AI's Customer Service Impacts, and How to Evaluate Tech Partners

speaker-0: Welcome to this edition of our MDM Amplified podcast, which is all about elevating vendor voices and solutions that serve distributors. This episode is sponsored by Canals. I'm Mike Hockett, MDM's executive editor. MDM is about to release the results of a major new AI research project we conducted that involved surveying over 400 distribution leaders about how they're seeing value with AI investments segmented by its top application areas.

We're now far past asking if and how many distributors are using AI and getting much more granular with sharing lessons learned, numerous case study examples, and industry input on what's actually moving the needle in terms of productivity and margin growth and where. As always, we seek to cut through the noise when it comes to AI and distribution, and this research absolutely delivers on that. Canals is one of the sponsors of that research, which will be presented on stage at our fifth annual shift conference during May 12th through the 14th in Denver, where I hope you can join us.

Built for distribution, Canal software solutions automate sales order entry, accounts payable, and purchasing workflows with the most accurate AI for the industry, working to eliminate manual work that slows your staff down. They have numerous distribution clients, particularly in the electrical space, which are mainstays on MDM's annual top distributors lists, and Canal sponsored the customer service portion of our research. My guest in this episode is Canal's co-founder and chief technology officer, Erez Arnan.

We explore our research results as they pertain to customer service, but what I love most about our conversation is the insights Erez shared about how to evaluate potential software vendors and specific things that are red and green flags. Given that he leads a technology provider, it wasn't something I expected him to cover, but I'm glad he did as the need for best practices on vetting software providers has never been higher. Enjoy. Ares, welcome to the podcast.

speaker-1: Thank you, good to be here. speaker-0: Before we dive into some of the results of MDM's new major AI applications research study, we always like to start with a guest introduction. So if you don't mind, can you introduce us first to who Canals is and who it serves, and then who you are in your background? speaker-1: Yeah.

So Canals automates manual workflows and wholesale distribution. That is sales order entry. Distributors, manufacturers will typically have teams of inside sales reps who all day, every day are receiving orders from their customers and they need to do their data entry. Canals automates that, hooks up to their email, phone calls, finds, identifies those orders and quotes and does the data entry for you.

Accounts payable. They typically have teams of... people working in finance who receive invoices from their vendors. again, they need to do a bunch of data entry and spot discrepancies and chase people and so on.

Canals automates that. Purchasing and receiving, when they buy stuff, they receive acknowledgments, advanced shipping notices. And again, there's data entry on that. And then also noticing discrepancies.

you, sometimes the vendor will send the wrong thing. And if you don't spot it on the acknowledgement. You only notice when the box actually comes in with the wrong stuff and then there's a huge delay. You need to reorder your customers waiting.

Canals makes that go away. And accounts receivable. They have people who are interfacing with their customers and they receive remittances from their customers, which they then need to do data entry on, compare against the data that's in the ERP, follow up with customers who are short paying or late and so on. Canals automates that as well.

speaker-0: That's great. And then can you go over how you came to where you are in your position there at Canals? speaker-1: So my background, I spent my entire career in Silicon Valley. First, two companies that were founded by other people, one of which was acquired and then another one where I joined with a very, we were four people when I started there.

I kind of followed the CEO of my previous company, he was very successful. He sold billions of companies, billions of dollars worth of companies. So I kind of followed him. And then this one and the last one, both of those, I was one of the co-founders.

Neither was my idea. Both are kind of the same story where I knew investors from my previous company and they kind of knew me and they constantly meet people who pitch them on their ideas and who often are still in the early stages and are looking for a co-founder. So the VCs really like to make those introductions because then, you know, they've contributed and that's how I've met both Michael here at Canals and my previous co-founder as well. speaker-0: All right, well, we're talking about AI here, which continues to dominate any technology discussion that we have in distribution.

But even though we're coming up now on the three and a half year mark since chat GPT's public launch, it still feels like so much of the discussion is around concepts and possibilities. Whereas I think our audience of distributors for the most part is well past that and looking for practical insights and proven real world application examples now. So as a bit of level setting, what are you seeing out there in terms of where we are on that AI hype cycle in distribution and what you're hearing from distributor customers about the knowledge that they're looking for?

speaker-1: say we're probably at the peak. There are definitely distributors seeing value already at this point. And when we go to conferences, for example, at this point, it feels like most people already know who we are. Most people already know somebody who's using us and have heard good things.

