Chris Dreyer:
Many firms are getting comfortable using AI for the back office. You might trust it to draft a blog post. You might even trust it to summarize a deposition, but do you trust AI enough to put it on the front lines? Are you willing to let it handle the very first contact with the potential client?
David Ellis:
Experience for the customer that wasn't just human-like, but maybe even better than human-like in some ways.
Chris Dreyer:
We're talking about agents that can reason, negotiate, and believe it or not, emphasize with potential clients.
David Ellis:
It's funny that you bring up EQ because there's been several tests published lately and we've done our own internal tests and the AI is regularly beating humans now on emotional intelligence.
Chris Dreyer:
That might sound like science fiction, but real life practices are already using AI intake agents who never sleep. This is Personal Injury Mastermind, powered by rankings.io, the elite performance marketing agency for law firms. Today, we're speaking to David Ellis, the founder and CEO of Neto. We discuss how AI solves the capacity problem, the gold and old, leads you're ignoring, and the cost and efficiency that could save your practice six figures a year. I'm Chris Dreyer. Let's get into it.
I guess just out of the gate, before we get into details, let's start by clearing some of the fog. Why do you think it's so important to have this conversation about AI agents now?
David Ellis:
Well, I mean, AI agents are going to change the world, and I think anybody who doesn't believe that is being somewhat naive. So I think it's important to figure out what that means for your firm, for your business, and not be scared of it, but figure out what the strategy is and know that even though it might not be quite what you want it to be today, it might be even as quick as next week because the technology is changing and growing so fast. So it is just responsible to have a strategy and to be keeping track of what is going on with the technology.
Chris Dreyer:
I have started ... Talk to me about your product. Let's set the tone and then I'll dive into questions about it.
David Ellis:
Yeah. I mean, it might help to know a little bit of my background because it kind of tells the story. Yeah, I exited a business that I was one of the owners of four years ago, and we were a communications management company from a good portion of the Global 1000. So we managed phone systems, contact centers, conference room estates. Left that business and asked the question, "Well, how do I take what I already know and all the experience I have, but pair it with something super cutting edge?" And so AI was the obvious choice. So what would happen if I could build a call center and a contact center and have no humans and just have AI do it? And obviously we're not quite there yet. We don't need no humans, but the idea is how do we supplement, augment, reduce the amount of human capital that you need and work towards a world where humans get to do what they need to do and not what they have to do, right? So that's what we're building.
Chris Dreyer:
Talk to me about that, the tone, the empathy, because then I've heard, and I think even on your site, it talks about the depth of your studying and how much you've reinforced the empathy component. I know in my background, not on the call side, but on the live chat side, that was some of my frustrations when I would look at these chats and be like, "Oh, that question right after that one? Can we say, Hey, I hope everything's okay? Can we take a moment?" Talk to me about that side, the empathy, the EQ from the robot EQ, I guess.
David Ellis:
Yeah. It's funny that you bring up EQ because there's been several tests published lately and we've done our own internal tests and the AI is regularly beating humans now on emotional intelligence tests.
Chris Dreyer:
Wow.
David Ellis:
Is it really emotionally intelligent? No, but it emulates it incredibly well. It's a really good faker. And I mean, that's quite frankly what we need it to do. But yeah, when we started the company three something years ago, one of the questions we asked was, is there a way that we can build AI agents that emulate all of the best parts of the human nature and the human conversation and can we remove all the worst parts of it? So can we leave the empathy, sympathy, compassion, kindness in the conversation, remove all of the self-interest, judgment, laziness, "I'm tired. I came into work with a hangover," out of the conversation. And would that deliver a experience for the customer that wasn't just human-like, but maybe even better than human-like in some ways?
So we weren't trying to recreate a human-like conversation. We were trying to create something that actually was maybe even a better experience. And so for three years, we've been on this journey of how do we build empathy and compassion and understanding into these agents. It's been a really fun journey and I think we've done it really well. We're handling conversations in a lot of different industries actually, in everything from, "Hey, do you want to put a new roof on your house?" To, "I'm an addict and I just had a really bad high and I need some help and I need to get into treatment." We have AI agents that are actually handling those type of conversations with incredible empathy and understanding and encouragement, but without any of the judgment. So we're actually seeing humans in a situation where they're more comfortable sometimes talking to the AI about these situations than they are a human because the AI doesn't have the judgment. They're not embarrassed to talk to an AI about the fact that they just got high.
