AIMinds #008 | Mohsen Amin, Founder & CEO at Contacti AI
About this episode
Mohsen Amin is the CEO and Founder of Contacti AI. He is a born entrepreneur who got the knack for business when he was just eight. Growing up in a house full of hustle, watching his dad's ventures, Mohsen started making moves early on. Whether it was teaming up with friends to wash cars or flipping baseball cards for cash, he was always onto something.
As he grew, so did his ambitions. Starting with selling dialers, he eventually built a global operation spanning four continents. But it wasn't all smooth sailing. When COVID-19 hit, he felt the pain firsthand, losing a big contract with a Fortune 500 company. But instead of folding, Mohsen got creative.
Teaming up with a buddy and an investor, they founded Contacti AI, the first of its kind automation solution that keeps businesses running, pandemic or no pandemic. Mohsen isn't just about making a buck; he's all about using tech to solve real problems and make businesses tougher and more efficient. His journey from kid entrepreneur to global big shot is a testament to grit, smarts, and never giving up on making things better.
Listen to the episode on Spotify, Apple Podcast, Podcast addicts, Castbox. You can also watch this episode on YouTube.
This time on the 8th episode of AIMinds, Demetrios meets with Mohsen. They spoke about the future of call centers powered by AI, the challenges in developing seamless AI interactors, and how his company, Contacti AI, is pioneering this space alongside their partner, Deepgram.
Key takeaways from their conversation:
AI's Future in Customer Service: Not just a tech fantasy anymore, AI is transforming the heart of customer service operations. It's not just about replacing human operators, but creating a whole new level of service efficiency - Mohsen's AI system stands out with its capability to handle interruptions coherently, thereby simulating real human interactions in a convincing manner.
Flexible yet Tailored Learning: Every business is unique and AI systems must respect and mirror that uniqueness. Mohsen reveals to us their approach for employing customer-specific data to tailor their system training, making their AI truly fit-for-purpose and business-specific.
AI-Powered Resilience: With the pandemic highlighting the vulnerability of staffing models, Mohsen emphasizes how AI can secure business continuity by stepping in during unexpected staff absences.
Fun Fact: Contacti AI’s partnership with Deepgram has enabled them to develop a feature that correctly pronounces names - a significant advancement in AI technology.
Show Notes:
00:00 Call center expert empathizes with customer challenges.
03:56 AI eliminates frustration and saves time.
07:17 Consolidating systems into one, facing manpower challenges.
12:47 AI system understands interruptions, doesn't disrupt conversation.
16:53 Automated system fills medical appointment cancellations efficiently.
18:20 Highlighting value of reducing appointment cancellations.
24:11 AI conversation challenge resolved with minimal latency.
27:56 Seamless call continuity, Deepgram integration, tech advancements.
28:49 Deepgram technology features.
More Quotes from Mohsen:
Transcript:
Demetrios:
Welcome back to another AI Minds podcast, where we explore the companies of tomorrow built AI. First, I am your host, Demetrios, and this episode is brought to you by Deepgram, the number one speech to text and text to speech API on the Internet.
Demetrios:
They're trusted by the world's top conversational AI leaders, startups, and enterprises like Spotify, Twilio, NASA, the one that puts rockets into space, and Citibank. Today we are joined by none other than Mohsen, who is the CEO and founder of Contacti AI. How are you doing today? Good.
Mohsen Amin:
How are you?
Demetrios:
I'm excited to talk to you, to tell you the truth. And like I was mentioning before we hit record, I wish I would have worn my suit, because those who are just listening don't realize how sharply dressed you are. But I want to be in that level right now, now that I'm seeing you and seeing how clean you look.
Mohsen Amin:
Thank you. It's like my Superman cape. When I wear it, I'm ready to zone in focus for the day. It helps with mindset.
Demetrios:
So talk to me a little bit about your story and where you're coming from. What have you been up to?
