Pod Insights The Playbook With David Meltzer Making Facial Recognition Affordable and Practical About this episode In today’s episode, I sit down with Michael Doherty, founder and CEO of Verifyfaces, to break down how facial recognition can be affordable, fast, and practical without relying on expensive AI. We talk about his shift to facial geometry, turning faces into mathematical data that runs on existing cameras while cutting costs by up to 90%. Michael explains how this approach improves search speed, reduces operator fatigue, and enables real-time alerts for schools, retailers, monitoring centers, and domestic violence prevention. We also cover rare use cases like biometric-triggered emergency alerts and how smarter architecture can prevent incidents before they escalate.
Transcript 0:00 Speaker 00 Welcome to Entrepreneurs the Playbook. I'm David Meltzer and I'm here with Michael Doherty. He is the CEO and founder of Verified Faces. It's a space that's filled with competitors in those places because they don't have the technology and the ease of use in order to facilitate, to make it easy to identify people and who they are and their backgrounds. Michael, welcome to the playbook.
0:25 Speaker 01 Thank you very much, David. Pleasure to be here.
0:28 Speaker 00 Well, let's talk about the issues concerning faces. There's a lot of different people with a lot of different products and a lot of different technologies to use facial recognition. You are using Verified Faces specifically in a different way to help people identify problems before they happen and also to alert people of those problems.
0:56 Speaker 01 Yes, certainly. Well, firstly, in facial recognition, just about everybody uses AI. Now, the AI is very good, but it's also very expensive to run because of its data requirement. And in doing so, that really restricts the market. So you predominantly only see facial recognition now at airports, you know, high-end businesses like casinos and government applications. But in our line of business, we started this journey. We actually run a refuge. And there's a lot of vulnerable people in there. So we, to bring facial recognition down to a price where it could actually start to help people that are at domestic violence or are targeted by dealers,
1:40 Speaker 01 we had to develop a system that could be affordable, easily deployed and easily managed. So we pivoted early and we moved away from AI and we moved to facial geometry. So we actually turned the face into a mathematical equation, which is, turns out to be equally, if not as accurate. But it's easily deployed and it's 80 to 90% less in costs. So our mission to be able to deploy a system that would be able to work on any existing infrastructure, use your existing cameras, wires and hardware and use it, we've been able to achieve.
2:18 Speaker 00 And the searchability becomes an issue when you were talking about the processing and the data and the memory and the energy and the expense. When we're using something like Verify Faces, we have to search these huge databases every single time there's an identification, every time there's some data. How does your system work in creating efficiencies in just searching specific data in order to facilitate finding if that person is contained within that alerting system, for example?
2:56 Speaker 01 Yeah, the network architecture is quite unique. Now, if you look at a normal video system, it takes around about 55 frames per second. And if you've got several cameras, that's a lot of data going through. So an AI system will be searching every frame, and that can add up to tens of millions of frames depending on the site, which takes time and a lot of data to analyze. What we do is that we search three seconds of film, and we pick the best image out of that, and we crop that image so it's down to about 30 to 40 kilobits and create that as a timestamp. So just like you're searching then for a Google piece of data, you're searching for the image rather than searching through the whole library of millions of frames.
3:41 Speaker 01 That makes it quick and efficient.
3:44 Speaker 00 Yeah, I was gonna say, it's so smart to limit and to direct the searchability in order to create this great value without having a huge expense or time. And then furthermore, the applications, you know, I think one of the things when you were starting on GoFund Yourself that we had excitement for was the multitudes of applications that can be used, that can use Verify Faces. What are some of the key applications that you're using it for today, and what can it be used for?
4:20 Speaker 01 Yeah, well, because now it's so usable and economical, it opens up so many more markets than what people are just using it for. And for example, door control, access control, the ability to open a door with a face has been around for years, but it hasn't been widely accepted because of the the price that's involved. And we see an opening there for tens of millions of doors, you know, globally as it's retrofitted and new developments come on because they're looking at probably paying a dollar per month per key and through an application through ourselves. And the other big calling we're getting is through monitoring centres.
5:02 Speaker 01 With monitoring centres, they're all about reducing operator fatigue. Now, with gun control coming in and gun detection, the A I is getting very good. But what's happening is that the monitoring centres are getting hundreds, if not thousands of alarms a day, which have to be manually viewed and suppressed. And now what they do with us is that when they detect a gun, they also pick up the face and we'll say, no, that's a security guard that's registered to carry a gun, suppress the alarm or vice versa, it's not allowed. So that's a very big market there on itself. We're also been working now for a number of years.
5:44 Speaker 01 We're very proud to say we're the only company which is accredited to trigger a nine one one alert on a biometric match. So just as an alarm is triggered on a fire alarm or a duress button, it can now be triggered on a biometric match. So we see ourselves opening up big markets there for domestic violence when there's a court imposed restraining order. And we can integrate with the house or works existing video management system. And when the person that is the perpetrator turns up, it would send an alarm through to the police with a copy of the photo and also the source photo.