And so it's really interesting to see where the world is going. I think that humans are going to be able to practice their human experience with AI in the future. Kind of like a doctor doesn't go practice their medicine on a live human in the beginning. They practice on dolls or cadavers, right? It's going to be the same way. I think humans are going to be able to practice the human experience with an AI and it'll actually improve their human experience. A little bit of a rabbit hole there.
Chris Dreyer:
Yeah, no, I think it's so important. I read John Malone's Cable Cowboy and he was talking about like they didn't invest anything in customer service because they didn't need to. They were the only option, right? PI, they could go to another firm immediately after if they don't have the right experience. Talk to me about the product and how it could help PI firms.
David Ellis:
Yeah. The product is really simple. I mean, we build agents that handle inbound, outbound calling, inbound, outbound texting and email. So any sort of communication you need to have with your customers that require somebody to get on a phone or send a text message. Primarily, it's being used for simple reception duties and then very advanced intake duties. So the AI is able to answer the call, get the pertinent information from the customer, get all of the information about the case, whether it's a single incident or a mass tort case, whatever that is, and then actually go through and qualify and reason and decide, this is a case we take, this is a case we don't take, this is a case we refer. Maybe this is a case we don't usually take, but the case is so solid, we're going to sign it anyways. The AI can make all those determinations in real time on the call while updating your systems, just like a human agent will.
And then even to the point of sending out representation agreements, walking the customer through that, getting it signed, setting up next steps, it's really quite intuitive, speaks a bunch of languages, knows what contract to send out for what case. If you're referring, it could send out a referral contract with another law firm's signature on it, whatever you want. It's very intuitive, very deep. And it talks very well. It talks well enough is what I would say. Right now, AI is a game of compromises. So we're always trying to find what's the best voice quality with the most accuracy combined with the cost to run it. There's all these compromises we're making to deliver a full product. And I think we've done a really good job of nailing it for where technology is and where the requirements are for the law firms.
Why most personal injury leads go cold due to limited follow-up depth—not lack of intent
Chris Dreyer:
There's a handful of companies that do the inbound, right? The logic, the walk it down the tree, and most of the time it's their pre-qual and they'll do a live transfer. And if they don't connect, they'll take down the information is what I've seen for a handful of companies.
I want to stop the conversation right here and just give everybody a bit of a warning. This conversation is going to go deep into the weeds for a few minutes. If you know me or listen to the show, you know I love kicking the tires on a system, see if it actually holds up under pressure. David is about to break down the specific mechanics of the chase from lead to sign contract. It goes far beyond following up. David describes the process here that is so granular and so persistent that it would be unrealistic and very expensive for a human team to do, but with AI agents, suddenly almost anything is possible.
So talk to me, I guess, first on the outbound component. Do you tie it into a form where it can auto dial kind of like an Amazon Connect outbound? How does that function?
David Ellis:
Yeah, so it can function a couple different ways. So let's say we get an inbound lead, somebody fills out a form on a website that goes into your case management system or your lead docket, whatever that the CRM is, and then it will immediately trigger Neto to do an outbound call. So you get a speed to lead component where somebody fills out a form and within seconds you're calling them, which is, as you know, whoever contacts first is most likely to win that case, right? But then the other really powerful thing is using it for chasing leads that you're trying to get ahold of and get them signed. Most humans give up after one to three contacts. Most people sign after six to 12 contacts, and we're actually seeing a ton of signings happen after 30 to 40 contacts right now. And most law firms that we're finding, 15 is on the far end of how many times they're reaching out.
David Ellis:
So our AI will follow up. But the other cool thing about the follow up with the AI is that it learns when the best time to follow up is. So whereas humans usually come in and your call center staff, first thing in the morning at 9:00 AM, they go through their call sheet of leads they've got to follow up on, right? And every day they call at 9:00 AM and every day at 9:00 AM that person's dropping their kids off at school and they're not answering the phone. So our AI will call them at 9:00 AM, then it will call them at 2:00 PM, then it will call them at 7:00 PM and then the next day it will alternate that again and it's hunting and pecking and then also learning from other data on when the best time to call certain demographics are actually. So we're getting a very high call through rate and getting people on the phone that the call centers aren't getting on the phone.