Mohsen Amin:
Yeah. So the past 20 years, being in the call center space, one of the biggest challenges, not only for me and for my customers, is that employees get sick. They decide they no longer work for the company. Crazy things happen, like COVID, the pandemic. All of a sudden, there's a shutdown, and you have 30% of your staff just sick in bed. And a lot of those were unexpected. I never saw that. I'm 41 years old in my 41 year old.
Mohsen Amin:
I've never seen a pandemic like this where all of a sudden you cannot go to the office and work or 50% of the people that you know are in bed sick. This AI came out at the right moment, right after the pandemic, because God forbid something else happens in the future. I don't think there'll be an interruption of business purely because there will be AI to make sure that all businesses, if all businesses adapt to AI, which I hope they do, that all businesses will be able to function even when their human counterpart is temporarily not available.
Demetrios:
So you have this deep expertise in the call center land, and I imagine you've seen a lot of the challenges. Maybe you faced them head on, and you've noticed them because you were working in the call center, or you had empathy for the customer who was calling in. I know that you mentioned that 2024 and beyond. The ideal scenario is that we are going to stop having to yell at the receiver and say things like, no, that's not what I meant. No, I didn't want to be transferred to another arpent. Can you break down what kind of artist you do have? Because I know there's a lot.
Mohsen Amin:
We have all experienced the IVR systems. When you call a bank or you call your phone carrier and you just want a simple answer and all of a sudden you're saying stuff and then it's repeating back to you what you never said, and now you're yelling at a machine, you're having basically a one way argument with the machine and at the end of it said, I'm sorry, please repeat what you just said. It's like I just said it 100 times. Then you calm down and you say it a little bit slower than it says, I apologize, I did not get that. Or in the middle of what you're about to say, you have a cough and then it like, I'm sorry. And it's like you get frustrated, like what's going on? I just need to speak to an agent. I just needed this information. I've experienced that.
Mohsen Amin:
I know you've experienced that. I know the worldwide, everyone experienced that. And with AI, at least with what we're building with Contacti AI, we're going to be able to eliminate that frustration purely because of the technology that's out there and our amazing coders that have been able to combine the technologies to bring what we have in store. But more importantly for businesses as a call center owner for a couple of decades, recruiting, hiring, training is a big part of time and expenses that you have to allocate. And with AI, you're able to eliminate that time to have to recruit and train, which means that if a customer calls me up and says I need 100 people, or I have 10,000 calls coming in a day, and I also have 3000 emails and I also have 4300 text messages coming in and I need these to all be replied to as fast as possible because the faster we reply, the faster they become customers. Or if they are customers, they're going to be happy customers. And now with AI, we're able to do that simultaneously. At the same time.
Mohsen Amin:
Everybody's emails, phone calls, text messages, chats can be all responded to at the same exact time. So you'd have 15 20,000 people on a website, or making calling for customer service, or wanting to reschedule appointment, or sending an email, wanting a response, and within minutes or less seconds you get a reply where right now, we're used to emailing and then getting a message, usually from Zendesk that's probably one of the biggest companies that everyone uses, that says, we've received your message. Here's your ticket number. Please give us 24 to 48 hours or one, two, seven business days to reply back to you. All of this is now eliminated with contact AI and the technology that's out there.
Demetrios:
And so was that the inspiration? It feels like it's twofold, right? That you had the pain of being a contact center owner and recognizing that employees get sick and employees can feel overloaded because there's never enough supply. As there is demand, it feels like there's constantly more demand, and then you have the ability to service, and then on the other end, you're recognizing that, hey, the current solution of us waiting days and that being the norm doesn't feel like it needs to be the norm anymore.
Mohsen Amin:
Absolutely. So about a little over a year ago, I started building my own CRM. This is where the whole thing started. So there's a bunch of CRMs out there, and they're all just either extremely expensive, so you multiply $500 to $1,000 a seat times 500 employees, or they're all just missing one thing. It's like this thing has a conversational AI. This thing has. You can send texts this way you get called, but you can't email. It's like they're all just missing something.