So that's definitely spreading. But also we're definitely still at the stage where it kind of feels like a gold rush. Every other week, it feels like I hear about another company in the space. And I'm also hearing more and more cases of distributors who are actually getting burned.

And that was not a thing. ⁓ even a year and a half ago, I feel like that was very rare and now it's a lot more common. If you ask me, what's the knowledge that, that distributors are looking for, I think one big one is that how do you sift through all the noise? There's actually advice that I give distributors that I think they found helpful and I think will be helpful.

⁓ so first of all, I'll say it's very hard even for me. If I'm just looking at the, way all the different providers in the space speak and their websites, I, we all look exactly the same. I cannot tell the difference. Everybody can speak.

Everybody speaks very well. Um, everybody has a nice demo. Everyone has, you know, really strong numbers. And I kind of know behind the scenes because I talk, you know, I talk to the distributors who come to us after trying somebody else.

But I'm like, if it wasn't for that, if I was just going by what's out there. Impossible. So the advice that I give distributors who ask me is evaluate, evaluate things that you can verify and don't trust anything that you can't verify. And so an example of that, the number one thing I say is make sure you see the product live.

Don't trust demos. It's so easy to put together a nice demo. You pick the one case where the tool works really well. You want to see it live.

should, the way you want to do that is say, say you're evaluating us for the sales or entry product. You would come to the meeting with a couple of your own POs, real POs from your customers. And you would ask us, can I see it working live? And if the provider says, sure, share those with me and I'll show, I'll share the results with you in an hour.

That's a flag. Cause again, it's so easy to fix whatever mistakes. to mix and so they should be happy to share their screen send it in and you watch it working live. That's maybe number one.

speaker-0: Seconds. Yeah, I like the idea of almost giving them a quick test. Not something that they can fully prepare for, but something that they can that they have to react to in real time. That's pretty good.

speaker-1: Yes. The two, the second probably biggest one is references. ⁓ I, I, ideally the vendor has enough traction where you already heard about them from someone you can ask them, but, even if you don't, if they're newer, that's fine. You can ask the vendor for introductions actually.

And of course anybody they connect you with will say nice things, but you're looking for the difference between, yeah, it's good. recommend it. And it's a no brainer. You got to do it.

You know, this was the best thing ever. And. You can also ask specific questions like one to 10. How would you recommend it?

And you're looking for the difference between a seven eight and a 10. You can ask how long did the integration take? What was that process like? You can ask for, do you have any specific metrics?

Are you able to quantify the ROI here? Is there anything that could be better? ⁓ so that would be number two. Number three is, and I do this when we're considering software vendors for ourselves.

And I also do this whenever I hear about a company in the space. go on LinkedIn. I check out their team, specifically the engineers. And you're looking to see, right, how many do they have first of all, and then the caliber ⁓ or the experience where they've worked that sort of thing.

even if you, I know that a lot of people, know, in distribution, might software engineering resumes is not what they're used to evaluating, but I really think it will not be hard most of the time to get a sense like. If you look at our team on LinkedIn, you're going to see about 50 engineers. You're going to see people from Google, Meta, Microsoft, Uber, Amazon, ⁓ former founders, know, CTOs of their own startups, competitive programmers, coaches, 10 years of experience, 20 years of experience versus, you know, sometimes I'll hear about a company in the space and I open it up and it's like five engineers.

And it's like, you can't do this so well with five engineers or, 90 % of the team graduated in the last two years college and and I think about myself, you know back when I was that age who's who's directing them and so I usually it doesn't have to be perfect. You're just looking for a filter and I think usually that's not going to be too hard. speaker-0: And in one way, it sounds like that's a lot of work to do a lot of homework, a lot of vetting to do, but at the same time, you are saving yourself a lot of time by finding out exactly what are you looking for and really narrowing it down to what are those check boxes that really have to be marked off.

