Martinka Consulting's Getting the Deal Done Podcast
Martinka Consulting's Getting the Deal Done Podcast
How AI Can Improve Productivity in Your Small Business with Thynk.ai founder John Long
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Most of us have had a (bad) experience with Chatbots used by cable, phone, and other companies.
Enter Thynk.ai and their revolutionary use of AI for customer service, ordering, troubleshooting, and more.
Founder John Long discusses situations where their service and improve productivity.
One example is how AI can handle 80 phone calls at the same time, immediately providing product info, quotes, ordering, delivery dates, and more. Plus, an unlimited number of simultaneous interactions via text of email.
John Martinka
Jessica Martinka
Contact us via either website or give us a call and be sure to check out our blog pages with new posts weekly.
https://nokomisadvisory.com/
https://www.martinkaconsulting.com/
https://www.gddpodcast.buzzsprout.com
https://www.youtube.com/c/JohnAMartinka/videos
425-515-4903
My special guest today is John Longrith. ThinkAI. John John is based out of uh the Salt Lake City area. And this is the Getting to the Deal Done podcast. And I'm excited to have you here, John. Uh looking forward to our conversation based on what we chatted about a few weeks ago.
SPEAKER_01Yeah, thank you, John. I'm excited to be with you, excited to be with the Escape Artist and be able to uh visit with you today. So thank you for your your kind uh offer to let me join you today.
SPEAKER_00So AI is everywhere. Headlines, uh apps, you know, Microsoft has got their part built into some of their pro some, if not all of their products. Um so tell us a little bit about how think AI came to be, your, you know, how what led you to it, uh, et cetera, et cetera.
SPEAKER_01Yeah, so it's it's interesting. They say um, you know, that you usually find what you're looking for when you don't expect it, right? And so we uh my business partner and I, we were we were working on another company, and uh we said, hey, you know, wouldn't it be great if we had an AI assistant that could help us in sales? And uh and we said, yeah, that would be great if we could find someone that could help with the prospecting and qualifying of leads and scheduling and all the things that go with that. And so we started looking for something. This was about, I don't know, 18 months ago now. And uh we were looking and looking and looking. We we you know interviewed company after company, and we just said, gosh, I don't think what we're looking for exists. Uh but through our searching, we found pieces, right? We found pieces that you know could handle text messaging and pieces that could handle emailing and uh voice and you know, these different pieces, but nothing that brought it all together and really gave you that that complete experience. And uh my business partner, Dane Oborne, he's got 17 years of experience in technology product development. And uh so I said, Dane, what if we just built this? Like we've built all sorts of software over the years. Let's just build this. And uh, long story short, we we built a tool initially for us to use in our other business. And um after we got it built, we said, wow, this is actually pretty good. Um, maybe maybe others would be interested in this. And we started sharing a little bit with others, and we realized, hey, we've got a hit on our hands here. And um, hard pivot and and went all in on Think and on what we're doing in the world of AI. And uh think AI was born, and we've been doing that now for about a year. Uh, publicly launched in January of this year, uh, so about eight months ago. Um, but uh the tool's been in in development and and in use now for about a year.
SPEAKER_00Okay. So is your background tech and software development uh or are you a business guy, both?
SPEAKER_01A little bit of both. So uh right out of college, I started my own business in the world of real estate, real estate investing, um, did that for a number of years, then got into heating and air conditioning and ran a heating and air conditioning business for a few years and then sold that. And at that time I got into tech uh in sales. And so, you know, it was a little bit of a different hat for me to wear. I wasn't the business owner or the entrepreneur, but uh in sales, but uh it was with a company that wanted a um that they were starting a brand new division and a new new vertical that they didn't have any experience in. So I felt very comfortable as kind of that entrepreneurial, but uh in inside a company and uh helped grow that business and then have been in a couple of other roles since then, but uh uh have spent the last 15 years in in tech.
SPEAKER_00Okay. So AI and uh what tell us what your niche is in AI.
