Episode 367

Saam Mashhad

EP 367: Saam Mashhad on Operational Lag | AI Workflows


EP 367: Saam Mashhad on Operational Lag | AI Workflows

Operational lag drains profitability long before a case reaches demand. Every stalled file slows cash flow, lowers case value, and weakens the client experience. In this Toolkit episode, Saam Mashhad, Co-Founder & Chief of Product & Legal Operations at EvenUp, explains how AI-driven workflows replace bottlenecks with speed, accuracy, and consistency across the entire case lifecycle. From knowledge-graph extraction to smooth handoffs between pre-litigation and litigation, this conversation outlines how modern PI firms keep momentum high and outcomes strong.

How AI Workflows Speed Up Case Value and Firmwide Operations

  • Why faster operations directly increase case value and personal injury law firms competitiveness using AI-driven workflows
  • How data exposes delays, treatment gaps, and idle demands before they compound and lower case value for personal injury settlements
  • How pre-litigation and litigation teams can share context using AI workflows without repeating work or losing momentum

Learn more about AI Workflows:

Guest Details

Saam Mashhad is the Co-Founder and Chief of Product & Legal Operations at EvenUp, the AI platform used by 2,000+ firms to accelerate PI casework and outcomes. A former litigator, he focuses on faster resolutions, cleaner demands, and tighter continuity from intake through litigation. EvenUp has powered $10B+ in settlements and uncovered $400M+ in hidden case value — earning a spot on the Cloud 100 and a $2B valuation.

Chris Dreyer and Rankings.io Details

Chris Dreyer is the CEO and founder of Rankings.io, the elite law firm marketing experts - for all your digital and traditional needs. 

Transcript

Chris Dreyer:

This is a special Toolkit Tuesday, where we bring you the tools to sharpen your edge on the competition. I'm Chris Dreyer and this is Personal Injury Mastermind powered by Rankings.io. Today, we're talking about the silent killer of profit inside most PI firms, operational lag. Every day demand sits idle, you're losing momentum, margin, and client trust. To break it down, I'm joined by Saam Mashhad, co-founder and chief of product and legal operations at EvenUp, an AI-powered platform built to help firms close cases faster.

EvenUp works with over 2,000 law firms and has powered over $10 billion in settlements. They even uncovered $400 million of value hidden in case files. Now with a $2 billion valuation, they're redefining how personal injury firms move cases and maximize outcomes. Today, we'll dig into how smarter operations lead to 30% faster resolutions, higher settlements, and a practice that runs like a business built to win. Let's get into it.

Saam Mashhad:

The big challenge in personal injury and the reason why just using GPTs out of the box, notwithstanding the privacy and security issues, is that for you to be able to recreate a template in the context of a specific case, you need to be able to read the underlying documents, make sense of them, and then fill in the gaps. That's really the most challenging part of this entire equation. This process of extraction and organizing the information that is underpinning your case, that is the difficult technical issue that we solve for. We basically look at your data and extract it into that knowledge graph that we then use for these generations downstream that make it particularly helpful.

Chris Dreyer:

People will hear Elon Musk when he's talking about Grok, they always do a column for a context window. Is this part of that equation, like, "Hey, the context that it's pulling from is this whole knowledge base and not just a shorter time period that maybe some of these LLMs are restricted on"?

Saam Mashhad:

We're not in a place where simply increasing the context window is going to solve all of our problems. The way to think about model performance, it's a function of two things. It's a function of, it's called the number of input tokens. So if you input one page versus 500 pages, the model performance will degrade. The second variable is around the number of actions you are requiring the model to perform. So if I input one page into the model and ask it to extract the injuries, you can imagine that as being a high-performing task. If I input 500 pages into the model and ask it to extract every single date of service and the corresponding ICD codes and the corresponding treatments, that's when the performance degrades very, very rapidly.

Chris Dreyer:

You use the word tokens. Would you equate that to the same nomenclature as a token from a content perspective on, say, Google search instead of the words like the tokens and the relationships, or would that be different?