⁓ so yeah, it is a lot of homework, but it sounds like it's going to save you a lot of time and effort in the long run if you get that done. So let's get into, and you already touched on this, but what really sets canals apart here. And seemingly every technology software provider in this industry has AI capabilities infused into their solutions. And if not, they're probably falling behind at this point.

But whereas so many of the known providers in this space have had to layer on AI into their offering or shoehorn it in, what I see is what sets canals apart is two things. One, The company was founded with AI at its core in late 2022. So it's been AI powered since the start. And two, its AI models are trained specifically for the industries of distribution, manufacturing, and construction across the key product verticals that comprise most of what MDM's audience is comprised of.

And that's industrial MRO, mechanical, electrical, HVAC, and plumbing. So it's not a retail native offering that then added a distribution version. It was there from the start. Can you expand or touch on why those two elements are so important right now for this industry?

speaker-1: Yeah, the industry piece, I think that's helpful because every vertical has, first of all, the different cases that you see there, different idiosyncrasies and things like that that you need to handle. And also both in the data that you're getting, but then also the use cases. for example, we have pretty deep support for wire cuts and reels, takeoffs, plumbing schedules, non-stock sourcing, direct invoice. billing that is a big one for accounts payable, a lot billing.

So, so I think that's helpful, but it's, it's really just a matter of resources. I don't think if there was a vendor that was more broad than us, I definitely wouldn't count them out just based on that. It's just a matter of you need more resources to be able to give the same amount of attention to, to, to each piece. But I would evaluate them normally and I think it's helpful.

But yeah, but I wouldn't discount somebody. AI native piece, I actually think it doesn't matter at all to be honest. speaker-0: Sure. Well, beyond those two components then that are true for canals, is there anything else that you see as what sets canals apart from the sea of SaaS providers that are out there in this space?

speaker-1: I think number one by far is the team. And maybe an analogy, if you're talking about the AI native piece, there was a period, right? Not too long ago, there was a period where it looks like Google might be in trouble. People really thought they were in trouble.

ChatGPT was all the rage. People were starting to switch away from Google to do search. And it really looked like they might be in trouble. But now it doesn't really, people don't really think that they're.

trouble anymore. They released Gemini. It's awesome. I've actually switched most of my usage to Gemini.

so how did they do it? OpenAI is AI native and Google isn't. And I think the answer is they have the best software engineers in the world. OpenAI does too.

Both of them, have the highest tier of software engineers. And so if it's technically possible, then they can do it. And if they have a legacy system, they can change it. That's the beauty of software.

The challenge here is that the problem is really hard. that we're trying to tackle here. And I think every distributor knows that variety of cases that they get and how, you know, seemingly crazy some of them are. Someone draws on a piece of paper, they take a picture, it's upside down, they cross some stuff out, there are typos, messy product catalogs, conflicting part numbers, duplicates, and a lot of it really does come back to LinkedIn.

If you look at the person who leads our product matching team, for example. He teaches a course on information retrieval. And before Canals, he led a data and matching team at an e-commerce product company, so product search. And before that, he was a data scientist at Mercado Libre, which is like the Amazon of Latin America.

so searching for products is his thing and 15 years of experience. And still it's a very hard problem and we're not at a hundred percent accuracy. So I think the team and the talent was it. speaker-0: Absolutely.

All right, well, let's get into the MDMAI ⁓ research here, which is the core of our discussion here. And so to provide an overview, in early 2025, MDMAI delivered a major AI research report that we then called, In Pursuit of Practicality, 52 Real Use Cases for AI in Distribution. And that project involved interviews with dozens of distributors across different verticals to identify where these companies are actually putting AI to use. And per its title, we detailed 52 of them in a resulting paper that I highly recommend our listeners check out, and I'll link to it in the promo blog for this episode.

But that 2025 project did a fantastic job of identifying what's possible with AI in this industry. So now here in 2026, we followed up on that research by going much deeper. And at the end of 2025 and here in early 26, we shifted the focus from what's possible to where to actually start and expand, identifying the five application areas that we've seen proven to have the highest value use cases. And that's in customer service, pricing and margin, inventory and demand, sales enablement and logistics and delivery.