SPEAKER_01Yeah, so we we help businesses find uh you know specialized processes, and then we help them design an AI assistant to help with those processes. So uh we work with a number of different businesses, but it's easier for me to give you a few examples and kind of paint the picture. So we work with um autoglass replacement companies, and we've got an AI assistant that we've built specifically for that industry that can receive a phone call or a text or an email. It doesn't really matter how it comes in, it all comes to the same AI brain, right? But our AI assistant can then talk to the person and say, hey, what type of vehicle do you have? And they can give either a VIN number or make model and year. Um and our assistant can then go and look up the autoglass that they need for that vehicle. If there's a few options, they can talk to them about the options. Um it can price it, book it, order the glass, schedule the appointment, and essentially handle all of the back office type of work for that that company. Um and so very, very customized use case, but AI allowing them to automate significant portions of their business.
SPEAKER_00Let me inter let me stop you right there and ask. If I'm the customer and I call the glass the glass shop and say, look, I got a bad windshield, got hit by a rock. Well, I know I'm not talking to a human.
SPEAKER_01You know, it's uh it's getting good enough that a lot of people don't don't know that. They think they're talking to a person. Um, others will still pick up on it. Um, you know, the voice AI has improved greatly, even just over the last six months or 10, 12 months. Um, you still have a little bit of challenges with latency where there's sometimes a an unnatural pause, um, either a little too long or a little too short, and and we're constantly working on trying to make that more and more natural. Uh, but a lot of people uh have no idea that they're speaking with AI. Okay.
SPEAKER_00And how many to follow up, uh how many people can your AI app talk to at the same time? And I say talk could be email or text.
SPEAKER_01Yeah, so um we this is where it gets a little interesting. We can receive 80 simultaneous phone calls and manage 80 simultaneous phone calls at once. Um, texting and emailing, there's no limit. Um, but we can do 80 simultaneous phone calls, and that actually can get ramped up from there with a few changes on the back end. So I think we can go up to like 500 simultaneous phone calls if we needed to. Most businesses don't need that many. Yeah. 80 is more than most businesses need.
SPEAKER_00That's right.
SPEAKER_01And yeah, that's that, you know, there's your productivity right there is uh uh Yeah, and and the thing that's interesting is that you know, a lot of the things that businesses businesses spend time on, um, and when you spend time on it, you spend money on it, right? A lot of times the the things that you're spending time and money on are not necessarily what make you money, right? For for an autoglass company, what makes them money is installing glass, right? Now, they need to talk to customers, they need to answer their questions and schedule them and do all of those things. Those are all necessary, but none of that is a money-making effort. The money-making effort is the actual installing of the glass. And so the idea is we try to help businesses do more of what makes them money, right? Do more of what really matters and let us handle the busy work, the repetitive type of work that um tends to be the less desirable parts of the business anyway.
SPEAKER_00Yeah. And I can imagine I would imagine, you know, your AI when in sourcing the glass and everything is much quicker than any human on a computer looking it up.
SPEAKER_01Yeah, it's it's instantaneous. I mean, it's on the call. It's able to say, okay, make model year, awesome. Here's the glass that's available. These are the different, you know, do you want it with the blue tint or a green tint? Do you want it with defrost or not defrost? Yeah, it can give you the options and instantly tell you what's available and then price it, collect payment, move forward, take care of the entire process.
SPEAKER_00Well, thank you for letting me digress into details on that one. So, what's uh the next next use uh case study?
