Why "No Alerts" Can Mean Success: Colin Morris of MaintainX on AI, Digital Logs, and Predictive Maintenance
Manufacturing Happy Hour
Predictive maintenance powered by AI has been promised to manufacturers for years, and in many plants it still hasn't delivered. In this episode of Manufacturing Happy Hour, host Chris Luecke asks Colin Morris, Senior Director of Solution Consulting at MaintainX, why predictive initiatives stall, what data has to exist before AI can help, and where AI actually fits in day-to-day maintenance work. Colin's position is that the data mostly exists but is scattered and unstructured, that the foundations are often skipped, and that AI's biggest near-term value lies less in exotic predictive models than in supporting technicians' routine work, especially scheduling, voice capture, and knowledge transfer.
Revisiting the data question
Chris frames the conversation around an interview from roughly two years earlier with Nick, a MaintainX co-founder. At that point AI discussions in this space were just beginning, and Nick had argued that AI only works in a maintenance environment if it has good data. Chris asks whether manufacturers now have that data and infrastructure.
Colin believes many customers and prospects do have the data. The problem, in his view, is its form: it is highly unstructured and spread across disparate systems. Centralizing it in an easy-to-understand format that AI can "layer on top of" remains a major challenge for the industry. He also notes that some companies pull from data sources that are on-premise only, and moving that data into a cloud solution can still be difficult.
Moving beyond the predictive-only mindset
Asked about other gaps between the promise of AI-driven predictive maintenance and what frontline workers actually experience, Colin describes a longstanding default in the industry. When people think of AI in maintenance, they think of predictive or prescriptive maintenance: how can the machine tell me what's broken, and how can I make sure it doesn't break again? Historically, approaches such as vibration analysis and fault classification have required substantial resources, expertise, and time.
Colin sees a larger opportunity in using AI to supplement technicians' everyday work. His questions are practical ones: how to make it easier to repair assets, and how to surface information and data quickly, instead of building up "a very particular set of individuals with a very particular set of skills" just to get predictive insights.
What the platform does, in plain terms
Chris asks Colin to explain the job as if talking to someone over a drink. Colin jokes that after eight years in the industry, his wife still doesn't know what he does. His description is that MaintainX is a tool that makes maintenance people more effective, presents digital work instructions for performing tasks in a facility, and simplifies asset management. In his shorthand, the goal is making sure machines make widgets "faster, better, and easier."
Chris offers his own framing from when he worked in the space. Manufacturers talk constantly about digital transformation of the factory or enterprise, but Chris liked to describe a "micro transformation" at the maintenance level. That means getting rid of Post-it notes and spreadsheets that track recurring tasks every 28 days or every week, and moving everything into a single cloud-based source of truth that is accessible from a desktop or a phone.
Why predictive initiatives stall: no news can be good news
On why predictive projects get stuck, Colin says people often expect predictive or prescriptive analytics to "solve world hunger." In practice, results depend on where in its life cycle you begin monitoring an asset. If the asset is in a healthy phase, there may simply be nothing wrong with it, and you can go long periods without detecting an issue or surfacing any insights.
He compares this to a car. If it runs well, you don't take it in, and you can't say it is about to fail because you haven't heard or seen a problem. If monitoring begins while an asset is performing well, nothing will happen. In his view, people often fail to understand that "no insights are sometimes good insights": things are running well, or repairs are happening at the right cadence. He describes a kind of confirmation bias in which people really want the system to flag something, when "no news is sometimes good news."
Chris agrees and relates it to software in general. The less he has to think about a tool, the more confident he is that it is doing its job. A quiet system can show that processes have moved the team from reactive into proactive work.
The skipped foundation: knowing what's in your facility
Chris asks whether companies skip foundational steps when they digitize maintenance. Colin says some people expect to stand up an application in seconds. The biggest challenge still facing the market, he argues, is that many facilities lack good digital records of what equipment they actually have. That can mean a period in which a company has to outsource the work or send an intern around the plant to collect data that exists nowhere else. Only the facility itself can make sure its data is accurate.
