Ep 54: From Legacy to Leapfrog: AI Innovation at Avnet with SVP & CIO Max Chan

Evan Reiser: Hi there, and welcome to Enterprise AI Innovators, a show where top technology executives share how AI is transforming the enterprise. In each episode, guests uncover the real-world applications of AI, from improving products and optimizing operations to redefining the customer experience.  I'm Evan Reiser, the founder and CEO of Abnormal AI.

Saam Motamedi: And I'm Saam Motamedi, a general partner at Greylock Partners.

Evan: Today on the show, we’re bringing you a conversation with Max Chan, Chief Information Officer at Avnet.

Avnet is a $20 billion global technology distribution company that plays a critical role in the electronics supply chain; supporting the design, production, and delivery of devices worldwide. 

Theres’s three interesting things that stood out to me in my conversation with Max:

1) He shared a highly structured framework for AI adoption, breaking it down into out-of-the-box, embedded, and custom use cases.

2) He showcased a powerful example of gen AI in action: helping engineers co-design circuit boards by surfacing real-time alternatives that balance performance, availability, and lead time.

3) Max highlighted an underappreciated opportunity for AI to help with knowledge dissemination across the enterprise: using AI to enhance onboarding, training, and institutional memory.

So Max, first of all, thank you so much for joining the show. Maybe just kind of kick us off, do you mind sharing a little bit about your career and kind of what you do today as CIO?

Max Chan: You know, it's been interesting. And I love to tell people, first of all, quickly about Avnet, right? Because it's not a household name. Not a lot of people actually know who Avnet is. But I can guarantee you, for every one of you who are listening in, you'll probably have a device in your pocket or in your house, in your office that we would have touched. Because Avnet is one of the largest global technology solutions company focusing on distributions and we support our customers at every stage of the product lifecycle, right? From idea to design, from prototype to productions and go to market, right? We really have a comprehensive portfolio of components, electronic components, semiconductor that we represent, big names that people would know of, and we ultimately help customers bring those technology to the market.

So we are a 104 year old company. We are founded in 1921. That's why when people ask me about my career at Avnet, I've been with the company a little bit more than 10% of that time because my boss, the CEO, has been with the company for like 43 years. And many of my team has been with the company for 30 years, 29 years, 35 years kind of thing. So it's super interesting to be with a company with such a long tenure, and even IT globally have an average tenure of like 12 years. So I'm like, okay, I beat that average a little bit. 

But with Avnet, right, I started as the business unit CIO first, for the first three and a half years that I was with the company and then later taking on a corporate role for a little bit before being offered to take on the baton as the global CIO. I have responsibility for the run of the mill IT that you can think of, right? Operational IT from infrastructure to applications and, and also have the cyber security or information security reporting into my organization, into me as well. And I'm also responsible for visual enablement and most recently being tasked to drive enterprise AI strategy for Avnet, right? 

Being a hundred plus year old company, we strongly believe that AI is going to change the way that the business is run and potentially disrupt the entire industry. And being tasked with all that responsibility is quite interesting and exciting as well. And on the side, I also have P&L responsibility for one of the acquisitions we made into digital SI. So yeah, a little bit of everything.

Saam: One common thing we still hear from some technology leaders is like, Hey, you know, this AI thing is super cool. but, where's the real impact in my business? Like, yes, I'm using ChatGPT. My, my kids are using it for their homework, but how do I tease it apart and kind of implement it in ways to transform my business? And so maybe just to start max, like, can you talk about a few ways that you're using AI generative AI specifically today at Avnet and sort of the transformations you've seen?

Max: There are two dimensions that we are looking at it. First is that the different type of generative AI applications, right? So the different type of, we bucket into, three types. First, we talk about out of the box, a generative AI tools, things like just chatGPT, or perplexity, or co-pilot, whatever you are leveraging that ultimately helped with a particular purpose, be content generations, image generation, creations, et cetera, et cetera, right? So that is, we call it out of box. 

The second bucket is what we like to call embedded AI that is the race that every one of our software vendor, including Abnormal, I'm sure, is trying to incorporate into the solutions that they offer. And we already have so many of them coming at us. And many of them are taking that as an opportunity to increase the price of the renewals. But that's a completely different conversation here. 