SPEAKER_01We uh we do a lot with sales, um, sales, scheduling, um, you know, those types of processes. Um, we I'll tell you a fun one that we're working on right now. Um, it's not not been released yet, but we're working on it, and it's a um a virtual sales rep for a car dealership. And so, you know, I I think, you know, like most people, you probably have been to a car dealership before and and you feel like, oh, I I stepped out of the my car, and it's like the piranhas are coming, right? And the sales reps start to close from from every uh part of the parking lot, and and you see them coming, and they're not quite jogging, but you know that it's like I gotta be the first one to get to this person so that they're mine, you know. Um, and we're changing the way they do things now. We're gonna allow you to drive onto a car dealership parking lot. There'll be a phone number displayed, and you can call that phone number, and you now have a virtual sales assistant that's able to walk you through the parking lot, ask you about the type of vehicles you're interested in. You know, are you looking for a sedan, a truck, an SUV? Oh, an SUV. Okay, we've got a number of SUVs. Is there a certain brand you're looking for? Oh, a Toyota. Well, we've got three different Toyota SUVs on the lot right now. Uh let me tell you what we have available. And uh, and as they're walking around, if they see a vehicle they're interested in, they can give the last four digits of the VIN number. And the AI sales agent will tell them all about that vehicle, the mileage, the you know, the background, the history. And then if they're interested in uh taking something out for a test drive, they can say, Yeah, I'd I'd like a test drive. And that will alert someone in the office to bring the keys out, get their driver's license information, and then send them out on a test drive. So it it takes the the part of car shopping that everybody hates, removes it, gives you a very personalized experience, and uh allows you to do it, you know, uh at your pace on your time, your schedule, et cetera. And uh, you know, really makes that a wonderful experience.
SPEAKER_00Yeah, uh I I I understand. I've I've been uh on both ends of the you know the the grabby salesperson. And one time I told him my lease isn't up on this, my current car for six months, and you know, he insists on writing out what it how he could get me into a car today. And uh the last car we bought was uh one you know was uh fixed price. Uh you could order it online, you can go to the dealership, doesn't matter. That's the price is the price, which was really an easy shopping experience.
SPEAKER_01Yeah, yeah. That was actually my last shopping experience as well. We ordered it online and they let me know when it was available, and I went and picked it up.
SPEAKER_00Yeah. Okay, that's that's interesting. What about uh what about B2B sales? Have you done anything with that?
SPEAKER_01Oh, yeah, we do a lot with B2B sales. That works great. Um, you know, B2B sales is interesting. Um, we we're still working on the prospecting side of things where it's getting past gatekeepers to the right people. But um, you know, anyone in the B2B space where you're generating leads, um, man, our our system works great for you know catching those leads, communicating, answering questions, qualifying them, because that's one of the challenges in B2B, is just making sure I'm talking to a qualified sales uh you know uh opportunity. Um but we can do that. The other thing that we do that's interesting on B2B is uh customer service. Um and so we work with a lot of businesses that use our AI for their customer service. It's tier one customer support. Um, but uh their customers can call in, take care of you know the majority, usually it's around 80%, you know, the old 80-20 rule. Um, but about 80% of the customer service calls can be handled by AI. Um so do we a lot we do a lot uh a lot there as well.
SPEAKER_00Okay. What's uh what kind of industries can you do that for B2B?
SPEAKER_01Um the ones that are using it most heavily right now are um we have some in the tech space that are using customer support. Uh we've got e-commerce. Um the e-commerce one is interesting. Uh one of our customers that uses it there uh it's e-commerce. They're selling um kind of more high-end health equipment, things like red light therapy and um you know uh red light saunas and uh anyway, things like that that are uh sometimes require a little bit more of an educational type of process for getting people up and running on them. And so uh our agent has been trained on all of their products and how they work and the benefits and how to use them the right way, and and it's able to walk people through as they receive this new product that's kind of new to them, how to uh how to manage that product.
SPEAKER_00Okay. So let me switch a little bit and uh I guess this could apply to business or consumer, but what do you what does your AI do differently than say if you're dealing with the your cell company or a cable company with their bots online? Uh it sounds like what you have is quite a bit more advanced than that.
SPEAKER_01Yeah, so a lot of times when you're dealing with like a chat bot or something like that, um, it typically is designed on an architecture that is a um a decision tree architecture. Um think of like a choose your own adventure book. When when you and I were younger and we used to read those choose your own adventure books, you need to get to a page and it would say, you know, if you want them to, you know, to jump down the the mineshaft, go to page 47. If you want them to, you know, leave, go to page 52, right? And it would walk you through this story in a unique way. Um, that's how a lot of chatbots are are designed. And so you have to follow whatever path they've outlined. And uh a lot of times, you know, similar with phone trees, where it's like, you know, press one if you'd like to speak to the accounting department. And then you press one, it's like press one if you'd like to speak to accounts receivable, press two if you'd like, and you're like, oh my gosh, I just want to talk to somebody. Well, our agent solves that. Um, you're immediately put in touch with either via chat, text, email, phone call, uh, with a knowledgeable AI agent. It knows the business, it understands the processes, it understands the the way the business works. And so immediately you're put in touch with someone that can answer your question. And so you can discuss in a very conversational way, hey, I've got this need. You can explain your need. The agent is able to understand that and then supply the correct information for you immediately. So rather than having to jump from one decision tree to the next to the next to the next, it just takes you immediately to solving your your need.