Saam Mashhad:

For our purposes, just think of it as letters or words. The more letters or words you input into the model for it to do things with or to understand and do downstream transformations, the harder of a time the model will have. And to be very specific, what the models are really, really good at, the large context window models, is the needle in the haystack problem. So if you feed it 500 pages and you ask it to find something very specific within that 500 page, it's quite good at doing that. But if you ask it to do an extraction and then ask it to create a relation between another entity that it's extracting, you're really asking it to do three things. You're asking it to extract two entities, store them in its memory, and then draw a relation between them.

The complexity of that task is much more complicated, and as a result, the accuracy diminishes. And when I say accuracy, I mean two things. The first is the classic discussion is around hallucination, so the model making up things. My impression is that that's less of an issue nowadays. What's more of an issue is completeness. So if you want to create a medical chronology, for example, you want to make sure that it's not missing any dates of service or if you want to calculate medical expenses, you want to make sure it's not missing any medical expenses. So it's how you define the system to make sure that doesn't miss things in an area of thousands of pages or many, many documents, that's what the frontier work is that we're doing.

 

Why faster operations directly increase case value and personal injury law firms competitiveness using AI-driven workflows

 

Chris Dreyer:

You work too hard and spend too much to let good cases stall out. What most firms don't realize is that every delay after intake compounds across the entire pipeline costing you and your clients money. When each lead costs thousands, every inefficiency eats straight in the margin, and the higher your acquisition cost, the sharper your operations needs to be.

Saam Mashhad:

How expensive the lead is, is going to very clearly correlate with the sophistications of the operations of the business, or of the law firm. Think of it this way, if I am paying $5,000 for a lead, then I need to make sure that I do what it takes to break even on that lead. And the things that I would do to break even on that lead would be very different than if my lead cost me $1,000 or $2,000.

What that means is there are things that very clearly will diminish the value of a case, like gaps in treatment and not making sure that the client is getting the care that they need and making sure that you're getting your settlements in a timely manner. These things start to really, really matter a lot more in very competitive environments. The more efficient law firm is going to be able to pay more for leads. So the operations of the law firm are directly related to how much of their total budget they can pay on acquiring new leads, and that fundamentally changes the market dynamics.

Chris Dreyer:

So if the CAC, the cost to acquire the case, is higher than the legal work to get the value, it has to be better because otherwise they're underwater. But if they're getting the cheap leads, maybe on the front end, maybe sometimes they don't work up the case and they leave value on the table?

Saam Mashhad:

That's exactly right.

Chris Dreyer:

I've had this hypothesis of the prices getting beat down on the pre-lit and tech plays into here, but putting more emphasis on the trial skill to get extract maximum value from the asynchronous side of things to level out the overall CAC, and that kind of reinforces what you're saying as well.

Saam Mashhad:

We have this multiple called the settlement multiplier. The settlement multiplier is basically how much settlement dollars are you getting for every dollar of bills that you produce? Historically, we would value cases by multiplying the specials by 2, 3X. First of all, if we look at settlement and verdict data, these heuristics of 2 to 3X are nowhere to be found, but they remain very popular heuristics. But what we did find is that firms that are willing to litigate cases and move cases into litigation more frequently will have a higher settlement multiplier. If you have a higher settlement multiplier, it means for every dollar of bill you are recovering a higher settlement. And the gap there is in the non-economic damages.

So basically the insurance company is looking at your law firm and saying, "Okay, they produce $1 of a bill, but I respect this firm. I understand that they're willing to stand behind their cases and litigate them, therefore, I'm going to pay more non-economic damages." Whereas firms that never litigate cases will have a lower settlement multiplier. Some are even below one, meaning that they don't even get the totality of the medical bills that they're sending to the insurance company. And it is very important to litigate cases or at least a subset of cases to make sure that the insurance company understands that you are willing to stand behind your cases. The hesitation usually stems from the need to file a complaint and doing all the discovery work, but if the cost of doing that is significantly diminished, then you should not hesitate to move into litigation and therefore increase the value of your cases.