Mohsen Amin:
So my journey started with, like, you know, my staff are using several different systems. Why can't it all just be in one? Why does it have to be slack and this and that? Why can't it just all be in one ecosystem? And the ones that have it, there's limited to what it can do. So my journey started with, all right, I'm going to build my own CRM system. This is going to allow basically everything to be in one place. So one system, super easy. All the technology that's out there with conversational AI and all this will be in just one place. So as I started that journey and I was super excited, I still came to the realization there's still a human problem, there's still a manpower problem with training and recruiting. And this was obviously, this all happened during the pandemic where I was like, people are getting sick.
Mohsen Amin:
25% of the staff can't come to work. So I was in Canada visiting with my wife, her cousins, and my friend and investor, Amir, who is the investor of Contacti AI. I went to his house. We're just talking. I was, listen, like, I'm building this CRM. But my next thing, which at that time, I was starting to interview coders and developers and companies, was to create what I call at that time was AI humans. And I was like, there's got to be a way. So if all of a sudden 25% of the staff get sick or 5% sudden quit, whatever it happens, I can have a replacement immediately, or they can work side by side with the.
Mohsen Amin:
I'm going to call humans and AI, he was mentioning. Yeah, you know what? I'm actually working on something AI, something completely different. And I like your idea really? Well, are you open to possibly collaborating and working with the coders that I have? They're super intelligent, they graduated from UBC, they've won all these coding championships. And I said, sure, because at that time I was interviewing developers and coders. So about a week later, came back to the States, texted Amir, and I was like, do you want to have this conversation with your coders? And you set up the call. And then that first day we had that call, we just all connect. It just made sense. I explained to them what I was trying to accomplish and what the goal was of what I was trying to accomplish, and they said, yeah, we could do it.
Mohsen Amin:
I was like, what? Because everyone else I've been talking to, like, oh, they've made it sound so complicated. Well, they said, yeah, it's complicated. We can do it. We see your vision. We can do it. And then I was like, Amir, these guys understand me. It's very hard sometimes to speak to coders and developers the way that my brain goes in a hundred million miles when I'm coming up with ideas, because it's just what I've experienced. And it was a fresh breath there.
Mohsen Amin:
I was like, I think this is going to work. So last year we started, we'd have full cut, we'd have zoom calls, and I would just dump information from. I was like, this has to happen. This has to happen. It has to be this way. And I just showed like, okay, perfect. Got it, got it. And then they started creating prototypes.
Mohsen Amin:
We do test of the prototype. Like, this is missing. That's missing. It should be like this. And they just kept creating better prototypes. And here we are today, where contact the AI can do what I visioned for it to do, which is literally to assist businesses or replace some staff where it becomes extremely efficient. Maybe not replace staff, but put those staff members somewhere else where they'll be able to bring more value to the business and allow the AI to handle those small tasks that there's a lot of them, but to be able to handle it efficiently and immediately, time being the essence, purely because we live in a world of instant gratification, and I don't think that's going away anytime soon.
Demetrios:
Yeah, I do like what you said to me before, too, and you wanted to highlight that the idea isn't replace, it is making sure that the staff is able to work on the highest value activities and they're being augmented instead of being replaced. Because we all know that AI has advanced a ton in the last 2510 years, but it's not at that point where it is fully autonomous in 100% of everything. But that doesn't mean that you can't try and push it to the max and see where it can be autonomous and where you do need to get the human in the loop. So I would love to hear a little bit of like, what are the features that you thought were must haves as you were breaking it down with these coders? And I do like that serendipitous story. Feels like everything was just put right into place and all the puzzle pieces landed. And tell me, what are the different features that you were like, we must have this, because I've experienced this pain as an owner of call centers, and I know that it's going to be a huge value add if we can get the staff leveled up in these ways.
