Episode 355: Your Team's AI Enablement Guide for 2026

Foreign.

Is the year you stop tinkering with AI and you start to scale it across your company. So what does that actually look like for your team? How do you scale it? Where do you even start? Well, on today's show, we're exploring what it takes to move your team from scattered AI experimentation to fully integrated workflows. Welcome to the Skewcast, the podcast for innovators and maverick thinkers in the promotional product space. My name is Bobby Lehue. I'm GLAD you're here. Eku Malcolm is our new director of AI operations at Commonsku. He spent the last four years at Office Space Software, embedding AI as a structural operating principle, scaling the company from 18 million to 42 million ARR while cutting manual work cycles by 30%. But more importantly, he's doing this work with our team right now in real time as we figure it out together. So today we talk with Eckoo about what leaders are actually asking for help with when it comes to AI. Some concrete examples of what changes when workflows happen, what tools and resources Eckoo recommends, and a few predictions about where AI is heading in 2026. Today's episode is brought to you courtesy of us at CommonScu. Over 900 distributors powering 1.8 billion in network volume, rely on CommonScue's connected workflow, process more orders, connect your team, and dramatically grow your sales. To learn how, visit commonscue.com now. Here's my chat with Eku. Eku, welcome to Skewcast.

Thank you for having me.

As I mentioned in the intro, we're going to cover today what leaders need to know about AI moving forward, particularly as we talk about developing AI and making it a priority for our team in 2026. So let's just kick it off with first question. What's the one thing managers and leaders are asking your help with when it comes to AI Right now, I think

what's really interesting is that people asking not so much about the tools, but how to start. And I think we're all at this precipice of understanding AI. We all know that we're supposed to be using it, but we're not sure how to get started with it. And so I wouldn't say that leaders are necessarily confused about what AI is. I think they're overwhelmed about how it fits into their own ecosystem, which is really interesting. And the real ask is, how do I introduce AI without creating chaos or fear or shiny object distraction, that kind of thing. And so what's really important from my side is Helping people understand the guardrails that we need to put in place to run AI effectively, but then how to move from that kind of experimentation phase to how do we get to fluency, how do we get to enablement at scale? At the end of the day, leaders don't want more tools, they want fewer decisions. So how do we do that in a way that can help them as opposed to something that they feel that there is additional work? Which is the challenge.

That's a real challenge right now, because in our personal lives, we might be playing with all kinds of tools with AI, and then you're trying to figure out how to consolidate tools for the business use case, and then that way you can maximize everyone's education around them, ramp up people faster. It's a unique challenge that we've not had in some time in the market where you have this proliferation of tools and options.

Yeah. And I mean, I think that's the thing, the AI ecosystem right now, there's so many options, and so what we use personally may or may not be the most effective for a business lens. And that's something that we're also learning and evaluating. You know, ChatGPT was first to market and made it incredibly accessible. But that may not be the tool for. For your specific use case anymore. It may be. It's a good starting point, but I think that's something that we've almost gone from one extreme to no extreme, where we started with one, and it was first to market and it was really prevalent, so everybody started there. So now there's AI tools around every block, and so how to pinpoint and to think about, and it's something that we're facing at commscue is how do we make sure that we're using the right tools for the right tasks. Right. And that is an evolving journey because honestly, that changes day to day when a new model comes out and, you know, they can change the way that images or visuals are done. That's completely different than, you know, three weeks ago, which was completely different. And so my message to leaders is that you're not alone if you're feeling like you can't keep up. And I think that's normal. And so the biggest thing is, you know, starting to do your research and starting to ask questions and starting to. To think about what is the actual use case that you're looking for as opposed to, you know, do I need a specific tool? Yeah, those are kind of questions that I'm starting to ask myself, but I also think are important for our leaders to ask as well.

We're both passionate about this topic, so we're going to balance all around. So one of the things I just realized too, as leaders, we bring our own bias about AI to the conversation and to our teams. So, for example, someone may be opposed to generative AI, but they love agentic AI. Somebody may love generative AI for everything. And so what I am learning through this journey of working with AI and talking with other leaders is that everyone brings a bias. The last time I saw this was during the pandemic when everyone went to virtual offices. And then you had this moment post pandemic where people were trying to decide whether to come back in the workplace, whether to have virtual, whether to have hybrid. And what was interesting that I learned talking with so many leaders then was that whatever your bias was, you're going to prove it. In other words, if work from home in virtual offices is the best option, you're going to prove it by putting in the best practices for that. I'm feeling the same kind of energy around AI. Are you sensing that, that we all sort of bring our bias to this particular conversation?