Chris Dreyer:

Let's say you're a litigating firm, maybe you fur out all the pre-lit low value stuff and you only do the big cases because then your settlement multiplier is off the charts when this insurance company goes to look, from a data perspective, what they're going to do on the payouts. Some firms are pre-lit, okay, then whatever, they can't get the number, so then they move it to the litigation department. But from just a total data perspective, the insurance company is going to look at it as a whole. So wouldn't it be better to just offload the pre-lit and have a better multiplier on less cases?

Saam Mashhad:

There are certain firms that will focus only on catastrophic cases and their sum and multipliers will be through the roof. What you really want to be able to do is you are maximizing your gross dollars. And the best way to maximize your gross dollars is by having a balance of pre-litigation cases that never go to trial, but having enough litigation cases such that your settlement multiplier on the pre-litigation cases increases.

Now, a lot of firms don't want to do this because running a sophisticated pre-trial practice is very complicated. Some trial attorneys, they want to do litigation work, if that makes complete sense. They want to focus entirely on catastrophic cases, and that's fine. But if you're interested in really increasing your overall return, then you want to be able to get that premium that you're getting on your litigation and apply some of it to your pre-litigation inventory to maximize your overall return. And the second thing I'll say is, it's not just maximizing overall return, it's also diminishing the time on desk.

Chris Dreyer:

I think that's something that's not talked about enough, the cash acceleration formula. One thing that comes to mind once you start peeking outside of auto and maybe premises is the expanded complexity from the operation side versus that's why a lot of these trial firms keep them lean and mean in the pods. You got to be able to get it either in the pod system or on the assembly line to work it through properly, which I don't know the answer there, but I think what you're saying, that sounds right because then you get the cashflow from the pre-lit side.

 

How pre-litigation and litigation teams can share context using AI workflows without repeating work or losing momentum

 

Saam Mashhad:

That's right. And the thing that we see happening very often at firms, and I think one of the key areas of opportunity to explore, is making sure that there's great continuity between pre-lit and litigation. What we see often when we talk to the litigation team, it's almost as if they're taking on the case, doing the case all over again. They talk to the client, they ask the same questions, they do a full review of all the documents. And it's not only to get context on the case, which they need to, but it's really to redo some of the work that the pre-litigation side has done.

So our point of view is when the pre-litigation team reviews a medical record, produces a summary, generates any memorabilia that represents that case, it should be very easily usable by the litigation team such that they don't have to go and redo the case. The biggest hesitation in even taking a case to litigation, the cost for the team to ramp up. A lot of our product is focused on making it really easy for the user to understand what this case is about and to understand the details in great specificity and with support from the underlying documents very, very quickly. The most important aspects of any AI tool is not beyond accuracy and completeness, but how much time am I spending reviewing this output? That's ultimately one of the most important KPIs that you need to be controlling for, "How quickly am I getting from no context to context to getting my document out the door?"

Chris Dreyer:

Let's put some real numbers behind it. Every firm says they're efficient, but the data tells a different story. When you start tracking timelines and treatment gaps across thousands of cases, the patterns get clear fast, and that's where the opportunity really shows up, in the numbers.

 

How data exposes delays, treatment gaps, and idle demands before they compound and lower case value for personal injury settlements

 

Saam Mashhad:

30% is the happy number. We've seen 30% higher outcomes and 30% faster case resolution. So maybe taking a step back, the problem that we frequently see is not only in delays, one interesting statistic there is on average, 17% of cases that have more than a 30-day gap in treatment, and we're actually revising this number right now with new data and it's actually higher than that, even in the first six months of care, there are 42% of cases where the delay between the last date of treatment and the demand being sent out is in excess of 100 days. These issues apply across all cases, even small cases. So even in a small case, if you know when treatment is complete, you go and request the records and you automatically create a demand, that could save you one, two months of time from your overall case duration. And that could have a profound business impact on your law firm considering what your carrying costs might be and the value of getting cash early to your law firm.