SPEAKER_00And how how deep can they go before a real person, a human being gets on the line?
SPEAKER_01Yeah, that that depends on on our customer. Our customer makes that decision where they say, Hey, I want the the AI agent to be able to do everything, or I only want them to do some subset of what we do here. And so the customer is able to kind of make that decision and customize it. Um the the caller is also able to say, Hey, I I think I just want to talk to a human. If they have any frustrations with the AI, it's rare, um, but but they can be then transferred directly to uh you know someone on the on the staff.
SPEAKER_00So you know, I'm I'm new to this, so uh I hope this isn't too basic a question, but how would they know to ask to speak to a person?
SPEAKER_01Like I said, in some cases, uh you know, you can tell you're speaking with AI with some of the pauses and things like that. Um or if they have a more technical need, they might not get the answer that they need. Again, our AI is really good at kind of tier one tech support. If you get into tier two where it's like looking up accounts and giving you know specific information or resetting passwords, things like that, we typically don't have our AI going to that level. Um, and so you know, if you call in and you have a need, the AI might even say something like, I'm not trained on how to do that. Would you like me to transfer you to uh someone that can help? Okay. Oh, yeah, sure, that would be great. And it would it would go ahead and make that transfer then.
SPEAKER_00Very interesting. Okay. So you've your company is about a year or so old, been at in the market since the first of this year. Uh I know we talked about you've got 15 to 20 people. What do most of the are most of them tech AI?
SPEAKER_01So we've got about half of the folks that work with us are on our development side. They help with development, they help with integrations, they help with customizations. Um, and then we've got um a few that are helping with um the AI, AI customizations. So you've got kind of the software behind everything, and then the actual AI customizations where we're um you know, we're we're adapting the um prompts that we use. Um, and and so we've got an implementation team that helps with that, and then we've got sales and then uh company leadership. So a little bit of a mix, but um yeah, uh it it's a pretty lean team for what we're doing. Uh we're able to produce a lot with very few. Um, and some of that's because the the technology is just really well written.
SPEAKER_00Okay. So use the word adopt, I would think. I guess the question I was going to ask next, which I will, is uh similar would be how do you train quote unquote train AI for each specific uh business situation?
SPEAKER_01Yeah, that's a great question. So the the type of AI that we're working with, John, is called agentic AI. And many people are familiar with generative AI, like chat GPT chats, right? Where you ask a question and then it gives you a response. And sometimes you go back and forth. I like to tell people generative AI is like playing tennis, right? You hit the ball over to the AI, AI hits the ball back to you, and you kind of have this back and forth experience. With the type of AI that we are building, the agencai, it is able to make decisions on its own. It's it's more like the quarterback in a football game where the quarterback can pass the ball, they can hand the ball off, they can run the ball themselves, they might choose to take a knee, right? There's a lot of different things that a quarterback can do. And so as we're training our AI, we work with the business that we're working with and we say, okay, help us understand what sort of plays you would like your AI to be able to do. Do you want it to be able to answer questions about these products? Awesome. Let's train the agent on those products, meaning let's build a knowledge base of those products. Um, do you have certain processes that you follow? Help us understand what those processes are, and we then train the AI agent to understand, you know, the processes A, then B, then C for this business. And this is what A is, this is what B is, this is what C is. And so we do all of that with a prompt similar to the prompt that you might use in Chat GPT, but these prompts are like monster prompts, right? These are like 40-page long descriptions uh of what the AI agent is doing. The probably the best uh best comparison is if you were putting together um like an SOP document or a um kind of a detailed description of a of a job for an employee, and you handed that to them as like a handbook. Here's your job, do everything that's in this handbook. Um, that's essentially what we're doing. We're handing that to the AI and saying, this is what we want you to do. And and within that, uh, we can we can also train things like what type of personality do you want the AI to have? Do you want them to be kind of uh lighthearted and funny, or do you want them to be a little bit more stern and direct? Um, and we can uh train them on how how close to like a script you want them to follow. Like, do you want it to be word for word? Is this really important that they've got every single word exactly right? Or is it a little bit more loosey-goosey and it's like, hey, we're heading this direction. And as long as they're heading that direction through the conversation, we don't care what they say. And so we can we can kind of put guardrails as tight or as loose as we would want on these AI agents and uh and help kind of head in the right direction.