So to do all this, we surveyed, I think it ended up being right around. 450 distribution leaders and conducted many more interviews to understand where they are investing, what they expect, and what progress they're making with AI. And the result of that new research is a brand new paper that we're about to unveil that we're calling In Pursuit of Value, AI Priorities from Over 400 Distribution Leaders. So back to my earlier question, it This project serves as a very in-depth sequel to our previous research to deliver a wealth of data and insights on actual AI applications within distribution in those highest value areas.

And we're grateful that Canal sponsored this new research in the area of customer service, which ranked second out of those five AI opportunity areas by the respondents who chose it as their top priority for 2026. And the range between number one and number five was only 14 percentage points. So it's not like one was far ahead or behind the rest, but Erez, when you see how that priority selection turned out, was it about on par with what you might've expected? speaker-1: First of all, the different sections, if you ask me to imagine what's possible with this technology and distribution, think yes, very much so.

The specific priorities, I'm only an expert really on what we do. If I was a distributor, I would be prioritizing based off where I think, right, R1, how difficult would it be to launch something there and how much value I think there is. And probably your listeners know what those numbers look like better than me. speaker-0: make sense.

Well, then let's look at current AI progress. And that's one of the other key self-rating components that our survey had was to ask respondents to categorize their current AI progress in each AI opportunity area with from the options of it's not on our roadmap or we're exploring or that they are currently piloting. And looking specifically at the customer service section, This resulted with 33 % of respondents in that section saying that AI was not yet on the roadmap. 41 % saying that they are exploring AI and 25 % say that they are piloting it.

And while 25 % may still sound somewhat low, it was actually the opportunity area with the highest percent who said that they are actively piloting by a narrow margin. So again, Erez, when you see that result, what does that tell you in terms of the appetite for AI in customer service? speaker-1: Yeah, that one surprised me. Both of those numbers surprised me.

33%. I don't remember the last time I talked with a distributor and they didn't think that they would be using something like analysis eventually. And so I actually wonder, but there's 33%. Does it mean it's not on the roadmap because they think it's not relevant or is it just not on the roadmap right now?

Like maybe it's a 27 thing. Based on the conversations that I have, it seems like people kind of understand that right now we're in a place where you can get an early advantage or you can get an advantage by being early. And where, right, some of our customers, they'll tell us, no, got, my customer sent me this huge request for quotes. got them, I got it back to them in 10 minutes and they asked me, how did you do that?

And that's the world of today. But, but a few short years from now, it doesn't even have to be all of your competitors who are doing this. If even half of your competitors are getting back to the customer. in 10 minutes, you're getting back to them next day.

In any sort of competitive situation, it just feels like it's, it would be really hard to do business. It would be like trying to run a distributor today without email or something. The 25 % number that actually surprised me as well. On the other side, that seems high.

If I just look at where, where we're at, even though I said at the beginning, you know, it feels like people, you know, know who we are and they've heard him so on, still we have at this point. 100 customers. You know, maybe in some verticals, like electrical is our number one vertical. Over there we have 50, almost 50 % of the top 50 electrical distributors in the US and a little over 50 % of the top 20.

But still, that's only the top, you know, X. If you look at the long tail and then you look at all the other verticals, it still feels like there is such a long ways to go. So that number actually surprised me. speaker-0: Yeah, in one way, maybe the biggest thing that those stats that I write off tell me is that there is so much room for growth.

Even though there's been all this discussion around AI, plenty of appetite for it, the amount of companies actually involved with it is still, we're still in the early days here. So there's still many companies that are looking for advice, looking for examples of how to do this and do it right. So there's a lot of room for growth here. So that gets into then what distributors are expecting versus what they're actually seeing with AI.

And that was one of the key things that we wanted to address in this survey. No matter the industry, no matter the product vertical, the consensus here is that there is a major gap between expectations and realizations with AI. Now, research reflected that in each of those five opportunity areas, including customer service. So we polled our respondents to ask them how much or how many of them expect at least 2 % margin improvement from implementing AI, and then how many of them have realized at least 2 % margin improvement so far.