And I think we're, we're seeing the, the human dynamic at scale, which is really interesting. We're seeing people come out with caution, for lack of a better word, where they're, you know, we're seeing that personality kind of come out and they're hesitant to try new things. We're also seeing the extreme innovators, right, who are at the forefront and, and we're seeing it live. And the one thing that I keep coming back to with AI is that the only way that you're going to get more comfortable with is you have to use it, you have to try it. I think one of the things we're dealing with just at scale in general is that when people are reluctant to apply, there is a sense of a fear that comes with that. And so that first barrier is how do we overcome that? How do we make AI normalize so that they're comfortable experimenting? Because then from that point, then they can decide their personal level of comfort, depending on the bias that you're talking about. But I think that first obstacle is you try it, see what is comfortable for you, see what is capable they're capable of and what it's not. And I'm a huge believer in AI and I also think there are very specific tools that I will never use AI for because it's not designed to do that. And that conversation is really valuable to start to have it's like, how do we start to identify, yes, it's a personal bias, but I'm going to try it. And then that personal bias is now educated because we all know that when you're educated about a topic, you're more comfortable with it and then you can make your own decisions as opposed to kind of like the fear of the unknown, which is a lot of people are working through as they're navigating AI right now.

Yeah. I've noticed on the extremes, you have folks that treat AI as binary. It's either all good or all bad. And so on the extremes, I have talked with folks who say, I'm not getting into that. I think it's a mess, I think it's bad. I think it's. It bodes bad for the future. And then there are folks, I probably fall further on this spectrum where they're like, this is amazing. We've seen it do amazing work. And so I've noticed this even in myself, that I have to check my bias and understand, you know, that we bring our bias to this conversation. But I agree with you. If you are on that side of. I'm never going to invest in that. I realize you're probably not even listening to this episode, but it's, you know, it's really important to try these tools before you come up with, check your bias. Try these tools.

Yeah. The reality of the situation is, is if you don't try it, you will get left behind. And so what does that look like at this point? What I've used this analogy with a few people, and I think it is very present for this conversation, is that using generative AI is this generation's calculator. What we never thought was possible is now at our fingertips. You know, people from my generation and older. Oh, well, you're not going to have a calculator in your pocket. Well, now we have a supercomputer in our pocket. Right, right. And genes is that same element.

So, eck, tell me about one workflow that we've either built. For example, common skew's, prompt architecture and GPT. Yeah, whatever it is. You wanted to share some of the things you've tried to embed in our team or trying to embed right now.

Yeah, for sure. So prompt architecture is a big, big need, especially for people who aren't comfortable in using AI, if they don't architect it correctly, can lead them down a rabbit hole. That leads to trust issues. So one of the things that I did early on in my, in my time at Commonsku is I built a prompt architect using a custom GBT and what that does is educate in the back end. And I created an element where I said, I need you to be the best practice prompt architect. I need you to self educate on a monthly basis about what's new. That is one thing that is near and dear to my heart, I guess, and I'm very confident in saying is no one is staying on top of AI, not even like it's my job to stay on top of AI, and I'm always feeling like I'm working from behind. So I needed to harness the power of GPT to be able to actually say, tell me what I don't know. And so I built a custom GBT with the idea of I need you to help guide people to create prompts the most effectively that they possibly can. Built in some parameters around self education about quality control, about making sure there's no hallucinations, making sure that there is a clear path and a guide of actually saying, this is someone who doesn't know how to create a prompt. All they have to say is describe their problem or what they're looking to do, and then that prompt architect will actually generate a prompt that they can then feed in to GPT and it'll actually give them the specifics to actually go and execute that. And so that's leveraging infrastructure coming from the LLMs directly. Some of them are talking about, you know, the importance of having a role, the importance of having a specific use case of that giving context. Right. When you just give a general, hey, please design this. That is challenging when you can say, hey, I want you to think of yourself as this role with this use case and this is the parameters and the constraints that I want you to do it in. The output is a much higher quality so I would say has been effective so far. That was a very early example. And it was honestly, it took me probably about 10 to 15 minutes to do so. It was easy to create and now the output of that snowballs because people can use it effectively to create their own prompts in whatever they need to do throughout the company. So it's something that is open to our company and really didn't take a lot of time up front. But hopefully we'll have a positive implication in terms of people's output and what they're able to do going forward.

