The Digital Project Manager

How to Know If Your Team’s Ready for AI in PM Software

• Galen Low

Share episode

The AI hype in project management software is real—but is everyone ready for it? In this episode, Galen sits down with returning guest Olivia Montgomery, Associate Principal Analyst at Capterra, to explore the findings from her 2025 Project Management Software Trends Survey. They unpack the real reasons behind the surge in demand for AI-enhanced PM tools and the foundational work teams need to do before expecting AI to deliver real ROI.

Together, Galen and Olivia dig into what "AI readiness" actually looks like—technically and culturally. They discuss how competitive FOMO, billion-dollar marketing campaigns, and shifting economic investments are driving decision-making at the executive level, while the realities of adoption, data governance, and employee empowerment are playing out on the ground. They also take a thoughtful look at how PMs can avoid common pitfalls (like AI hallucinations) and begin to build workflows that align with both human and machine strengths.

Resources from this episode:

Galen Low:

When organizations are seeking out PM software with AI features, do they actually have an idea of what they want to get from them, or are a majority of these buyers just kind of following a mandate to do AI stuff?

Olivia Montgomery:

We're definitely leading with a bit more of a competitive FOMO that executives are having. There's a lot of marketing, vendors are definitely rushing to catch the AI wave.

Galen Low:

What needs to be in place before a team or an organization can actually benefit from AI features in project management software? What in your mind are like the foundational pieces that help teams and organizations accelerate towards the ROI of AI?

Olivia Montgomery:

There's two pretty big considerations that I think companies of all types and all sizes, all industries, all levels of maturity should be taking into consideration. Your technical readiness and your cultural readiness.

Galen Low:

Welcome to The Digital Project Manager podcast, the show that helps delivery leaders work smarter, deliver faster, and lead better in the age of AI. I'm Galen, and every week we dive into real world strategies, new tools, proven frameworks, and the occasional war story from the project front lines. Whether you're steering massive transformation projects, wrangling AI workflows, or just trying to keep the chaos under control, you're in the right place. Let's get into it. Today we are talking about the increasing demand for AI enhanced features within project management software, what folks want from those features, and what businesses and teams of various sizes need to have in place to achieve their goals. Back in the virtual studio with me today is Olivia Montgomery, Associate Principal Analyst for Project Management at Capterra. Olivia is a former PMO lead, a project management professional, a prolific speaker, and just the best kind of nerd when it comes to research backed insights on project management, technology strategies, and the human side of leadership. She's also just published her latest research paper, Capterra’s 2025 Project Management Software Trends Survey, which reveals a shift in how companies are choosing and using project management tools. That's exactly what we're gonna be diving into today. Olivia, thank you for coming back in the studio and joining me today.

Olivia Montgomery:

Thank you for having me. I'm super excited to be here. We got some hot topics to hit.

Galen Low:

Yeah, we're gonna get spicy today. For folks who have listened to the podcast for a while now, Olivia is just a fan favorite. We were just in the green room jamming about linguistics. Then we realized we're like actually linguistics and the liberal arts and LLMs and AI, like it all fits together. So we're gonna go a lot of different spots. We do tend to go down some rabbit holes, and I do hope that we do that, but just in case, here is my roadmap that I've sketched out for us today. To start us off, I wanted to just get one big burning question out of the way. Just that like uncomfortable, hyper-relevant question that everyone wants to know the answer to. And then I'd like to zoom out from that and talk about three things. First, I wanted to talk about what is needed in order for organizations to truly take advantage of AI features within project management software. And then I'd like to deep dive into like what AI features actually look like in practice and how to measure whether or not they're actually working or how they're being used. And then lastly, I thought maybe we could just talk about the impact near term and long term of some of the AI features in project management software and elsewhere in the way that we collaborate and how that's gonna impact project based work.

Olivia Montgomery:

I am ready to go. Let's do it.

Galen Low:

Let's start with the one hot question then. In your recent research paper, you've noted that 55% of all project management software buyers are now prioritizing AI features. That's probably not surprising to most of our listeners, given the pressure for businesses to do more using AI, but when organizations are seeking out PM software with AI features. Do they actually have an idea of what they want to get from them, or are a majority of these buyers just kind of following a mandate to do AI stuff?

Olivia Montgomery:

It's a mixed bag out there. We're definitely leading with a bit more of a competitive fomo that executives are having. There's a lot of marketing vendors are definitely rushing to catch the AI wave because there is this hype and demand that is going around, but we are slowly moving toward, slowly a more strategic intent. With AI, the kind of like initial like sparkle of it has kind of worn off. We are starting to see a bit more what these tools actually can do, what they can't do, how teams use them, when, why, where, and we're getting to the tougher questions. They're still pretty heavily driven by this like competitive fomo, and that is definitely driven by, I don't know if you've heard the news, but the spending for AI infrastructure, hardware, software, data centers for the US has surpassed consumer spending. So for the first time in history, the USD GBP is being grown by AI investments. By businesses not consumer spending. So that should tell you a lot about where the conversation is going. These investments are also getting billion dollar marketing campaigns behind them. And so that's what we're seeing and that's a, it's a big driver of this competitive fomo and there are a lot of executives and business leaders that are excited. They hear about this new tech that works really fantastically and they want it, of course. Now our hands are getting into it and we're kind of starting to see like, oh, okay, so marketing doesn't always line up with the technical capabilities that we see. And then the technical capabilities don't always line up with what your people and your team are skilled and trained and willing to do. So there's definitely a disconnect That's, you know, we're right in the middle of, and I'm sure probably everybody listening is feeling some kind of pressure some way about this.