Mohsen Amin:
So one of the number one things, I think, was with previous, I think that they're called AI or IVR systems that was out there when you'd call and you'd have to explain to them why you're calling, right? And if you coughed in the middle of it, or if you had to tell your kid to like, hey, be quiet for a second. I'm on the phone all of a sudden, whatever you were saying just threw off the system. The system now thinks you said something else, or you have to start all over again. I'm sorry, I didn't hear what you said. So what we've been able to build is, let's say in the middle of the conversation, you're speaking to our AI on the call piece of it, the conversation, we submit and you cough or you sneeze, or in the background, you say, hey, turn off that music or anything that happens, that interruption that can happen when you're on a phone call. Our AI is not going to get confused. It's going to know that you're not talking to the AI, so therefore it's not going to say, oh, can you please start over? It might make a joke of you, like, oh, yeah, kids are pretty loud. Aren't know if it's like, hey, Tommy, be quiet, and then it'll continue where it left off, where that conversation was.
Mohsen Amin:
All right, now that, did you need a couple more seconds to let Tommy know again, we can control how funny it needs to be and whatnot for the customer, but did you need to put me on hold for a sec so you can tell Tommy to be quiet and then it'll pick up where it left off. And that's one of the most important features because there are companies out there that have this conversational piece, and I've tested a lot of them, and immediately you cough or something gets said in the background. That AI is now confused. And that's, I think, one of the biggest accomplishments our coders and developers have been able to do is that part of it. And that's one part of it where when I was speaking to the companies that I was interviewing, that's one of the things I said. I was like, the interruption is going to happen in the middle of conversation. Or when two humans have a conversation, you might go off track for a second and then you need to go back to where you were, and they go, that's not possible, or that's going to be five, six years of development, all this stuff. But the team we have was able to develop that.
Mohsen Amin:
I mean, I can talk to the AI. Hey, I'm calling about, for example, let's say it's a doctor's office. I'm calling to cancel my appointment because of a cough I have. But I can't make Thursday's appointment because I have to go to the movies or something like that. You just say something, it'll go back to the canceling instead of just like movies, and it just kind of goes off in its own tangent of movies. I don't know if that was a great example, but if you've ever had a conversation with anyone, just like we're having now. Yeah, this is going to get edited. I know, but you go in left and right directions, but you come right back to where you left off.
Mohsen Amin:
And that's something that I noticed that seemed to have been difficult for a lot of companies to do when I was interviewing them. And till this day, once a week, twice a week, we have a team that searches what's out there and we'll do these demo calls, test calls, and they all fail because as soon as you go off topic, it just continues off topic. It can't go back to where it left off, you can't decipher between that. So that's one of the biggest accomplishments our coders have been able to do. That's the call piece of it. Then the other features is let's say you do call and you say I want to cancel my appointment and let's use a clinic like a medical clinic. Well immediately, what with our technology we're able to do is we can connect it to your CRM. So we'll use your ecosystem, right, your own.
Mohsen Amin:
You don't need to use a whole brand new system in your ecosystem. We can go and update the dates of the cancellations, send out the email, call everyone else that wanted appointment for that day, the one that first either emails back, texts back or messages back depending on if you want a bot or not. We then put that person into the calendar and then update everything for you. So a lot of medical clinics, their issue is when someone or doctor's offices, when someone calls in sick or someone says I can't make that appointment, well the doctor is still going to be there for 8 hours, they still have the staff, they still have the overhead, they need to fill in those gaps. Now the receptionist or whoever it is has to pick up the phone, make phone calls and figure out who can come send out emails which let's say you have a pool of 4000 patients, 5000 patients, it's going to take a very long time to message everyone and even if everyone replies back to you, it's going to take a while to figure out who said what and when, where. Now this can be done automatically, it just happens instantly. I need to cancel, make it for Thursday. The system just does everything, text, emails, everyone schedules everything in.
Mohsen Amin:
So now medical practices, just one example will be able to make sure that every time slot is filled up purely because now we can cancel but also schedule others instantly, which is huge for medical practices, but that can be other businesses as well.