Chris Dreyer:

So the staffing on the back end, you got your case managers, you got your paralegals, you got your different roles, right? Broad strokes, where does EvenUp fit?

Saam Mashhad:

If you think about the value proposition, we're trying to make the increased outcome of cases reduce duration, and there are very important elements in that that are controlled by different personas of the law firm. We're really end-to-end, and now with Mirror Mode, litigators as well are making use of the platform to help them get through discovery and emotions practice in a more streamlined manner as well. And that really is the core value proposition of the platform, being one platform for all the users and creating that continuity throughout the life cycle of the claim from pre-litigation to litigation.

One of the problems that I am really a little bit unhealthily obsessed by is care management. I think case managers are probably the most important constituents of the personal injury space because of how critical they are to the pre-litigation process. And I think they have incredibly difficult jobs with sometimes unreasonable expectations levied against them. We're expecting them to handle north of 100 cases. We're expecting them to get a bunch of inbound calls from clients asking where they're on on the settlement or having questions about the case. We're expecting them to reach out and call adjusters, call providers, negotiate liens, Medicare, sometimes write demands and request records. It just becomes really, really unruly.

And I think the big challenge now is if you go into the CRM of any law firm and you want to understand, "Is this case on track or is it not on track?" it's fundamentally impossible to do unless you go and audit the file. The thing that we're really focused on is, number one, making it really easy to understand how the case is going from a treatment and from a staff intervention perspective in less than five seconds.

Chris Dreyer:

On the agency side and the SaaS side, we're familiar with client success managers, right? It's very familiar in SaaS, and you onboard a client and you have an onboarding period where they have to get familiar with how to use your tool, and then it goes to maybe a different individual. In my business, we have a client success manager for the first 90 days to get the logins, get everything, and then it goes to an account manager for the continuation.

Why don't firms have a client success manager from the onboarding to get them started, get them treated, blah, blah, blah, get it rolling, and then it goes to the case manager on the ongoing? Have you seen that segmentation from a firm perspective? Why doesn't that happen? Why is the case manager just supposed to manage 200 clients and you don't have an onboarding segment?

Saam Mashhad:

The more steps you add to the process, the more complicated it becomes. And I think there is already such a need and desire to hire great case managers, and the skill set of a person that would do that initial onboarding and that initial setup is very similar to that of a great case manager. I suspect that that's why they tend to be bundled together oftentimes.

Chris Dreyer:

There's a lot of overlapping on the account manager, client success manager, like our agency, but I still see value in that first 90 days because it's so critical, especially if you're a lawyer, like the treatment and setting the expectations from a client service perspective, even from a review and reputation securing so they don't have buyer's remorse and then they go sign with another firm. Anyways, lawyers, why are you not hiring client success managers and you only have case managers? I don't know.

Saam Mashhad:

I think for law firms, this client success and this setting of the appropriate expectations happens from the very first conversation. So if we look at successful intake conversations, the intake specialist is not just taking in the questions, they're also building rapport with a potential client. These kinds of interactions are really, really critical to make sure that the case goes well thereafter and your law firm doesn't get dropped and the client responds to your second call to them. That's why we're really, really careful about attempting certain elements of the life cycle with AI capabilities out of the gate.

Chris Dreyer:

This has been awesome. For our audience listening who has more questions for you or wants to learn about EvenUp, what's the best way to get in touch or where would you like to send them?

Saam Mashhad:

You can visit our website, evenuplaw.com. You can also send me an email at saam@evenuplaw.com as well. I'm more than happy to have a conversation with anyone.

Chris Dreyer:

I'm Chris Dreyer, CEO of Rankings.io. From SEO to AI search, we help elite personal injury firms stay top of mind and first in the search results. If your firm's ready to move faster and scale smarter, visit Rankings.io. Let's build the kind of marketing machine that never leaks a lead.

Expand to read