Colin singles out parts information. Replacement components often have some of the longest lead times, and without good parts data a plant ends up waiting, or paying heavily to have something overnighted. A full foundation, in his view, includes maintenance plans and asset plans, knowledge of existing parts and inventory, and a plan for how other data sources connect. That means working out whether data will come from an ERP, a QMS, or an MES, and how those systems will talk to each other. He calls it a progression rather than an overnight journey. He also says it's fine to start with a standalone application, since "you don't need to have all the bells and whistles from day one."
The minimum parts data to capture
Chris asks for an actionable list: if someone is sent out to walk the floor, what should they record about parts? Colin's answer covers several items:
- Quantity: how many of each part you have.
- Location: where it sits in the facility, ideally down to aisle, row, and bin if such a system exists. This reduces the time technicians spend hunting for parts.
- Obsolescence: whether parts are still useful. Colin says customers frequently report things like a motor that has sat on a shelf for 25 years with nobody knowing where it goes. Such items should probably be disposed of or sold.
- Asset linkage: a direct correlation between each part and the piece of equipment it belongs to, which he calls a major process in itself.
- Classification and unit counts: good parts classification and accurate counts.
Colin notes that this last group becomes especially important once a company pursues integrations, for example exchanging data between an EAM and an ERP.
Eight years of change: from cost center to cost saver
Chris points out that the older label, CMMS (computerized maintenance management system), now feels somewhat antiquated, and asks about the biggest change over Colin's eight years in the field. Colin names the shift from treating maintenance as a cost center to seeing it as a cost saver. Historically, companies didn't want to spend on maintenance or on software for it, and they kept deferring. Now, as manufacturing returns to North America amid a labor shortage, Colin says companies can't let maintenance remain something that happens in the background. More of them are investing in maintenance and reliability to better understand their equipment and reliability metrics. In his words, maintenance has moved from "the forgotten child" to a key component of digital transformation projects.
The past two years: AI across the whole workflow
For the more recent period, Colin says AI has accelerated this change, with more solutions offering AI tools. The focus used to be predictive and prescriptive analytics, but he now sees AI applied across all segments of the work. He describes it as a tool in the tool belt for getting information and understanding what needs to be done. It doesn't have to act entirely on its own. With a human in the loop, AI can take over repetitive tasks it understands well. He expects another significant leap with agents and more tooling that simplifies the overall maintenance execution journey.
What an agentic future looks like: scheduling first
Colin expects humans to remain in the loop, but he sees agents taking the first pass at actions. The use case he hears most often is scheduling and planning. Schedules are known, and so is who will be available when. An agent could review all pending jobs, their prioritization and criticality, and propose something like "I think Bob should do these tasks and here's why," which a person then accepts or declines. That, he says, is where agents will start. Over time he expects more activities to become fully automated, possibly to the point where scheduling or planning is no longer anyone's full-time job and people simply approve plans or reject them with a reason.
Chris says scheduling is probably the number one topic that comes up on the show when people discuss where AI or agents make sense, and asks how schedulers feel about it. Colin calls it a very challenging job because it requires coordinating across teams with competing priorities. Operations doesn't want to take a line down for maintenance, and maintenance doesn't want to work overtime. He describes the scheduler as "playing traffic cop." Chris draws on a past role as a project engineer at Metal Container Corporation, which made aluminum cans. There, some people wanted the line back up as quickly as possible while others wanted new guarding installed, and Chris agrees the scheduling role is stressful.
Where AI doesn't make sense yet
Chris asks where people might rush in expecting AI to solve their maintenance problems when it isn't ready. Colin's answer is that, at this point, AI can do a lot, though not perfectly. He says he can't think of anything it doesn't do "really, really well in the grand scheme of things." His framing is that AI can handle about 80% of a task, with a human needed for the last 20%. He expects that to remain the case for the next few years, while noting that the technology keeps improving and that each new model release from OpenAI's ChatGPT or Anthropic is very impressive.
Tribal knowledge, and why voice matters
Returning to another thread from the earlier interview with Nick, Chris raises skilled workers retiring and taking their knowledge with them, and asks how the industry has handled it. Colin's answer is blunt: a lot of companies haven't. Many are letting tribal knowledge walk out the door. Some try to capture it with pen and paper or video and enter it into a CMMS or EAM application, but in his view most of the market hasn't really gone back to address the problem.