Demetrios:
Yeah. I want to mention two things to what you're talking about. First is incredible value add because the amount of stress that I can see being put on whoever it is, the assistant or the receptionist at these clinics every day, I bet two or three people cancel that kind of thing. Or statistically it's like at least one person's canceling every week and it's probably a lot higher than that. And even if you're putting conditions around canceling like you got to actually pay for the visit, if you cancel within 24 hours you're still going to get cancellations. And you still want to make sure that you're getting other people in who could benefit greatly from seeing a doctor way sooner than two weeks out or whenever their next appointment is because everything was booked up so I could see it for doctors or medical clinics. You can have it for chiropractors, you can have it for haircuts. It can be generalized across any vertical because everyone is fitting that same kind of mold where, well, I'm going to be here and I need to fill as many slots as I can.
Demetrios:
And I have all these people on the waitlist or this person that I couldn't give an appointment to because we were booked up for the next month or two months. So I can see how that is an incredible value. Add the other piece to it is when you were talking about how we humans like to go on tangents, and I can identify with that 100% because I am one of those humans that goes on many tangents, but the other.
Mohsen Amin:
Piece.
Demetrios:
Not just with tangents. But this is where I can relate to what you're saying so much because I've been on a call with the bank and they ask for my credit card number and I have to say it out loud and if I say the wrong number, or maybe it's not my credit card number, maybe it's just my phone number or it's some kind of long thing, and then I have to spell my name, whatever it may be.
Mohsen Amin:
I didn't even think about that. You want to say I made a mistake? Fix the last number. Fix the last letter. You have to start all over again. Think about that.
Demetrios:
Yeah, that's like the bane of my existence because it's like, oh, no, that wasn't a four. That was actually a five. Oh, wait, go back. Wait. Did it pick it up? No. Oh, man. I'm not going to be able to know until I get finished. And then even if I do say, like, start over, do I know if it started over? All that kind of stuff.
Demetrios:
So that's an absolute mess in my eyes. And I'm glad you're, you're right.
Mohsen Amin:
I didn't even think about that. That's happened numerous times, I think, just recently. The other one is this. You go to, let's say, a medical practice, and you are standing in line because the receptionist or the front desk is on the phone trying to help someone on the phone. So now you have a line of people waiting for these two or three people. So one of them help you, but they're busy on the phone, and that's where it's going to really alleviate that. Individuals at the front desk. Yes.
Mohsen Amin:
If it's like, if the system can't handle it, which I don't see why it wouldn't, then, yeah, get to a human. But now if 99% of it's being covered by AI, then those front desk individuals can give 100% of their attention to the people standing right in front of them. Happier customers and more.
Demetrios:
Not only that, but how much less chaos is there in the office if the phones aren't ringing off the hook?
Mohsen Amin:
Oh, yeah, a lot. So I want to bring that point up. There is answering services, right? My call center has an answering service, but someone takes the information and then emails it over to the front desk of that business, they still have to go and figure things out. Right? So yes, the customer got to speak to someone, but nothing was done where with this AI, the appointment can get done, it gets scheduled. Like the whole thing just gets done. So they don't have to go and do it. They really don't have to do anything. They just have to smile when the person walks in like, hey, how can I help you? My attention is yours, which everybody loves, right? Everybody wants that.
Demetrios:
Yes. That's the customer experience. That is beautiful. So as we're finishing up here, I would love to hear what some of the challenges you faced when building this company have been, because I can only imagine that there are some standard challenges of just building a company and getting product market fit and getting users all that fun stuff. But then you have this added complexity of being on the cutting edge of what is and isn't possible when it comes to AI.
Mohsen Amin:
So there's three blocks that consist of what we have. One is speech to text, which is Deepgram that we use. Then we have the LLM model for finding the proper response. And then there is the text to speech. So as I mentioned previously earlier, that one of the biggest challenges was this interruption, or when an interruption happens when you're speaking to AI, making sure the AI doesn't get confused and can stay on track. Whether it's the customer goes off on a tangent or yells something in the background or coughs that it doesn't have to restart itself all from the beginning, that it starts from where it left off. That was one of the biggest challenges. And being able to have this all happen because there's different systems that we're using to build this and to be able to make sure there's no delay, especially on the voice piece of it.