So in a sense you've calibrated AI to automatically create better output for our team, no matter where they are, no matter what their skill level is.

Yeah. The only Word I would challenge there is automatically, there is still a point of reference that needs human intervention at the start. They need to actually create, to use the prompt architect to say, hey, I would like this prompt architect to help me design this prompt for so and so solution. And the reason why I call you on that is very important because this is the difference between agentic AI and autonomous AI and generative AI. Generative AI is all about creating content. It's creating something that didn't exist before, something that's going to save you time, efficiency. The reason why I'm being specific about the word automatic is because that is what agentic AI is working towards. The ability to create an agent that will automatically do a task for you. And so that there is a difference there that I think is important to call out and is are you looking to create which is creating a trigger from yourself or are you using a systematic trigger that's going to say, hey, I want this agent to run in the background and generate that email and then send that email autonomously without my intervention.

One quick jump back into platforms because we're talking about GPT. Yeah, and custom GPTs. Right now there are organizations that are kind of split, I would say our audience is kind of split down the middle. Some of them are what I call Google houses and some of them are Microsoft houses. So some of them are going to have Gemini tools baked into their normal workflow because they're all. They work in Google, they work in Gmail, they work in Google Docs. And then there are other companies who are hardcore hardwired into Microsoft and so they have a lot of copilot and other tools like that. What we are seeing emerging, we're seeing to keep up with this sort of horse race about who's got the better tool right now is so confusing because one day it's, you know, one day it's this and then it's that and it's just so, so massive. Between Cloud, GPT, Microsoft, Grok, you name them, you just go through the list. DeepMind, you tend to have this horse race. But would it be best for companies to just embrace the tools that are embedded in their normal workflow first? I mean, how do you rec. Then leaders sort of go, okay, there's 20, 30 tools out there. How do we coalesce this down into something we can use for now?

Yeah, it's a really important question. So a quick clarification. Gemini is baked in from a Surface level, but most people don't have the Pro edition, so they're not choosing to pay for the enterprise package of Gemini. And so that is a really important differentiation because Gemini at scale, I would just argue, and especially with what it's done recently, if you have that Pro package embedded in your goog workspace is incredibly powerful and maybe enough. But if you're not paying for that additional level of service, and that's not including your Google workspace, its benefit is limited a little bit. And so that's something that there's a business decision that people would want to have in terms of do they want to pay for the. The Pro edition of Gemini to be embedded in their workspace. That's one. The second thing that I would say on that is, honestly, I would start with one. And there's so much proliferation out there, but. And all of them have pros and cons. And we can get into that if it makes sense. But what I would recommend is that being conscious of the fact that something like a gbt, most of your team is already using it. We just did a survey and we found that 94% of our team is using GBT.

Yeah.

And so that's something that they're already comfortable with. And it's probably going to get you 80% of the way there for most of the tasks that you want to create. The question becomes, is it the right tool for those specific use cases? And that's something that's probably worth doing some research on. You can use a GBT to do that research. But starting to ask those questions, say, oh, like, I want to be a financial analyst. Well, maybe GPT is not as. As good at that. So that's something that asking GBT direct is like, what are the best tools for this? What I would really recommend is start with one, like, start with one model and, and let it build for a while. Get comfortable with that a little bit and see how that's progressing. And of course, just like any software, there's going to be outliers that you need something else for. But start with one and really see what you can get out of it. Do the education element, use the enablement part of it. Really try to get your users comfortable with it and then kind of adapt from there. GPT is easy because everybody's already comfortable with it and we're all using it kind of in our personal lives.

Yeah.

But it may not be the best tool. And so that's something that, as you get more specific, maybe a cloud becomes more valid for coding, for example, or for deeper analysis. But what I will say is that I use for my journey with AI is GPT almost exclusively for probably about a year and a half or two years. And it did most of what I needed it to very well with a few kind of key elements. I was like, for a long time it couldn't do design and visuals very well. So I knew that and I said, okay, so I'm not going to use GBT for that. But for anything content that I was creating, it was very helpful. Yeah. So that's kind of what my advice would be. Start with one and then go from there.

Let's speak a little bit to your role because this is something that's going to interest a lot of folks that are listening that haven't done this yet. I've noticed some businesses in the promo space start to do this where they've brought in a specialist or someone who is going to share a role. But you are Kaminsky's director of AI Operations. As I mentioned in the intro, it's a role most companies don't have yet. But what should you mentioned a survey and that was one of the first things you did. You also instituted an education day for our team every Friday. Speak to this a little bit. So what should a distributor look for in someone like you to help bring on an AI expert and kind of share a little bit of your role in what you're doing right now so they can sort of learn from that?