Chris Dreyer:
They pre-qual statutes are good. Injuries probably don't do liability at this point. You're probably hooking them out of the gate, but let's just say you're going to sign that engagement letter, whatever you want to call it, the retainer. Is it then done? The reason I'm asking this question is the fall off rate is a very important percentage. So you take a call at 4:00 PM on a Friday and you sign the engagement letter, but you wait until Monday to introduce them to the lawyer. They've already got the new firm. So how do you translate that to reduce fall off rate or is that maybe like down in a future thing to try to solve? Talk to me about that.
David Ellis:
No, it's not. I mean, we do leave it up to the firm. I mean, every firm has their own process. We encourage firms to give a concrete answer of what's next. And somebody will call you in the next 48 hours is not a concrete answer. So the best thing we do is actually, my favorite is to schedule a video conference follow up or schedule a phone call follow up at worst, but I think people want a definitive. They don't want to be waiting sometime in the next 48 hours. They want to know, "I have a serious incident here. I need somebody to help me. I'm looking for a lawyer to be my advocate. I want to know they're going to call me tomorrow at 12:45."
So our goal is to get firms to schedule a concrete follow up next step that you can count on. You can put it in your calendar. We think that's the best way to limit fall off, but some firms do just go, "Yeah, we'll call them back in the next whatever," and we do that as well.
Chris Dreyer:
I'm trying to find as many intake problems as come top of the dome. Let's say you're going through the pre-qualification setup and box truck, commercial policy. Do you have an escalation, a barge in team? Do you have bells and sirens and the text messages go out to certain people? When you're talking about those A level S-tier cases, what happens there?
David Ellis:
Yeah. Again, up to the individual client. Most clients will have us do an escalation. If there's a death or a commercial policy, we'll immediately transfer it to the call center for humans to take over. Or a smaller firm, we might ring the managing partner's cell phone in the middle of the night. We can do all of those things. It really depends on what you want to do. We do find that the longer firms use the tool, the more comfortable they get with it and they start having it escalate less because the escalation's actually just another kind of stumbling block in the process. If we know it's a good case, let's just sign it. Why would we transfer it at that point? So it really is up to the individual, but I would say the fewer things, the transfers, handoffs we can have in the process, usually the better off in the customer experience. But it's really up to the firm.
Chris Dreyer:
Okay. So say they do sign, it's pre-qual they sign, right? Great. What about the answering the objection of, "Hey, can I speak to a lawyer?" Is that just another process of like, hey, then it goes to lawyer, or can you set it up to try to rebuttal that?
David Ellis:
Yes, and a lot of people do ask to speak to a lawyer. And so if we don't have any information yet, we usually try to say, "Hey, I'd love to get you to a lawyer. I need to ask you a few questions first and gather some information and then I'll be happy to transfer you." Now, if we found out it's a commercial policy or a death case or something, a really big case, then yes, the AI knows to treat that more sensitively. And if they say, "Can I talk to a person or a lawyer?" The AI is going to transfer that immediately. It gets more lenient with its rules when it's a really high quality case.
Chris Dreyer:
Okay. So you got your chase sequence, you do the auto dial, you don't connect and you try maybe 30, 40 times. Is it based on those follow-up sequences based upon the data of like, "Hey, day one, our connection rate, we're most likely to sign, we're doing five contacts, a pre-text, a email, la Facebook Messenger." What does that kind of look like and how has that evolved on the chase?
How AI agents schedule outreach, adapt timing, and boost contact rates across intake
David Ellis:
Yeah. So we usually do a sequence of, we call, we leave a voicemail and the AI generates dynamic voicemail, so it's not leaving the same voicemail every time. And it's custom to that. If we have information in the case management system, it's pulling it and going, "I'm calling you about this type of case that happened on this date." We're making it very personal and we're randomizing that. And then if we leave a voicemail, then we're usually sending a text message if we're allowed to because if they have opt-in for that, we send a text message that says, "Hey, I just shot you a call, didn't get you, left a voicemail. Either call me back or text me back on this number." Just that type of sequence. And then we can do emails too, but quite frankly, the email pull through rate is so low and in my personal opinion, it's almost not worth it anymore.