Colin says MaintainX is heavily focused on voice as a way to capture data from users. His reasoning is that people no longer want to type on their devices. The days of good physical QWERTY keyboards like those on BlackBerry phones are over, and typing can be difficult for some workers. He expects voice to become increasingly prevalent in the industry and says it is already in use.
He describes two workflows. The first is creating work. Instead of typing, a user taps a button that opens a voice transcription service MaintainX built in-house. The system captures what the user says, enters it, and can automatically surface work instructions. Colin highlights translation as something AI does well. A worker who speaks Spanish at a business that operates in English can speak Spanish, and the system will create the work instructions in English.
The second is work order completion. After an experienced technician finishes a preventive maintenance task or a reactive work order, they can give a spoken high-level summary of what they did, along with any recommended changes to PMs or other observations, rather than typing out notes. The transcription service captures this and generates a summary. What Colin calls unique about MaintainX's approach is that it also adds the content to a knowledge base. The aim is to transfer knowledge from a 25-year veteran to a technician with two or three years of experience, which he says is increasingly typical now. Since manufacturers have to do more with less and with less-skilled labor, he believes anything that helps upskill the workforce will pay off in the long run.
Chris says the translation capability speaks to the diversity inside manufacturing facilities and welcomes anything that makes the industry more accessible.
Adoption and wrench time
Chris asks whether easier creation and closure of work orders is increasing work orders or usage. Colin says only that MaintainX is seeing a lot of adoption, which is all he can share, along with very positive customer feedback about not having to go back to a computer to enter data. He cites a figure that historically about 60% of technicians' time has gone to administrative work. Even a 10–20% gain in efficiency, and thus more wrench time, would be a big saving for customers. He believes the gains are probably much higher given all the software's functionality, but he doesn't give a specific number. His concrete comparison is between speaking to a device for 30 seconds or less and getting an automatic summary, and an older technician like "Bob" who might need considerably more time to type the same thing on a keyboard or phone.
Is manufacturing getting cooler?
Chris introduces a new hypothetical worker, "Sally," someone new to the industry, and asks whether manufacturing's growing technological profile is drawing in people who might otherwise have gone to build apps in the San Francisco Bay Area, a theme from the earlier interview. Colin thinks so. He points to a resurgence over the past five to ten years and says many people want to work for companies like SpaceX, Blue Origin, or Tesla, North American manufacturers of a kind that he says hasn't really been prominent in the past 20 years. In his view, locally built technology and ongoing digital transformation projects are making manufacturing "a little bit more sexy."
He adds that many people in the industry see a real opportunity to help it mature and adopt more technology, and that this gives them purpose. It does for him: his first real job was in manufacturing and he hasn't left since. He calls it an awesome industry and says the current resurgence makes it a good time to get in.
What's exciting Colin now: agents
Asked what excites him most, including in developing MaintainX as what Chris calls an AI-first work execution platform, Colin again names agents. He expects a large shift in how people do their jobs day to day, because much of that work can be automated. He compares it to earlier transitions from typewriters to computers, and from computers to email and connectivity, each bringing large efficiency gains. He sees AI as the next major innovation that will make work easier and faster, and says he is excited about upcoming AI functionality at MaintainX, particularly agents, though he doesn't describe specific features.
When Chris asks what he should have been asked, Colin says he thinks they covered the core topics. The episode ends with the two planning to continue the conversation over another beer at Left Field Brewery, the former-factory taproom in Toronto that framed the discussion.
What's exciting you right now? Agents are really exciting to me. It's going to be a huge shift in the way that we do our jobs in a day-to-day because we're going to be able to automate a lot of it. AI is just the next huge innovation within the industry that will help to make things easier, make things faster, fundamentally.
Maintenance folks are going to want to listen to this episode. In today's episode, we're covering artificial intelligence impact on predictive maintenance. Colin Morris of MaintainX knows this topic well, and we'll explore why a bunch of predictive initiatives fail so that you can avoid those mistakes as you begin to put AI-driven predictive maintenance into practice in your own operation.