Yeah, for sure. So my role has two distinct components to it and both are very important and neither of them are tool based to be actually to be upfront. So one of them is operational excellence and operational rigor. You need people, as you look at bringing someone like myself into a role, you need someone who understands what an operational workflow looks like and that ability to kind of this plus this will equal that. There's a tactical element of my role that we're still building up where I'm going to be the person who's working directly on building agentic workflows with our company. And so doing that kind of tactical work is going to be something that's important. So that operational excellence and operational rigor is one key element. The second key element is enablement. And so what you're describing is I created an AI survey, kind of a comprehensive list of. These are the questions that I wanted to know about how people are using AI where they're comfortable, where they're not. We made that anonymous so that we could just get right, real feedback from our team to say like this is how you feel about AI right now and so we came out with some really interesting findings about how people are feeling about AI at Common sku. And so that's been something that was very important is we needed a data set. We needed to understand what, what we were working with. The second thing that you mentioned is I created an AI office hours. And so the way that I built this is the first half is first 15 minutes or so is about really sharing wins. What are people experimenting with? What are they having success with? What are the challenges that they're facing? One of the biggest elements about creating an AI culture is creating group enablement and creating that kind of culture of like, I did this with AI, this is exciting group think. And these are the challenges I faced. So that was one element. The second element of every one of those AI office hours is a thematic curriculum, essentially, for lack of a better word, where I'm working on a 15 to 20 minute presentation on a topic regarding AI. I've built out kind of 12 topics to start that are early stages, but with the full understanding that this is going to evolve. And one of the things that, you know, I was talking about even yesterday is like, we need more specifics on how we can use Gemini. It is embedded in our workspace. How do we use it effectively?

Right.

And so that was an example. There's a tactical element, that operational rigor and then there's the enablement piece, which is. Is important as well.

What, what I'm seeing you address, even with our own team, is there's this interesting balance right now between risk and also caution.

Yeah.

So we. You mentioned guardrails a few times. I've noticed there's a tendency for us to want to be very cautious about tools and naturally so. And especially for Komsky, we're a data driven company, so, you know, there's data that we have to keep locked down. But speak to that a little bit how you feel like in this moment in the market, there are folks that probably going to, just because they're biased, they're going to be more cautious. There are people going to be more aggressive. How do you feel like right now, even as you watch our own team struggle with. It is important to have sort of an 80% risk right now because you're trying things and 20% guardrails. Or is that too simplistic?

I don't think it's too simplistic. I think that it is. Be mindful of what those guardrails are. So the, the most obvious guardrail is customer data. Like you have to make a decision of if you were just using random AI tool, pick your poison, how is that data educating its backend? LLM models are consistently educating themselves. And so where is that coming from? So one of the key elements that I would think about is, you know, getting into that paid license can be very important because the thing that I've learned about freemium is that you're sacrificing something so it's free for a reason. So getting into a paid version of Gemini, getting into a paid version of GPT Claude, that at least is allowing what you create in your workspace not to be educating the masses. So that is a key element. The second thing I would use is just having that kind of common sense approach is that making sure that when you say yes to connect your Gmail or connect your, your email address or whatever, is that be aware of what those consequences are. And the best example to me of this is notetakers. We've all seen now note takers jump onto calls and incredibly valuable. But it is also something that you have to think about, where is that going? Like, where is the content going? If it's a paid license, you're probably protected. If it's not, where is that information being used? And I guess I always go back to, you know, the analogy is that if you wanted, if you're going to put something on the website or onto the Internet, are you comfortable with it being publicized on the front page of a newspaper? That's always the guardrail that I start with. And there's variations of that, but essentially I come back to is like, okay, so from a customer data perspective, I'm only going to use that in a lockdown environment where I know that it's not going anywhere. If it's something that I'm just asking about normal business practice or, you know, something like a standard operating procedure, that might be something. I don't care if it's like, it's not unique to, to how we operate. I'm more comfortable with that. And so I might go to a more open source option in that perspective.

Then let's jump to where things are going or where we think things are going. Yeah. You're a student of AI. I'm a student of AI. And for those that are listening, I mentioned this in the intro, but don't forget we have, we have the AI promo brief, which is our newsletter that we sort of look at the emerging tools and things that are happening in AI through the lens of the promo world. But since we're both students of AI, where do you see the big changes going in the upcoming year. Do you see Agentic? Do you see design, which in your mind as we move forward?