SPEAKER_00So it sounds like what you're you're doing with your AI or customer service. Um if you compare it to the chat GTP, the the Claude, et cetera, Gemini, where they supposedly have had them scan all kinds of stuff off the internet and books, et cetera, et cetera, to get this knowledge. You've done that in a very focused way. Is that is that a good analogy?
SPEAKER_01Yeah, that's a good way to put it. Theirs is slightly different because they're training on language. So they um they use a process where they'll give the AI a sentence, remove one word, and say what words should go here. So it might say the sky is, and then they remove the last word. And the AI says, well, you know, 84% of the time when we see the sky is, the next word is blue. So I'm gonna guess blue, you know, right? And they're training the AI on how to communicate, how to how to talk. Ours, our AI. Already knows how to talk. It's been trained in that way, but now we're trying to train it on the specific details for this company. And we keep all of that knowledge uh in a private uh database so it doesn't get shared with you know with the world, right? So private AI? So yeah, it's in a private uh repository that's not open to uh you know to other other AI tools.
SPEAKER_00Yeah, I've been uh well a previous guest who's not in the tech field though encouraged me to start researching into a private AI for myself, and I could feed whatever you know books I've written and articles in that that uh any anyone on our team could could access without publicing.
SPEAKER_01There's a number of options for that type of thing. I will say this, and this is important for listeners to understand, um, AI is moving really fast. And a year ago, 18 months ago, uh there was a lot of concern that if I shared something with the chat GPTs of the world, and it could be chat, it could be Claude, it could be Gemini, it could be Grok, um, then now it's it, like I've shared something and it's now loose, right? And everybody has access to the knowledge I shared with it. Um and that for the most part has changed. That's no longer how they're training. Uh they're training these models on all sorts of data that is already in the public domain. It's already out there and accessible. Um and the the discussions that we're having with our our AI is not being used uh for the training, for the most part. Um and so it is interesting. There are some quote unquote legacy beliefs with AI. And it's the reason I give the air quotes there is because legacy is, you know, 12 months ago. Yeah. Right. Uh, but there are some legacy beliefs that are still lingering uh with AI in terms of security and privacy and things like that that uh really are unfounded in today's day and age. Uh, but again, it's moving so fast, it's not uh it's not a surprise that some people still have concerns in those areas.
SPEAKER_00Yeah. So what's a good comp good, not in you know, not specific company, but what's a good industry, size of company type of thing for that your firm think AI can really help with?
SPEAKER_01Yeah, so we we we work with large companies, small companies, solopreneurs. I mean, we work with with a lot of different sizes, but our bread and butter is kind of the uh a little bit larger than a mom and pop, right? We're not looking for uh someone that wears every hat. Um, but uh, you know, a company with you know 10 to 50, maybe 100 employees, where your employees are still having to do um multiple tasks. You maybe are starting to specialize a little bit and say, well, I'm now in the sales and marketing department. I do sales and I do marketing. I'm doing both. Um but we can go into companies like that and we can help offload a lot of the repetitive kind of tasks that people don't love doing and load that onto AI and give your people like this. We we tell we tell our customers, we're gonna give your employees an Ironman suit, right? We're gonna turn your Tony Starks into Iron Man and allow them to really focus on the things that that matter most in their world, and then take the things that are still important that need to be done, but that are repetitive and process driven, and allow AI to help take those pieces. And the thing that's so fun with that size of business is it allows them to not have to increase um payroll, not have to increase costs really, but it allows them to sometimes two, three, four, 10x the outputs that they're able to produce, with everything else being the same. And it it ends up being really um dramatic improvements, driving dramatic improvements for these businesses, and they surge past uh glass ceilings that they've had sometimes for years and years and years. Right. We work with companies that are like, man, I've been stuck at $5 million a year for 20 years. We've been doing $5 million a year for the last 20 years. And we start working with them, we get phone calls, and they're like, oh my gosh, we just hit 8 million. You guys, we never thought this was possible. Right. And so it's really fun to see some of those opportunities unlock for these businesses.