Mohsen Amin:
Right. That's the other part. The challenge with AI where right now you speak and then you have to wait three to 5 seconds and then it replies back to you and it doesn't sound like a real conversation. So that was our biggest challenge. But when I say we are coders who are geniuses, they were able to figure that out. So the conversation is just fluid, right? There's not like a three to five second delay after you say something for the machine to think and all while it's logged into the CRM system that you use in your ecosystem to make the appointments or look, oh, this person can't do this appointment look up times. Like, if the person on the phone says, I need to do Thursday at 03:00 because I can't do today at two, it's done instantly, not like, oh, please wait and have it like 30 seconds later, a minute later come back and give that answer to you, it's done within seconds. So that was the biggest challenge that we faced, and I know other companies face it, but we were able to hone in and that latency is to the point where it's like two humans having a conversation.
Mohsen Amin:
It's just out there. So that was our biggest thing.
Demetrios:
If it's not getting ten X better and you're not feeling that advancement, then it's like, all right, I imagine lots of the customers that you're talking to say, well, we already have a solution and it's not ideal, but it's going to be very painful for us to lift and shift, so we're just going to stay with what we've got. And so you really have to make that value prop like, no, it's not three to 5 seconds. We have latency under control. We have these tangents or the context, the ability for the AI to stay with the person as they go off on the branches of the tree and all that fun stuff. So it makes complete sense. And I love that you're doing that and you've focused on that.
Mohsen Amin:
Yeah. The other part know, obviously, we all know Chad GPT and there's Claude and there's other companies out there like them. But one of the things is it's very broad. Like, they're great systems. I'm glad they're around. But because it has so much information, you really have to hone in its training to make sure it doesn't say things it's not supposed to say or do things it's not supposed to do because again, it has so much information. So the way our developers were able to develop it is we are going to use the customer's data. So once we plug in our system, contacted AI to your CRM, it's going to go read your emails, read your text, it's going to understand your business, how you communicate, what responses you give.
Mohsen Amin:
It's going to look at that. So now it's based on the information from your own system, not billions of other trillions of other information that you have to figure out what not to say, what not to do. And I think that's another crucial piece, is we're using and learning from your specific system. Just like when you train someone, you want them to know about you. You don't care. They worked at, let's say you're a doctor's office, but they worked at hairdresser, right? That's great. You work there, but that information is not necessary for where you're working now. You need to know the information here at the doctor's office, or have you ever dropped a call and then you've had to call and restart everything all over again just to get back to where you were before the call dropped?
Demetrios:
That is the other bane of experience, that.
Mohsen Amin:
Right.
Demetrios:
Oh, my God, I'm getting anxiety just.
Mohsen Amin:
Thinking, no longer, if your call drops on the call piece, just call right back and it'll start right where you left off. No more having to go through that entire. Explain why it just will continue right where it left off. And really, the last thing we want to make sure that is out there, that we are using Deepgram, we're going to continue diving deeper with our usage of Deepgram. I want to make sure that's very clear. And we're really excited for, I know Deepgram has new, they're always developing new technologies or advancements, so we're constantly looking to that and seeing how we can utilize that and put that towards our system to make it even better. And I want to add one last thing. The names, remember, in the beginning you said, how do I say your name? Because you didn't want to butcher it.
Demetrios:
Yeah.
Mohsen Amin:
And most people end up butchering my name to motion. Whatever reason they say it becomes Mosh N. So what we've been able to develop because of Deepgram, from what I'm reading here in my notes, what our developers have done is it's going to be able to correctly pronounce names. So Mohsein will be Mohsen it will not be. I think that's extremely cool. Yeah.
Demetrios:
Incredible, man. Well, I'm very thankful for you all. Being in the startup community, it's great to see the growth that you've had and how you are honing in this product. So I wish you all the best and I appreciate you coming on here.
Mohsen Amin:
Thanks for having me.