Yeah, it's a really good question and honestly, I don't know if I have a good answer for you. So I'll give you kind of my thoughts, my expectations. When Gemini came out with Nano Banana, it changed the game from a visual standpoint. This was something that every LLM had not been doing well. And all of a sudden, literally on the overnight, it went from no one was doing this well to one tool was doing this better than everyone else. And it wasn't close.

Right.

I think Agentic is climbing. People are looking for ways to make their lives easier, to make their business life easier, to make things flow more effectively and there is value and it's able to just consolidate so much more information than you have at your fingertips. You are seeing big companies invest in AI. So whether it's a HubSpot, whether it's a salesforce, you're seeing these kind of big massive players. They are spending millions of dollars investing in R and D on AI and so they are getting the most use case out of them. To answer your question, most directly, AI is embedded in everything that we do now. Phones now. It's AI functionality is super important. So I think what you're going to start to see is as we've gone so far in one direction of like AI proliferation everywhere, you're going to start to see more tactical approach AI going forward. So you're starting to see how people are using it. More specifically, they're porting guardrails up because now there's a caution element to it. We are open, but we are now starting to question, and I think this is the human condition, which is a good thing, is that we are going to protect ourselves from ourselves a little bit. And so now there is. We were so excited by the prospect of AI to start. Now it's kind of going the other way. And so now we have to kind of recenter. What I will say though is the models are changing so fast and so in such a positive way. I saw it this morning. Gemini is passing math tests, which it wasn't able to do before, and GPT is connecting to tools that didn't have that access. So it's not really a fair answer, but I think it is. The true answer is that it's evolving constantly. And so we need to understand it is if you are in something that is repetitive, if you are in a workflow that is manual, if you are in a workflow that is. That is challenging. Ask the question if it can be done with AI, and the answer is probably in some version of yes. Now, where human intervention, it will always be at the start, it will always be at the end. But a lot of those middle points can be helped or addressed in some way, shape or form with AI.

I must mention that we have a couple of big strategic initiatives for next year. One is AI and one is automation. Both of which are meant to help automate and create intelligence around the back end of your work so that you can focus on better sales and more progressive sales and working with your clients and more creative and things like that. But ecu, let's wrap up with this as folks that are listening are typically leaders in suppliers and distributors. They're leading teams. Whether it's large teams or small teams is immaterial. They're leaders and they want. They're going to head into 2026 and they want to be more aggressive with AI. What would be your final tips to them as they head into 2026?

Yeah, doing whatever version of that survey for your team, however big your team is. Right. Small teams, big teams. But getting an idea of where the landscape is for people and their comfort with AI. It's a great starting point that even came to us from our survey recently is there was things that came out of it that people were surprised by with their comfort or their experience with AI. So that's the first step is definitely understand your perspective. The second thing that I would say is really identify those pain points. What are those pain points in workflows that are causing grief right now?

Yeah.

How are people spending their time in terms of ways that they don't enjoy? That's a great place to start. The third element, kind of going back to the guardrails and the caution conversation. But understand what you're comfortable with from an AI perspective. You know, and a great example of sales, right. Is that people look at sales and they say, oh, I need to be in person and I need to actually have a human interaction with my sales team at and especially with distributors. And my understanding of this industry so far is that people really pride themselves on. And that's amazing.

Right.

So then we have to figure out where does AI contribute to that. But then you're putting a clear guard rail up, which is really important for you and your company to say, okay, we're not going to have AI do SDR calls, for example. We're not going to have them do business development. Another one is interviews so now you could absolutely have a full interview process that's done by AI. It's not as valuable potentially. And so now you're putting up a clear guardrail as a company to say, okay, I'm still going to have a human interaction with a zoom call or in person interview, whatever that looks like. So starting to think a little bit about what don't you want AI to do and then work backwards is also really helpful. But I'll kind of finish the answer with just saying this is that again, if it's something that is causing pain, if it's something that is repetitive, if it is something that is manual, I would be very interested about using AI for it because the doors of efficiency that it will open, it will allow you to focus your team on the things that they want to focus on as opposed to the things that they have to focus on.

That's a great way to end this, Ecku. Thank you, my friend. You're going to be back with us again because I think you're going to be recurring guests this upcoming year as we go through different models and evolutions as things happen. So stay tuned everyone. We'll have Ecku back on, but thanks, my friend. Thanks for joining.

Awesome. My pleasure. Thanks so much.