Last but not least, is truly custom AI, generative AI solutions. That is what we, I talk about driving the enterprise AI strategy, right? That is what me and my team are really focusing on because that is where we look at how we can create solutions that ultimately benefit the customers externally as well as the employee internally, right? So out of the box embedded and custom.

And then the other dimension that we are looking at it is the Avnet generative AI maturity cycle. We start from, well, I call it stage zero, which is we think of it as personal productivity. So you can leverage those out of the box AI or embedded AI to improve your work. If you are in IT. how you can leverage that to help with coding, development, testing, et cetera, right? For a sales person, right? Leveraging that to consolidate meeting notes and help you create reports from your engagement with the customers, et cetera, that ultimately fits CRM and all. So those are for personal productivity, right? Coming from out of the box or the embedded versions of generative AI. 

Then we look at role-based improvement. Augmentations leveraging generative AI to help you code better, to help you as engineer come up with a particular circuit board design better, coming up with alternatives that you can choose which is the best one that allows you to present to the customers. That truly helped with, you know, accelerating our ability to go back to the customers, either provide a more comprehensive quote, a more complete quote, a better design in a much shorter timeframe, right? So that is what we call the augmentations or accelerations within a particular task or role. 

Saam: There's a lot I want to follow up on, but maybe one to start is just the use case you described around circuit board design. I thought that was interesting. And it's like a good example of like AI being used in a very specific way at Avnet. Like, can you maybe take that use case and like talk about how it works? Like, are engineers co-designing with AI tools? Are you using like full AI to actually generate end-to-end designs for this circuit board? you know, maybe help bring that to life.

Max: At this point in time, we are looking at more of a acceleration for engineers. So yes, engineering coming out with a particular design, because at the end of the day, right, when you do a circuit board design, depending on the applications, you need to understand what the footprint look, physical footprint look like, right? What is the output? What is the capacity? How much power that it can actually withstand, et cetera, et cetera. 

And the generative AI with the data that we have in the backend, right? Allow us to very quickly say, all right, now, let's say I want to reduce the power consumption by a certain percentage. And by the way, because of the data that we have, knowing the line cut represent the available to promise that when this particular components is going to be made available and how that could impact the go-to-market timeline of the customer, right? Leveraging generative AI to come back with what are the alternate, what is the pin-to-pin replacement and what is the potential output that still meet the minimum requirements that the customer has for that particular, that be it, a dashboard in the vehicle or drone or whatever it is, right? And really allowing them to look at, okay, here are the three different alternatives. 

That significantly reduced the time that the engineers take going back and forth with the customer. Because now the engineers, instead of coming with, here are three designs, the customer look at it, come back with some recommendations or question, they go back to the drawing board, created again, they can immediately in front of a customer say, hey, look, these are the things that we can do if this is what you, these are the parameters that you're changing. And to the extent that now we are also trying the ability to open this up to customers to do it through self-service themselves.

Evan: So, kind of sticking with the of applied AI, right? And sorry, this is going to be a bit of a tee up, but I promise it won't be too long a question. 

So I had a CIO call me the other day and said, I'm trying to give some examples of the board about like, to motivate like why specifically we should invest so much in AI and how we're going to get some big wins. And so the question I got was like, hey, can you tell me like, what are like the top, some top five easy wins, right? To help us hit the ground running. So I want to come back next board meeting and be like, we nailed one, two, three, and we got a lot more coming. Right? And I had some ideas, right, but I don't know, I don't work anywhere close to these industries. 

So I imagine for you, Max, if you guys have tried applying AI in different places, there's a couple of things that are probably work obviously out of the gate, right? Like people can edit emails, edit documents, send a screenshot of their calendar, chatGPT, and get something. 

I'm curious, more like industry or specific to your business, if you made a list of some of the areas you've applied AI so far that were highest impact to the business, by dollars an hour saved, divided by time to get going or energy to get going. If you ranked all those top, highest ROI AI projects that you guys have seen some results, at least early results from, what would be kind of like the most, what would be like the most unique one, right? That maybe like one of your peers, right? At a similar company, it'd be like, wow, that's actually really clever and I thought about that. Like any kind of like surprising or unique applications that have seemed like high ROI or quick wins?

Max: A few of the use cases that was quite interesting, one is specific to a, we call it the vertical library database. We have a group of sales engineers managing huge compilations of vertical knowledge, be it EV batteries or drones or whatever not. And it actually has all the different offerings from the different line cuts that we represent and the solutions and everything else there is. 