Chris Dreyer:
Delivery is so bad now.
David Ellis:
It's just really bad. But what we're finding, we have one ... I'll give you a story of a firm that we've had on for a couple of months now. They're using their humans to reach out to the potential cases 15 times, and then after 15 times they give it to us to do. So we're getting the garbage, right? Last month on 500 follow-ups that they weren't able to get ahold of after 15 times, we got ahold of 40% of them and we signed 15 cases out of that 500.
Chris Dreyer:
Wow.
David Ellis:
So that's a lot of captured value for a firm. Now, let's say even 50% of those end up falling off. That's still seven new cases out of data that you had just deemed as garbage. And it's just the AI is pleasantly persistent. It just keeps going until the person tells us to go away. And we're finding a lot of what we refer to in our offices as gold and old.
What happens to conversion rates when personal injury intake responds within seconds instead of minutes
Chris Dreyer:
I need to highlight what David just shared. He said they took 500 leads that a human team had burned through, the leads that were effectively marked as zero value on the balance sheet, then AI stepped in and signed 15 cases. If your average case fee is $15,000, that's $225,000 of found revenue straight out of the garbage can. This is the concept of capacity versus utilization. Think of your staff's time like inventory. With humans, you have to leave 40% of that inventory on the shelf in case the phone rings. If you utilize them 100%, you miss inbound calls.
AI breaks the rule. It has infinite inventory. It can work your old leads at 100% utilization and still has zero wait time for new callers. That is how you turn a cost center into a profit center. On top of that, these AI agents have the potential to help your intake team recognize cases that might have flown under the radar.
David Ellis:
And one of the things that I like about the AI is like just the capacity issue solved.
Chris Dreyer:
Does it allow you to sift through volume where maybe your human is only looking to ... You've been in a fender-bender, are you injured, type of deal?
David Ellis:
Yeah. It can definitely find fringe cases and find things that, again, like I think I said earlier, it can find cases that maybe it's a case you don't usually take, but it's a really good quality case. So it will do something with it, whether it's sign it or refer it. So we can disqualify, but we can also say, "Hey, this is a case that I'm going to have our team look at and then we'll get back to you." So it can do a referral to your own internal team.
So if it finds a case like that, yeah, it can do that. It might not have the ability to fully qualify it because we haven't given it a script or a training to do that, but it's smart enough to not just say, "Oh, we don't do that." We have some firms that, they're a PI firm, but they get a ton of family law calls for prenups and they might have a referral partner for that. So we can go, "Hey, we don't do that, but we have a referral partner. Let me have them reach out to you and we can send it out to referral." A lot of different options.
Chris Dreyer:
So the agents could do that in different jurisdictions, different practices, and maybe instead of just giving the bar ID number, be helpful to the consumer?
David Ellis:
Yeah. We have one firm that we refer out to over 200 partner law firms based on criteria.
Chris Dreyer:
Wow. I wish I knew that name.
David Ellis:
I won't tell you.
Chris Dreyer:
No, I'm not asking. I'm not asking. What are the weaknesses of the AI agent today that you said maybe in a week it solves versus the human component?
David Ellis:
I mean, yeah, the human component is obviously you have limited capacity, right? You have people that you can't hire, they don't show up to work, they call in sick. It's unpredictable. Humans are unpredictable. With AI, you have unlimited capacity. That's the beautiful thing. But the downside of the AI is it's still AI. It still isn't a sentient being that can make ... It's going to hit situations that it can't handle, and so it needs escalation.
Now, you could argue that humans need that too, because that's why we have management and things, but the AI probably needs it to a higher degree. And then the AI still doesn't talk like a human. I mean, if you push it hard enough, you will be able to tell that it is AI because it's a computer with a processor and processors are not totally consistent all the time. So you will get some hims and haws in the AI delivery. So it's still technology. It's still early. The good news is though, the AI that we have today, worst AI will ever have. So if it's good enough today, just imagine in six months or a year when it's going to be like.
Chris Dreyer:
What about the context window? Let's say we have a prospect that contacts in and they're a real case, but for whatever reason they don't sign and they're still within the statute and they contact later, 40 days from now, like do we have to go read through the process or can it remember that they talked to so-and-so with the injury in the rear end collision when they call back the next time?