Colin Morris, a long-time industry colleague, a first-time Manufacturing Happy Hour guest, welcome to the show.
Great to be here, Chris. Really appreciate it.
And you and I have hung out before in your home of Toronto, Ontario, up in Canada. In the spirit of Manufacturing Happy Hour, if we were back there right now, where would we be having a drink, having today's conversation?
I think I'm going to go with Left Field. They've got two different locations. Their original location is in the east side of Toronto, former factory, so goes along well with this conversation.
I love it. So, in true brewery fashion in an old industrial facility repurposed to make beer instead of whatever the product that was there before. So, let's say we're hanging out in Left Field. I'm interested to get your take on this question around artificial intelligence because I'll lay some groundwork.
We have featured MaintainX on Manufacturing Happy Hour before, but it's closer to 2 years ago now than it was like a recent interview, and Nick Cos, co-founder, we were having a conversation around where digital maintenance, where work execution was at that time, and there was a big focus around kind of those initial discussions about artificial intelligence and he was saying it's like, "Hey, for artificial intelligence to work in our environment, it requires good data." So, I'm very interested to hear your take, Colin. Do we have the data, that infrastructure there? Do manufacturers have that now to make artificial intelligence in the context of work instructions and manufacturing in general work for them?
I think a lot of customers and prospects within the industry do have that data. The biggest challenge that we face right now is it's super unstructured, it's in disparate systems, and so ensuring that we have all of that data centralized in an easy-to-understand format that AI can, you know, layer on top of is still a huge challenge within this industry. People are pulling data sources that, you know, maybe on-premise only, and so figuring out ways to get that into a cloud solution can still be pretty challenging in today's market.
Okay, so I'm going to follow on with that. Are there other gaps between the promise of AI-driven predictive maintenance in this case and the operational reality? You talked about some of the technical highlights there, but what does this maybe look like in practice to the folks on the front line?
I think historically within the industry when it comes to AI, people have always thought of predictive or prescriptive maintenance. They think how can my machine tell me what's broken and how can I make sure that it doesn't break again? And so historically, especially for things like vibration analysis, as well as, you know, fault classification, all of that, it requires a lot of resources, a lot of expertise, and a lot of time.
And so I think that is always been the default mindset is I need to have those types of tools in place where I think that there's a huge opportunity to layer AI on top to help supplement the day-to-day of our technicians. How do we make it easier for them to repair assets? How do we get it easier for them to surface information and data versus needing to build out a very particular set of individuals with a very particular set of skills to get those predictive analytics and insights that they need immediately as an example.
Yeah, we'll dive into more of these details as we get further into the conversation, but I'm going to ask more of a baseline question first as if we're hanging out at Left Field. Assuming not everyone is spending their everyday lives inside a digital maintenance platform work execution, how do you typically describe what you do as if you're hanging out with someone over a drink, right? I know some of our audience is going to be familiar with this, but there probably others that are a little less familiar than they should be.
I think my wife still doesn't know what I do and I've been working in this industry for eight years.
So, I would describe it as we are a tool that helps make maintenance individuals more effective, presents digital work instructions on how to actually perform different tasks within a facility, and helps to simplify asset management fundamentally. So, what does that mean? Making sure machines make widgets faster, better, and easier.
Yeah. Yeah, it's funny when, you know, since you and I have worked together in this space before, when I first started working in this arena, I would describe it as, you know, manufacturers talk a lot about digital transformation, right? How they transform their operation, their factory, oftentimes their enterprise. I always liked talking about how you can start with like a micro transformation, right? Imagine just doing that digital transformation, but at the maintenance level. Empower those folks that are out there turning the wrenches, get rid of the Post-it notes, get rid of the spreadsheets that say, "Hey, we need to make sure we do this maintenance task every 28 days, every week, whatever it may be," and put it in a single, you know, point of truth platform that you can access from your desktop, from your phone, etc., cuz it's all cloud-based. That was my way of describing it as a way to make it a micro transformation platform for the maintenance department. So.