6:26 Speaker 01 And we've taken that one step further for incidences, say where there's a shooting on. We can also send first responders a floor plan, access controls to the fire entrances, et cetera. So things which are available, but now they'd be available before they actually arrive on the scene, rather than having to muster the information.
6:50 Speaker 02 Washington Gas has a 99.8% reliability rating. You know who else in the DMV has that kind of record?
6:59 Speaker 02 Nobody.
7:01 Speaker 02 We earned it by investing in system upgrades to make our network safer.
7:07 Speaker 02 Reliability is just one way Washington Gas is here for the DMV. Visit hereforthedmv.com for more.
7:20 Speaker 00 And we can use a limited database from local, state, or federal agencies to identify threats, perpetrators, and wanted or warranted people. And they can utilize specific data and people for specific locations. And then even on top of it, have an alert before there is an incident to be able to have that person apprehended, or at least question. What other governmental uses of the Verifi Faces are available today that our government could be using?
8:04 Speaker 01 We've got a large aquatic center in America at the moment, and they've actually put a database in of people that have got not allowed in the swimming pool area, pedophiles. So it sends an alert through to the sheriff's office.
8:20 Speaker 01 We've got an interesting use case where a courthouse has actually put it in, and they're saying when somebody's turning up to court, quite often their friends come with them. And in many instances, the friends have got outstanding warrants. So when they enter the courthouse, the security guards are able to greet them and do the relevant procedures.
8:45 Speaker 00 Wow, that's incredible. Now, on top of that, I imagine because the expense is less, that there's a lot of places that would like this beyond the door systems. Because the expense is so much less now and the capability is so great. I can imagine a lot of schools, whether it's via the same protective measures or even identifying weapons, now have a much more non-invasive and economical way to protect our children.
9:21 Speaker 01 Ah, yes, it's, you know, in those incidences, we stand alone because in schools, everyone's got tight budgets and they can deploy this in a matter of days. We also, in retail for loss prevention, we did trials with a spa retail store. And over a period of six weeks, they were able to reduce their retail loss by 40%. And one of the interesting factors that they found out is that they didn't have a lot of shoplifters. They just had the same people coming back all the time, operating in a gang form. And they do a thing called love bombing. So what they would do is that when they identified somebody had stolen, they would put the face into a gallery.
10:03 Speaker 01 Now, that gallery, you know, because they're changing staff and they're open long hours. And previously, they'd never be able, the person would always get through. So this way, we'd take the store manager. The store manager would go up to them and say, great to see you back again. Lovely to have you. Now, I'm going to stay close by. So if you need anything, I can look after you. As soon as people realized that they were detected, they went for an easier target.
10:30 Speaker 00 Yeah, a lot of deterrent measures are just to create some sort of interference. I know that's true to be in cybersecurity, where there's so many people that they can attack that if you have even one layer of security, one measure of security, it deters people from even trying to create some sort of perpet-
10:54 Speaker 00 Situation of a crime or incident. So that bends to be really important. What are we missing? I know, like I said, you were on GoFund yourself and had some great success and got great advice on the show. We certainly see some great things and a great investment opportunity. What haven't I asked that is of great value to Verify Faces?
11:21 Speaker 01 Um, well, we see our market going forward. We have competitors in people with accuracy. But we don't have any competitors in the network architecture, price and ease of operation and putting out there. So that's why the company is at a stage now where we're looking to expand quite quickly. We're delivering our product at a manufacturer level. We're not going in at the corporate level.
11:50 Speaker 01 And we already have now several global contracts.
11:58 Speaker 00 And just moving forward, we are utilizing crowdfunding and other measures. You are raising money. Where can people go to fund your project or learn more?
12:09 Speaker 01 Yes, we're on the Republic Capital Board and Verify Faces dot com. And we're looking to raise a total of $3 million against the company. We have two valuations done at the moment. One is pre-launch at 34.85 million US. And the other one is post-launch at 95 million. We consider we're somewhere between the two, but at the fundraising stage at the moment, we're giving investors the chance to get in on the lower bell.
12:41 Speaker 00 What an opportunity for everybody to get into the most economical and powerful way to verify faces and alert people of that verification, which is a nuance that only Verify Faces can alert 9/11 of potential problems that exist. Utilizing a set database is efficient, effective, and statistically successful, providing protection pre-incident, which makes it a great value. Check them out on the Republic platform. Check out Verify Faces. Credible entrepreneur, innovator, Michael Doherty. Thank you so much for joining me. This is David Meltzer with Entrepreneurs, The Playbook.
Previous From Pain to Traction Next How Changing Meaning Changes Everything Page updated November 14, 2026 · transcript and summary are machine-generated