The challenge with that database is that number one is so big. So I'm sure you know where I'm going with it. And number two, while comprehensive, if anyone wants a report, people who have created that database basically have to go in and figure out how to generate the report for them to help them take it to the customers. We leverage generative AI and we took all the different vertical database there is, right? And fed it to our generative AI platform that allows now them to open it up to, to just anyone who wants to query it for things, right? 

They can now go in that, okay, my customer is looking for creating a drone that is this particular size. They're looking at this kind of things. How do I get started, et cetera. It immediately goes through not just one database, but across the different databases, right? And come up with, okay, this is the power. This is the board design that you can look at. And here are all the different follow-up that you can have that ultimately they can take and feed it into the engineering design tools that then generate a board design that ultimately turns into a bomb, a bill of materials for, for them to then go create a prototype. 

That, while, you know, it's not rocket science, it has really helped with productivity and also expanding that knowledge base that was only held by a small group of engineers who understand what to look for to now allowing anyone and everyone to use it. 

The other area of use case, which is not a surprise at all, because I've seen many companies going into is customer service, how we leverage generative AI to help customer service agents to better serve the customers and potentially allowing customer to directly interact with the AI agent to get better results beyond just the chat board that we they use, they are used to, right.

So those two, we are seeing great successes, and there are a lot more that is in the pipeline that we are working on.

Saam: I'm curious, like, was there a use case that you thought would be promising that you experimented with or rolled out where it like, it flopped and like it didn't work. And if there's any learnings from that as it relates to thinking about AI more broadly.

Evan: Especially when we're like, you're like really bullish, like there's no way this could be like a bad investment and it kind of flopped and then it was… anything that was surprising would be really interesting.

Max: You know, I know that this particular use case would have worked for other people, and that's why we push for it. And it's really on contract generations, right? 

We're like, hey, generative AI, the ability that it can summarize and understand the sentiments and generate new content. Why don't we take all the contracts we have, feed it to the platform and allows it to come up with new contracts or really help us decipher the RFP that comes up and come up with a response, right? 

So the team went out and created something like that. This was early days. This was probably like 18, almost 20 months ago now. One of the very first use case that we have done. And we came up with solutions that truly could either interpret RFPs that come in or generate contracts ready to go out. 

The response, while the solution work, the team felt that, hey, you know what? In the business that we are in, more often than not, we have to deal with a lot of different exceptions. And since we have to go through it and get through the exceptions anyway, it really doesn't help us as much as we thought it would. And we decided to just put it into the back burner and not deploy it, but the solutions was done. But that solutions actually became a great learning for us to apply on the vertical database that I was talking about. 

So that particular one, I thought that it was a sure win. It was a low hanging fruit that we could just roll it out and implement it across the board. So that is an example.

Evan: We talked about AI applications in the business right across the team. I'm curious for you, Max, I'm sure you've obviously experimented with some of these tools. I think all of us are using, at least lightly, some AI tools, our own personal workflows. Is there anything that you've done using AI for your day-to-day work that's been surprisingly helpful that you'd recommend, you know, some of your peers try experimenting with.

Max: Yeah, in fact, the extent that we, I find it hard not to go through a day without actually leveraging tools. We are a Microsoft shop, so we are very much into co-pilot. What happened is that, you know, we were lucky enough to be one of the first EAP, right, with Microsoft, the early adoption program or early access program that they roll out. And allowing us to leverage and also work closely with Microsoft to deploy that. 

The ability to leverage Co-Pilot on everything that I do. I cannot imagine going to SharePoint and look for another piece of document these days. Anything I need, I'm at Co-Pilot, I need to prepare a board deck on this particular topic. Get me all the data, information that is out there. And by the way, summarize it in five bullet point that I can put on one slide for the board, right? So that has been super helpful. The ability to pick up on any meetings, late into it and know exactly what was discussed prior to it and jump straight into being productive on a meeting that also is great.

Outside of copilot, leveraging other tools like perplexity and even chatGPT enterprise, et cetera, that has been great for me personally. Every now and then I use it to create code that has nothing to do with work, but help with analyzing certain things that I would like to analyze with Python or whatever. But that is probably more of a hobby as a hobby and then as a CIO. 