Yeah, exactly. So, let's go into some of where this space is right now. And if I think about predictive initiatives, they don't always work out. They don't always go through. There are reasons they get stuck, they stall. What are the reasons, what are the pitfalls you sometimes see folks run into, Colin?
Yeah, a lot of the time people think, as I mentioned, predictive or prescriptive analytics are going to solve the world hunger. When it comes to building up these models and actually putting them into practice, it really depends on where you're capturing an asset in its life cycle. There may not be anything wrong with that asset. And so, you may go long periods of time without actually detecting an issue and surfacing insights. And so, I think people forget that, you know, your car if it's running great, you're not going to go and take it in and do maintenance. And you're not going to be able to say, "Hey, I actually think this is going to fail because I hear or see an issue." And so, if you capture an asset on that, you know, really good performance stage, nothing's going to happen.
And so, people don't understand that no insights are sometimes good insights. That means things are operating well, that means things are getting fixed or getting fixed at the right cadence. And so, there is a little bit of, you know, confirmation bias that happens at times where people really want something to happen, but no news is sometimes good news.
Yeah, it's a great way to describe it. I always like, when I talk about this with any software platform, the less I have to think about it, the more I know it's doing what it's supposed to do, right? It's where there constant hiccups, constant issues, that's when yes, maybe it is helping me reveal those things, but it's nice when it's showing that, hey, many things are working or we've started creating these processes that are allowing us to live in the proactive space versus the reactive space.
When someone is going through a transformation might be a strong word, but when they're digitizing, when they're putting in a work execution platform, are there aspects that get skipped sometimes? Are there some foundations to a facility that aren't there that people are maybe going into step three when they should have gone back to step one first? What do you see?
There are definitely people that think, you know, you're going to stand up an application in seconds. And so, I think one of the biggest challenges that still faces this market today is the fact that people don't have good digital record logs of what's actually at their facility. And there is going to be a period of time where you may need to either outsource or get an intern to walk around and collect that data because you don't have it anywhere. And so, that is, you know, one of the challenges in the process. Nobody can actually ensure that you have accurate data outside of yourself. And so, actually going through collecting parts information is usually important because that is typically one of the longest lead times is getting a replaceable component. And if you don't have that, you're going to be in trouble. You're going to be waiting. You're going to be paying a whole bunch of money to get something overnighted to your facility.
And so, actually building out not only maintenance plans and asset plans, but also understanding the parts and inventory you have in place as well as the other data sources, you know, are we getting data from an ERP? Are we getting it from a QMS, an MES? How are we going to actually have these systems talk together and be able to function the way we expect them. And so, it's not a journey that happens overnight. It's a progression. And it's fine to start as a standalone application, is what I would say. You don't need to have all the bells and whistles from day one.
Yeah, well let's go into this a little bit further cuz I like that you say, "Hey, you need to have an intern, for example, walk the floor, get a good digital log, understand the part information, understand data from other spots." You mentioned, you know, the MES, QMS as well. Quick question for you. Like, if we're going to give an actionable insight to the audience out there and they're going to send someone out to walk the floor, what is the parts information that they should be looking for? What's the bare minimum information that they should be capturing while they're out there?
Yeah. How many you have? Where it's located in the facility. So, actually understanding aisle, row, bin, if you have a system in place is going to reduce the amount of time that it takes your technicians to actually find those parts and use them. Ensuring that you don't have obsolete parts. A lot of the time we hear from our customers that they have, you know, a motor sitting on a shelf that's been there for 25 years and nobody knows where it goes. You're probably going to find some things that should either be of or sold. And so, actually ensuring that you have a direct correlation between that part you collected and the piece of equipment it goes on is a huge process as well.
I would also say, you know, ensuring that you have good parts classification, unit counts, all of that is going to be critically important as you build out that knowledge base. And it's especially going to be important if we start to talk through, you know, ERP integrations and how that data is going to exchange between an EAM and ERP as an example.