And within customer service, 46 % of respondents said that they went into it expecting at least 2 % margin improvement, whereas only 8 % have realized at least 2 % improvement to date with AI. And I will point out that customer service had the smallest gap between expectations and realizations but it was still significant. So, Erez, this result doesn't mean that these same distributors who took our survey won't end up seeing 2 % margin improvement, but only that they haven't seen it yet.

So with all that context, what are your thoughts here on AI expectation versus realization when it comes to customer service and what a realistic ROI timeline might actually look like? speaker-1: Yeah, well, I think first of all, people are getting burned and part of the expectation gap is that. ⁓ Now, we talk about us, translating it all the way to margin, that I don't know because I don't know the full cost structure and operation model of a dissuader. Maybe you can actually help me out with that.

Or maybe your listeners, I can share the numbers that I know and maybe that tells the listeners what they need to know. ⁓ If you look at Order entry specifically, which I think that's the piece that we do that is around customers. Maybe accounts receivable is also customer service, but that's also newer and I don't have numbers there. So for the sales order entry, there are three main numbers.

Number one is time savings. And this one is, hard to get numbers, you know, that I would sign my name off because you kind of have to sit with a stopwatch. You have to go on site, sit with a stopwatch with canals without canals, the same order. At end of the day, everything I'm going to give you, everything I'm going to say, are, you know, what customers have told us, which are really kind of their estimates or our estimates.

And I'll try to be clear about that. But on that piece, we actually have a case study on our website with Puget Sound Pipe. And they told us that even for their most experienced sales reps who are already, you know, used and fast at entering orders, they think it's saving them 70 % of their order entry time. Newer folks more like 90%.

Now, I think if you ask, oh, you know, all of our customers, not everybody's going to say it. Everybody's going to agree that it saves time. You're going to you know, different estimates. When we do, we did kind of back of the envelope calculations and we estimate that we save about an hour of time per 125 lines on quotes or orders.

And all this stuff, again, going back to the beginning, don't take my word for it, right? It's so easy for me to just say stuff. One is the case study on our website. signed off on it.

So that's there. But also if you were considering us, you would want to reference us and you would want to hear from customers directly. But we're to do some math here. So we need some numbers.

so that's that. ⁓ and also very important to mention to my knowledge to date, no canal's customer has laid off a single person. They are. It is the case that sometimes when people leave, they don't backfill the position, but for the most part, they're doing more with the same team.

So, so I know that's a concern and to my knowledge has not happened not once. So that's time savings. Next you have win rate. And this one, this one is also a case study on our website with Turtle, who is a well-respected distributor.

And so they're, they've, compared the hit rates of quotes that they return, that they process via canals versus not. And normally their hit rate is 20 % and with canals it's 57%. It's definitely a real number from Turtle, but it's also, I have one data point. So this isn't something that I would say, you know, use canals and like, but there is a win rating piece.

And what that's attributable to is one, you're getting back to your customers faster and two, fewer errors. That's that. Then error rate as well. This is one we don't have a case study for.

We just have two customers that have shared data with us where they pulled what is their return rate of orders processed through canals and without canals. And one of them said there are 17 % fewer returns from orders processed through canals and the other 40%. So again, a big range there. So if you...

look at those three numbers and maybe we want to be super conservative here and slash them all in half. don't know how, do you have a sense of how much margin that translates into? speaker-0: Yeah, it's tough to say because each one is its own somewhat of a case study there. But when you factor all everything combined here, it undoubtedly does tell a story of ROI that they're seeing there.

And it's just, yeah, it's hard to tie an exact margin number to that. Yeah, I'm not sure what the equation would be that that would come out to how those translate to margin. But I think the narrative there is what's most important is that you are seeing proven examples from these, like you said, well-known distributors who can point to as close as they can to a statistic that shows the ROI there. So I think that's pretty impressive.

speaker-1: And I have one more stat actually. And that is this was actually publicly published by United Electric. They did an analysis with canals, you know, where are they at and how much can they grow revenue with the same team? And they came to the conclusion the answer is 20%.

That's, I think it was, it was published in an article somewhere and they didn't say it was canals. They said an AI solution that they're using, but it's canals. And I also don't remember if it was. without growing just their inside sales team.