That being said, right? What we have done across IT is that we told every single team, like, okay, make sure that you have hackathon or similar type of activities with your team and every quarter come up with one to three use cases that you can go and POC and ultimately implement. And I make sure that I put it into the performance goals of each of my direct reports to drive that kind of things because IT professionals, if we are not leveraging it today, how can we be the change agent that helps the company adopt AI moving forward, right? And there are some great things that the team has come up with, through those hackathons and similar type processes that we implement either in managing incident tickets or helping self-service for our users and employees, et cetera, and also the provisioning workflow and all, right? So we continue to leverage that to help across IT to show people that, hey, this is how we do more because we know that the workload is going to just continue to come in.

Evan: We like to kind of end this show with a bit of lightning rounds looking for your like one tweet responses to questions that require usually a lot more than one tweet to to to answer, so Saam do you want to kick it off for us?

Saam: Yeah, absolutely. So Max, maybe to start, how do you think companies should measure the success of a CIO?

Max: At the end of the day is really what the company needs from a CIO, right? A CIO from one company is very different from another these days. As far as we are concerned, we look at it on the business value that the CIO create and deliver to the organization. Are we impacting the top line? Are we impacting the bottom line? Or are we generating cash for the organizations? This is how we do it and we have gotten quite used to it so that works for us.

Evan: What's one piece of advice you wish someone told you when you first became a global CIO?

Max: You know, I would have hoped that people have told me that CIO is not a technical role, right? Because my day to day is really, I feel more like a politician. I feel like more like a lobbyist. I go out there and talking to the business and helping them really to solve business problem and the importance of what looking ahead, right, in the transformations versus firefighting on the here and now, right? This has been a huge learning in my years as CIO that, you know, at the end of the day is really helping the business understands how technology and innovative technology solutions can help them achieve their goals.

Saam: Maybe switching to the personal side, what's a book that you've read recently that's had a big impact on you and why?

Max: The latest book that I read is by this Harvard professor called Karim Lakhani. I don't know if you've heard. It's called Competing in the Age of AI. I thought that it was a great cheat sheet for everyone in the CIO or CDO or CTO role, to go through because it really talks about pitfalls that you should avoid, how do you get started, et cetera. So as well as reading through it, I'm like, wow, okay, I think we have done some thing well and here are some other things that we could do better.

Evan: What is the most underhyped and overhyped AI use case right now.

Max: I think the over hype AI use case is probably in customer service. I think that is where everyone thinks that is the silver bullet at the end of what we have learned and the approach that we have taken, right? It's not a silver bullet. It helps with great accelerations and it has a lot of potentials, right? But it's not something that you can just apply and expect it to solve all your problems. 

I think underhyped, I don't know if there's one that is underhyped. Maybe in the ability to help with knowledge dissemination, right? Because a lot of people don't realize the power of leveraging generative AI especially, to help with onboarding, to help with continuous learning, et cetera, for employees. And I think all of us can do more of that. While we have done some, there are a lot more that we can do. And that's what we are partnering with the chief HR officer to try to leverage.

Saam: Max, what's an upcoming new technology or product? It doesn't have to be AI related that you're personally most excited about?

Max: Autonomous everything. I guess people tend to think of that as general intelligence, right? And how that translates into the physical realm. But I think that is going to truly help with a lot of things from a personal life to, to work life and everything else, right?

Imagine the most dangerous work, like a firefighter can be augmented with AI to get to dangerous places that you cannot get to do something as simple as, well, I know it's available right now, but it's not generally available, making dishes so that I can have a hot meal when I go home because most families have both, if I have everyone working, right?

Evan: Max, we appreciate you taking time to join us today, and looking forward to talking again soon.

Saam: Thanks Max.

Max: Thanks for having me. Appreciate the unorthodox podcast that we pretty much talk about anything and everything.

Evan: That was Max Chan, Chief Information Officer at Avnet

Saam: Thanks for listening to Enterprise AI Innovators. I’m Saam Motamedi, a general partner at Greylock Partners.

Evan: And I’m Evan Reiser, the CEO and founder of Abnormal AI. Please be sure to subscribe, so you never miss an episode. Learn more about enterprise AI transformation at enterprisesoftware.blog

Saam: This show is produced by Josh Meer. See you next time!