Yeah. Well, one thing I want to ask you based on your personal experience is you made the comment you've been doing this for 8 years. Your wife doesn't always understand what it is you do. And the last time we really talked about this, on Manufacturing Happy Hour, was 2 years ago. I'm very curious to hear your thoughts on what has changed over the past 8 years. What's the most significant evolution you've seen in this space from the time that we were often referring to it as, you know, CMMS, right? Computerized maintenance management systems, right? That term is a little antiquated in some ways. What's the biggest evolution you've seen over the past 8 years? And then I'll ask you about just the past 2 years shortly.
Sure. I think one of the biggest things that I've seen at least in manufacturing is the shift from maintenance being a cost center to being a cost saver. Historically, people didn't want to spend money on maintenance. They didn't want to spend money on getting these types of solutions in place. And so, they put it off longer and longer. And now that, you know, manufacturing is coming back to North America, we're seeing a labor shortage. They can't let maintenance be, you know, that thing that happens in the background. And so, more people are investing into maintenance and reliability to get a better understanding of their equipment and reliability metrics. And from there, we're seeing a shift from it being the forgotten child to one of the key components within a digital transformation project or any manufacturing environment.
So, we were just talking about how things have evolved over the past 8 years. What about just in the past 2 years? I mentioned we were just starting the artificial intelligence discussion in this space at that point. What are some of the biggest strides you've seen during that time?
I think AI really has accelerated this process. More and more solutions have different tools in place. You know, historically as we were talking about, a lot of the focus has been on predictive and prescriptive analytics. We're now seeing AI leveraged across all segments. The AI can be a great tool to have in your tool belt to get information, understand what needs to be done. It doesn't necessarily need to do everything on its own. We can have AI with human in the loop to help simplify processes and, you know, take away a lot of the work that is typically done by humans that AI is really good at. Repetitive tasks that it can understand and have skills and just do. And so, I think we're about to see another big leap when it comes to AI functionality with agents and more tooling when it comes to simplifying the overall, you know, maintenance execution journey.
Colin, can you paint the picture of what that agentic future looks like?
Yeah. In my opinion, we are going to continue to have humans in the loop. But agents can go out and take the first pass at an action. The one that we hear a lot of the time is when it comes to scheduling and planning. You always have a set schedule. You always understand who's going to be there when and why. And so, why not have an agent go in and schedule your work orders. Look at all the jobs that need to be done. Look at the prioritization, criticality and say, "Hey, I think Bob should do these tasks and here's why." Accept or decline. That is what agents are going to be to start. And then we're going to start to see more and more activities that are just going to be fully automated. We're going to get to a point maybe one day where there isn't a scheduler or planner or it's not their full-time job. And they are just going in and then saying, "Yep, looks good." Or no, that's not good and here's why.
Yeah, and for the Manufacturing Happy Hour audience out there that listens to this show on a regular basis, scheduling seems to be, I would go as far as to say it's the number one topic that comes up as far as where does an agent make sense or where does artificial intelligence make sense in general? Out of curiosity, how do the schedulers feel about that? From what I, you know, I haven't talked, like as I understand it, that's not someone's favorite task to do anyway. I'm curious if you've heard similar.
I think it can be a very challenging task. Why? Because you typically got to coordinate across a whole bunch of teams with competing priorities. Operations doesn't want to take down the line to do maintenance, maintenance doesn't want to have to do overtime and so trying to drive that
right balance can be difficult. So, I feel for a lot of people that are in that type of position because, you know, you're playing traffic cop. Right, to a lot of extent.
Right. No, it's a very good way to describe it. You are a traffic cop and you are looking at a lot of competing priorities, as I recall from being the project guy back in my days working for Metal Container Corporation making aluminum cans, right? Like there were people that wanted to get that line back on as quick as possible and there were people that want to get the new guarding installed and there are all these elements that come into play. So, the job of schedulers, I think stressful is a very good way to describe it.
We're going to talk a bit about the humans in the loop here again in a second, but I did want to ask. So, you just gave a great example of where agents and artificial intelligence make sense. Where doesn't it make sense right now? Where is the spot that some people might rush into and think artificial intelligence can solve their problems from a digital work instructions maintenance standpoint, where doesn't it make sense today?