I imagine they still need more drivers and so on or the entire team, but that was the stat, 20 % revenue growth with the same team. speaker-0: So yeah, it's almost impossible to find a direct correlation between one of those elements and margin growth. but that's what I would say is collectively there's, that's a great story, a narrative of ROI that companies are seeing in different areas. Even if it's not margin growth, it is time savings growth.

It is reduced errors. It is faster responses to customers that all lead to the ROI they're looking for. So again, just very impressive stuff there. Well, while I'd love to get into more components of that AI research with you, Erez, we just don't want to give it all away here.

And we don't want this podcast episode to be three hours long. So, but I do think what we just covered gives a nice sampling of what that project entailed. And we're going to roll out much more of the results and insights from that research in the weeks and months ahead. And that includes with an MDM webinar on April 9th that will spend a full hour exploring, showing those research results.

and I will link to that webinar in the blog for this episode, or you can click webcasts under the events dropdown on the MDM website. And following that, in person at our upcoming shift conference held May 12th through the 14th in Denver, NAW's VP of Research and Innovation, who is Patty Rausch, and I myself, we're going to kick off day two's programming with a main stage session. that likewise provides key takeaways and the juiciest results from that survey. So I do really encourage our listeners to join us at Shift, which is all about turning those concepts into action and change management.

And Canals is one of the companies that will be there at Shift as a sponsor. So we're looking forward to having them there. And that aforementioned webinar will include Canals co-founder and CEO, Michael Delgado. as a panelist to discuss those customer service results of our AI research and the accompanying paper that also includes plenty of commentary from him on this topic.

So it's been great to have Canals as a partner in this research. And Erez, you and Michael are talking with distributors all the time and anyone that goes to canals.ai can see some of those great company names that are clients of yours. This includes Turtle, you mentioned, Kirby Risk, Shade L'Oreesco, Johnstone supply, McNaughton McKay, United Electric, all companies that are mainstays on MDM's top distributors lists.

And neither you nor Michael have a, like you said, not a distribution background, which makes me curious to ask what led you to, to target distribution when you were founding Canals and getting it off the ground for an industry to focus on. speaker-1: Full credit to that goes to Michael. His wife is a distributor. And so he was kind of familiar a little bit and that he saw how fur business works and he recognized that there is room for technology there.

And so he started going to conferences, talking to more distributors, hearing about their pain points. And by the time I was introduced to him, actually, already knew distribution and he had a direction for the product. It wasn't fully fleshed out, but he had a direction for... me, as long as there are interesting problems to solve and it's something valuable, I'm happy.

Every company that I've worked at has been a completely different industry. And I think here, if we do what we are supposed to do here, everything, more or less everything will be a little bit more affordable for everybody. And I think that's pretty significant. So here I am.

I think if people are thinking, you know, why should they trust us, not from, you know, we're not distributors, what do we know? I would say it's true, we're not distributors. I think that's a good thing. I think if you want to work with a great distributor, you want to run by distributors.

If you want to work with a great software company, you want to run by software people. And we are a world-class software company. The problem that we're trying to solve here is very hard. And I think, I think that's important.

And I think at this point, we have a hundred distributors who are telling us what they need. If it's something we haven't heard of before, they explain it to us, we get it. That's not hard. Learning to build a successful software company, that's a 10 year journey.

was part of a startup two companies ago. was part of a startup that failed. And then the last one, it's a success. It's still running.

Let's say a standard success versus Canals, which is a remarkable success. And Michael has a similar story. One that failed, one that was, let's say a standard success and then Canals. And I think if I was trying to do this, if the me from a few years ago was trying to do this, I wasn't ready.

I wouldn't have been able to make it work. so yeah, here we are. speaker-0: love that you certainly do not have to be an expert and a long time tenured person in this industry to do great things in this industry. And I've seen that myself, seen new people come into industrial distribution and that are now leaders.

Whereas two years ago, they might not have had any odd interest or contacts within distribution. And now they're doing leading things that distributors look to. So yeah, I think there's that it's one of those things where this industry is comprised of so many. companies and executives that have decades and decades of experience, but you don't have to be one of those specific people to influence this industry in and serve this industry in the way that Canals is.