David Ellis:
Yeah. That is one of the cool things that we've developed and you are using context window correctly. Think of context window as short-term memory and then your database or your CRM, your case management system is the AI's long-term memory. So the context window, we use that for short-term and then as it gets full, we actually have a method to drop out of the short-term memory because if the context window's full, the AI stops working. We have to drop, right? But then we're storing pertinent information. We have an infinite amount of data fields our system can store. So we could learn the person has a golden retriever named Sally and then use that in context on future calls. It's not in the context window, but it's in the database.
Chris Dreyer:
Okay. Very good. Very good. Core integrations, Lead Docket. What about Litify Salesforce?
David Ellis:
Litify Salesforce, Smart Advocate, Neos.
Chris Dreyer:
Probably have issues with Captora because of their platform?
David Ellis:
Yeah. I don't know that we've done any integrations with Captora right now. I mean, and I don't know all the integrations we've done, but I know Filevine, Lead Docket, Litify, Smart Advocate, Neos, I know we've done those. And then all of the document signing tools, DocuSign, Filevine, PandaDoc. I mean, anything you can imagine.
Chris Dreyer:
What about specific use cases like, do you have any firms use it just for local services ads?
David Ellis:
Yeah. Yeah. Local services ads. One of the things we're working on that we're really excited for is doing monthly medical follow-ups with customers, calling them saying, "Hey, let me ask you the five questions we need to know. Have you been to the doctor? Is there anything new?"
Chris Dreyer:
That is a really good one.
David Ellis:
Yeah. And we will continue to build out product extensions that as we talk to more firms and find out, "Hey, this is something that we just have to waste a lot of human time on and we want to automate it." And so we're even working on things like calling insurance companies, doing verifications. Some of those things are a little more complex and they're not quite ready for the market yet, but we will continue to develop and build out new types of agents that will blow your mind and actually save you a lot of time and money as well.
Chris Dreyer:
What's the pricing compared to your setup versus like the manual labor? What's the pricing arbitrage here?
David Ellis:
Yeah. So I would say it costs about the same amount to onboard our tool as it does to hire a human. The onboarding is $5,000, but we do everything for you. It's totally white glove. We build out all the integrations. We get into your management system with your permission and we do everything for you. So you don't have to have a tech team. If you have a tech team, we'll work with them. And then the agents are like reception agents, $750 a month and a intake agent, which is more complex as $1,500 a month. And then there's some small usage charges, per minute or per message, things like that. It ends up being if you have a full-time agent on the phone 160 hours a month, which is like a full-time person that never goes to the restroom or eats lunch, it ends up being under $3,500 for the agent, but it can handle 20 calls at one time, like what human can do that? So it has some advantages for sure.
Chris Dreyer:
Man, amazing. What questions should have I asked you that I didn't? Anything that we missed?
David Ellis:
One of the big questions that everybody asks is what the difference between a bot and an agent is. And that's where I always like to go because bot is a swear word in our office. Bots are frequently asked questions. They answer. When you start to talk about agents, the difference that really is starting to happen in the industry is the agents are actually able to make decisions, understand, and actually execute business processes. And so that's why we're very specific that we don't build chatbots or voice bots. We build agents that actually can do stuff. They can actually send out contracts, sign contracts, negotiate even. They have very deep skill sets that actually blows a lot of people away. And so being careful that the vernacular is very important in AI right now because people are just saying AI and meaning a hundred different things and it's not one word to catch all anymore.
Chris Dreyer:
David, this has been amazing. For our audience listening and wants to learn more or get in touch with you, has questions about the pod, what's the best way to get in touch?
David Ellis:
Hello@Neto.ci. Just shoot us an email there and we'll reach back out.
Chris Dreyer:
Amazing. David, thanks for coming on the show.
David Ellis:
Yeah, thanks for having me. Great to be here, Chris.
Chris Dreyer:
Unlimited capacity, 24/7 availability, and it handles 20 calls at once for the price of one staffer. I believe in proof over promises. The proof here is in the 15 cases David signed out of 500 dead leads. But remember, intake is only half the battle. An AI agent can't sign a case if the phone doesn't ring. If you're ready to fill that new intake system with high quality leads, come talk to us at rankings.io. I'm Chris Dreyer. Thanks for listening to Personal Injury Mastermind.
