then AI does its work, and then you review it again.

It becomes this sort of sandwich where AI definitely goes faster, goes deeper, and can more cheaply do the work, but the human element is still super important.

You need to prompt it the right way, set up the guardrails the right way, and give it some of the tools.

Claude could call out to DecisionVault and ask DecisionVault to reach out to a client and invite them to start the questionnaire.

DecisionVault can follow up to make sure the client has filled in the questionnaire.

Claude can check back to see, has this happened? Where are we with this?

Those interactions are definitely what we're playing with and trying to figure out where we fit in and how we can connect.

Tyson Mutrux (42:10)
I love the way you think about things.

That's very cool. I feel like I think about things the same way.

You can check out DecisionVault.com if you're interested.

DecisionVault.com, Facebook.com/DecisionVault, and you're also on LinkedIn at LinkedIn.com/company/DecisionVault.

If people want to reach out to you, what's the best way of getting in touch with you? Social media? Email?

Wouter IJgosse (42:35)
Yeah, I would say [email protected].

Tyson Mutrux (42:39)
Very good.

If you have any questions about DecisionVault, check them out at DecisionVault.com or [email protected].

Great to have you on. I can't wait to see you out in Atlanta.

Hopefully everyone listening will join us at MaxLawCon in Atlanta and make sure they stop by your booth.

Really appreciate it.