So I think that's a great point to make there. Well, to finish up, let's look to the future. Given how fast things are moving with AI in this industry, it's really hard to look further out than just a couple of years, given just where AI was two years ago and where it might be two years from now. But when you do think of where this is all heading, what is your expectation for how likely distributors are to really leverage AI tools in customer service and what capabilities might be available to them, let's say a couple of years from now that either don't exist right now or that are just concepts at this moment?

speaker-1: I think they'll definitely be using it. ⁓ And I know there's those 33 % that surprised me. ⁓ But, but I don't think there's any doubt. You have to use it ⁓ for the business more work.

In terms of what's coming that today is a concept, think one thing that, you know, today, the systems are unreliable and they're not in a place where you would put them in an area where, workflow. where errors are very costly. And that's why every product has to be built around that and humans are in the loop for everything. There is kind of the promise of AI that is just as smart as us.

And then you can have these fully agentic workflows with people directing and managing the thing, but you don't have to review every little thing it does and not everything needs to be catered. Does that happen? When does that happen? That is a big Debates, even the people who are building these foundational models disagree widely.

So I do, I have my thoughts, but there are people who are much better than me to listen to about this sort of stuff. speaker-0: It's one of those things where there's going to be a lot of people that are wrong about a lot of things and they're going to be right about a lot of things. But right now, as fast as everything is moving, a lot of this is still a guessing game. But like you said, the one thing we know for sure is that there's no going back.

The genie is out of the bottle when it comes to AI. And it seems like the distributors really primed for margin growth for higher productivity. Are those ones getting involved right now? So yeah, there's no doubt in my mind that the the adoption of AI in this industry is only going to remain exponential.

yeah, just excited to see where this is all heading. speaker-1: think the one thing that we're already seeing is we're able to do, we're able to be more, to produce more in the same amount of time or produce the same amount in less time. And I think that trend is going to continue and we're just going to see more prosperity as a result. speaker-0: thing.

with that, Erez, any final words or parting advice for our listeners as they are all navigating their AI in their customer service journey? speaker-1: Well, I think I'd circle back to where we started. Everybody, they know they need to adopt AI. They just want to know, okay, what do I do specifically and where do I start?

What I would recommend is, first of all, if you know somebody who's using a tool that they're raving about, just go with that. Start with that. If otherwise, I would recommend just make a list of vendors that you're hearing about. for each one very roughly put in place, you know, the estimate, the ROI that you think might be there, and also the time to value.

I would lean a little bit towards ones that have a lower time to value because as soon as you have one project that works and you see that success and your team sees that success, I think that will create a flywheel and your organization will just start to move in that direction. So all else equal, would err in that direction. So you do that. It can be very rough.

This is just for prioritization. You sort that list and then you start going down the list and you evaluate and remember, right? What were the few things? References, see it working live and LinkedIn.

And LinkedIn actually you can do without, without even talking to the vendor. So that's really fast. If you do that right, you should be, you will end up crossing out a good number of those. And that's important.

I really do know distributors who have multiple distributors who have spent months of their teams time, money, internal capital and nothing to show for it. So protect yourself. But if you do those things and a vendor checks, you just go down the list and once you find one that checks the boxes, if you do those things, I think you're protected. And I would say just go for it.

Start with one and go from there. speaker-0: Yeah, and it's finding that starting point tends to be the hardest and maybe the longest part of the process here. But that's, the message that I've heard from a lot of executives in this industry is you just have to start with something. It doesn't have to be perfect.

It doesn't have to be anything big. Find one area that you want to apply AI to that you really think is going to make a difference. See what happens and then if it works, scale it from there. And again, it's sifting through with the AI noise.

And there is a lot of noise out there. That's the biggest reason why we did this AI research in the first place is because we want to provide concrete, real-world examples for our audience to see out there. And Canals is certainly helping with that effort. So with that, Erez, thank you again for joining the MDM Amplify Podcast.

speaker-1: Mike, thank you. speaker-0: Thanks for listening to the MDM Amplify Podcast. This episode was sponsored by Canals, and you can learn all about them at canals.ai.

Like what you just heard? You can find our complete library of episodes at mdm.com slash podcasts, or wherever you listen to podcasts.