I think at the place that we're at within the market today, it can do a lot of things. Can it do it perfectly? No. But, it can help to supplement what is typically taking a lot of people to do and help to enhance that process.
And so, I can't think of anything that it doesn't do really, really well in the grand scheme of things. I think it can do the 80% and the last 20% you're going to have to have a human in the loop, and that's probably going to be how it's going to be for the next few years, but it's getting better and better. Every time that they release a new model from ChatGPT, from Anthropic, the things that it can do are very impressive, to say the least.
Yeah. Yeah, and you gave the answer earlier where you're like, "Hey, I believe humans are still going to be in the loop, right? The agents can take step one in scheduling, right?" But, you still need, I believe you used the same example maintenance person that I've used many of times before, Bob, you know, is still going to need to be there doing the work.
Speaking of Bob, speaking of humans in the loop, one thing that came up when Nick and I were talking about this topic a couple years ago was skilled labor walking out the door, tribal knowledge walking out the door. Yeah. How are we capturing that? How have we been capturing that over the past 2 years and longer? Because since then, we look at the stats, right? There have been a lot of people retiring and walking out the door. So, how have we been solving that problem?
A lot of people haven't. A lot of people are letting that tribal knowledge walk out the door. There are some people that are trying to capture it, pen and paper, videos, and trying to input it into something like a CMMS and EAM application for work execution, but I would say the majority of the time the market hasn't actually gone back and tried to address this issue.
I know at MaintainX we're really focused on voice. How are we going to be able to capture data from voice from our end users? Why? Because people don't want to sit there and type on their devices anymore. It's not the days of the QWERTY, you know, nice keyboards that we used to have on our BlackBerrys. It's sometimes a little bit difficult for folks, and so I personally think voice is going to become more and more prevalent within this industry and we're already starting to use it.
Yeah, go a bit further. What does that look like when someone's in the maintenance shop describing their task and creating a work order from voice?
Yeah, so rather than them having to sit there and type things out, they can click onto a button, opens up a voice transcription service that we've built in-house. It's going to capture that, input it into our system, and surface work instructions automatically as an example. The huge thing here as well, and again, something that AI does really well, is translation. So, if I speak Spanish, but my business operates in English, I can speak Spanish to it and it will automatically create work instructions in English as an example.
The other way that we're doing it is on the work order completion side of things. So, as an experienced technician, I've gone out, I've completed my job, I did my PM or my reactive work order, rather than me having to sit there and type out all of my observations, all of my notes, why not just open up our transcription service, provide a high-level summary of exactly what I did, any recommendations on changing our PMs or any other observations, and let our voice transcription capture it, provide an automatic summary, and then the other unique thing I would say about ours is it adds it to the knowledge base.
So, we are taking that knowledge from that 25-year veteran, giving it to our, you know, two or three-year veteran at this point. That's what it is these days, and upskilling those technicians, because that's the biggest thing right now. Manufacturers have to do more with less, with less skilled work, and anything that we can do to help upskill that workforce is going to pay dividends in the long term.
Yeah, that's a very real picture you painted, right? Because there's a lot of diversity inside of a manufacturing facility, right? And the ability to speak in one language, have it translated to another for the work instruction is huge. I love anything that makes manufacturing more accessible to more people. And you know, on top of that, I'm curious, have you seen, because of this ease of creating work orders, closing out work orders, etc., I think this is something you track in the industry, are you seeing like an increase in work orders and usage as a result of some of this new functionality?
I would definitely say we're seeing a lot of adoption. It is all I can share at this point from our perspective, is we are definitely getting a lot of very positive feedback from our customers on the fact that, you know, they don't need to go back to a computer and enter data.
Historically, 60% of the time for our technicians has been focused on administrative work, and so even if we're able to get 10, 20% more efficiency, more wrench time, that's a huge savings for our customers. I would say we're probably getting a lot higher than that with all of the different functionality that we have within the software. But you know, there is a significant time savings and value for people to be able to talk to their device in 30 seconds or less and have it automatically summarize it versus, you know, Bob, who's a little bit older and may need a little bit more time to type it out on a keyboard or on their phone.