SPEAKER_00John, I'm gonna ask you for your final thoughts, but I want to say something first because it's just what you just told me triggered something from about 20 years ago. Oh, and I'm in the buy-sell world and uh as a company, and it was it was on autopilot at best. And the person who bought it was a service business, and they sold services in blocks of time. And the website was a brochure, and the buyer who got the company within two months switched the website from a brochure to an ordering system, freed up his people to not be playing phone tag or taking you know booking orders, uh, and doubled the business in two years. Sounds like what you were doing just reminded me of that in the next generation of using technology to again, he he increased the productivity of his people. Uh just like it sounds that's what you do.
SPEAKER_01Yeah, I mean, I like to sometimes you know talk about technology, and AI is just technology, right? Similar to like, you know, if I was to go to a tree trimmer from a hundred years ago and give them a chainsaw and say, hey, you can you can put your saw away, you now have a chainsaw. Well, if they didn't know how to use that chainsaw, they might still be sawing away on tree branches with their chainsaw, and it would take just as long, if not longer, to get those trees through. But once you teach them how to use the technology, the the increase is you know incredible, right? They can go from chopping trees down in weeks to to hours. And that's the sort of impact we're able to have for many businesses. Um, a great example. We had a business reach out to us uh not too long ago. It was on a Friday at 3:30 in the afternoon, and they said, Hey, we're running a sale this weekend. Can you help us send a message out to 21,000 people? We want to text and call 21,000 people. Well, reaching out to 21,000 people, if it was an individual person needing to do that, to send a text and to call them, that's like a year and a half, two years worth of full-time work to reach out to that. Well, we were able to reach out to 21,000 people before five o'clock that day. Wow. And the end result, um, I don't remember the exact numbers now, but I think they their their sale that weekend netted like $90,000. And it was a sale they put together at 3 30 on a Friday afternoon. And so yeah, it allows AI allows you to have speed in business that you've never thought possible and do things that you never thought possible. And you can do it with the same same team, in some cases, smaller teams. And I'm not a big advocate of getting rid of people. I think you know, most employees have a lot of domain expertise, and there's there's a lot of value locked inside employees' minds and and skill sets. Um, and so use them in other places, right? Don't don't don't get rid of them. Um, but increase the size of your business. You don't have to increase your payroll, just increase the size of your business by allowing AI to act as that Iron Man suit to let your people do three, five, ten times what they were able to do before.
SPEAKER_00Iron Man's a great analogy. So, how do people reach you and your company?
SPEAKER_01Yeah, you're welcome to connect with me on LinkedIn. Um, I'm fairly active on LinkedIn. Uh, feel free to look me up there, but you can also visit our website, uh think.ai, and it's spelled t-h-y-n-k. Uh think.ai, and uh there you'll have a chance to to visit with one of our AI agents, and they'll they'll help you get in touch with us. So we'd love to work with anybody that that has a need and happy to help answer questions that you might have about AI.
SPEAKER_00Yeah, John, this was fascinating, uh wonderful discussion. Uh uh I I I see what you're getting at. This is this is so different than the AI we read about or we use when we ask chat GPT or Claude or someone something, a question. Uh it's taking it to the next level and working, it seems.
SPEAKER_01It's working great.
SPEAKER_00Yeah. Well, thank you again. This is uh John Martinka with John Long with ThinkAI and uh on the Getting the Deal Done podcast. Thanks for being here, John.
SPEAKER_01Thanks, everybody. Take care. Thanks, John.