Yeah. Yeah, I mean anything that lowers the barrier to usage, right? Makes the user experience more seamless is no doubt going to have an impact on adoption.
We've been talking about Bob a couple times here. We're going to pick a new character in maintenance. So, let's say Sally. I used to always talk about Jack and Sally as my go-to maintenance characters when I was back in that space.
Is manufacturing getting cooler? Again, I want to hear your take on this because when I was talking to Nick, and I promise this is probably the last time I reference that last interview, but another interesting through-line is we were talking about people leaving the industry. We're talking about people coming into the industry as well. And we had been reflecting on our times living in the San Francisco Bay Area, where people were working on apps, and that there was a real opportunity for people to see the promise and the excitement and the necessity around manufacturing and get back into that industry. Do you see us making the manufacturing space a place that folks like Sally, who haven't worked in this space before, are now starting to jump in because we've become a more high-tech industry, etc., etc.? What are some of your takes on that, Colin?
I think so. I think that there are going to be more people that are going to be interested in manufacturing. You know, we've seen a resurgence over the past 5 years, 10 years at this point. I think there's a lot of people that want to work for somebody like SpaceX or for Blue Origin or for Tesla, some of these companies that are based in North America that, you know, historically hasn't necessarily been within the past 20 years. And so, I think people are starting to make manufacturing a little bit more sexy when it comes to the technologies that we're building locally, as well as the digital transformation projects that are underway.
I think a lot of folks that, you know, work in this industry see a huge opportunity to help this industry mature and adopt more technology, and so I think that gives a lot of people purpose. At least it does for me. I started, my first real job was in manufacturing and I haven't really left since, and so I think that it's an awesome industry to be in, and the resurgence that's happening right now is also a great time to get in.
Well, this leads me to another opportunity to ask a more personal question. What has kept you excited most recently? And you can even take us under the hood a little bit, because we've covered a lot of these topics from what I would call like a very good podcast-centric standpoint, right? General audience can follow along. Are there maybe some areas, it could be more of a macro area you're excited about right now, but as far as, you know, developing your platform as really an AI-first work execution platform, I mean, what's exciting you right now?
Agents are really exciting to me. You know, I think that there is, as I mentioned, going to be a huge shift in the way that we do our jobs day-to-day because we're going to be able to automate a lot of it. You know, I think we saw this resurgence as we went from moving from typewriters to computers. You get a lot more efficiency, from computers to emails and connectivity, and so I think AI is just the next huge innovation within the industry that will help to make things easier, make things faster, fundamentally. And so I'm really excited about what we have coming up in terms of AI functionality, and especially agents. And so I think we are going to see a lot of shift and change within what is up and coming.
As we get to the end of our conversation, what is something you wish I would have asked you that I haven't asked you about yet?
I don't know. I think we hit the core items, and maybe it's because it's almost 3:00 p.m. on a Friday.
That's true. That's true. The brain is running a little slower, but yeah, I honestly can't think of anything that I wanted to highlight, at least.
That's fair. Actually, I have something that I can ask you. Normally when I end these interviews, I ask the question like, "Hey, what's something you wish I would have asked you?" But I have something that I wish I would have asked earlier. I didn't really get like a full description on what characterizes Left Field, because I think we're done with round one and we're probably going to need to grab another brew. You know, take us inside there. What is a Friday afternoon like, like we're recording this right now? What does that look like?
Yeah, it depends on which one you go to. It's a dog-friendly location, so you can bring your pups. There is, yeah, wide open space. They typically have some board games. In the summer they've got, you know, cornhole and some outdoor games as well. But generally speaking, big hall, lots of different families, lots of different people sitting at different tables and just enjoying a nice cold beer.
Well, next time we do this, whether it's two years from now like last time, I think we need to be cracking beers in Left Field back in Toronto. I needed an excuse to get back up there into Canada.
Nice. Yeah, looking forward to it.
Sounds good. Well, hey Colin, always good catching up. I know I'm going to be seeing you around a bunch of shows this year as we get further into 2026. Thanks so much for being on Manufacturing Happy Hour.
Thanks, Chris. Really appreciate it.
Article published