Galen Low:

I thought it was interesting that you brought up the economics of it.'cause I hadn't really thought about it. In some ways, when you're talking about that level of investment, like surpassing consumer spending in terms of how economies are making money, imagine the pressure that these companies are under, they've invested in infrastructure, they're spending, I remember you know, we were all taught that like, oh, cloud storage is cheap now. Yeah. Not when it's petabytes of data. You know what I mean? Like there's this pressure that filters down and if you're a marketer in the software space. I salute you actually, because there's a lot of pressure to just sell as many licenses as you can because there's this immense investment that organizations have made and are not earning back right now. Like they're not profitable business right now. They're in an investment stage. There is this pressure to go out there and maybe even find the use cases because we're in this like spot where we haven't really started with the use cases. We were like, this has potential to revolutionize and change everything. Great. Let's go find some nails a hammer. And I think that's like, it's an interesting lens to put on it. That's actually not what I thought you were gonna say, but it's this top down pressure that trickles down into, okay, yes. I also now feel competitive pressure to be buying AI enhanced project management software. Because that's like buying a phone that doesn't have a fax machine in 1992. You know what I mean? It's like, no, of course, you know, I need my multifunction state of the art thing, whether or not I send faxes, because that's what everyone else has got. But I do like that idea that now we're in that spot where it's like, this is when we're figuring it out. And it's okay if you've gone out and bought AI enhanced, you know, software of any kind, because that'll help us figure out what we can do with it. But it does take that mindset and not that like. Oh, magic button. We hit install and we're good now. Yeah.

Olivia Montgomery:

Yeah.

Galen Low:

We figured it out.

Olivia Montgomery:

Exactly. We're starting to see that too. I was very proud in the survey we had the results that security came as like the top priority for buyers for PM software and that's super exciting to me. It usually is functionality, which of course makes sense. So to see security bump up for the first time that we've seen as the top priority is fantastic. That's telling me that teams are realizing like, okay, this. AI of any type that you're using is increasing the surface attack level that you have, especially depending on if you're using external LLMs. It's a, there's a big difference between AI features that are, you know, in your current system that they just rolled out and they're like, Hey, you know, we're like helping you with helpful suggestions or some extra auto complete, or will help you build a workflow a bit better. That's a little less risky than your employees using external LLMs too. Synthesize project meeting notes and all of that. So I'm definitely excited and proud of our community for putting security at the top of the list.

Galen Low:

That's actually a really interesting one. It also speaks a bit, I think, at least from my perspective, I think it speaks to like a the speed to maturity, right? Like I think there's a lot of things that have disrupted us in the digital space and in technology over the past couple of decades. We've always been really slow to kind of like figure out the important stuff and to like hear that. I'm like, okay, well yeah, privacy data security, that's getting taken seriously now and it's still pretty early days in the grand scheme of things. Like that's important. It also speaks to how they're using it because to your point, I was never really that fussed about data privacy around like autocorrect. I was like, that's probably fine. You now know that whatever I cannot spell or type on a touchscreen for the lives of me. Right? And then that's fine. But when I'm punching in, you know, all of my project Adam, my client information, like business reports, it becomes a lot more important. And what I like about it is it's almost like a, I wanna say speed dampener, but like, it's almost like now's the right time before we get too far too fast. Shooting from the hip and like throwing data everywhere and then going, whoops. Whereas now it's yeah, it's important. Also, maybe it speaks to that in business competition, right? It's like, okay, I don't wanna like share all my information. Like, you know, we hear about building proprietary LLMs and stuff. I'm sure we'll get into it, but that's it's really interesting that end privacy showed up as that much of a priority.

Olivia Montgomery:

I know it's great. So proud of us.

Galen Low:

For me, it speaks to like this foundational stuff. And I wondered if we can zoom out a bit because. Not every organization is ready to just wake up one day and go, Hey, let's buy some project management software that has like all the AI features we want. Sometimes things need to be like in place first. And I thought maybe I'd ask you what needs to be in place before a team or an organization can actually benefit from AI features in project management software. And what in your mind are like the foundational pieces that help teams and organizations accelerate towards the ROI of AI?

Olivia Montgomery:

Absolutely. There's taking a few moments, any moments to pause and be like, okay, are we ready? Are we ready to do this? Yes, I'm ready to buy. Yes, I'm ready for the business to inject AI and see all the gains that we can get. But when are we actually ready? And there's two pretty big considerations that I think companies of all types and all sizes, all industries, all levels of maturity should be taking into consideration. And that's your technical readiness and your cultural readiness. So Debbie, in first, like the technical readiness, do you have enough clean structured data? The AI is, whether you're using machine learning for predictive analytics or even your LLMs to help you generate, you know, reports. And synthesize data. Those need vast amounts of really high quality data to be able to give you an output that you're gonna be happy with. So you need to double check, do we have this data? Where is this data? What's the quality of it? Are we ready to share that data? Is it good enough to get us. Insights. You wanna audit your existing tools before you unleash an AI because it could be that you are unleashing redundant capabilities. Another tool could already have it. And you also, when you do a tools audit, you could also flush out any shadow it that might be happening if you have employees that have. Rushed ahead, and maybe they're using tools that you're unaware of while they're at work for your business. You wanna know about that. And so doing a full audit sweep of your entire, I advise your entire IT r infrastructure, your entire environment first to make sure like, okay, these are all the apps that are in use, this is where they're in use. And make sure that is you. You have clarity of what your ecosystem looks like. You also wanna have your governance in place. Before you give these people your people tools, you wanna make sure that they know what they're getting, why they're getting it, and what to do with it and what not to do with it. Really need to have clear policies of the types of information that can be shared, how to share information safely with the tools. Or ideally you can set that all up in the backend and you're like, okay, if you're in these two, you know, this is our knowledge management, this is our document management, this is our task management. They're all safe and secure. Have at it guys. Go for it. That's ideal case, but a lot of companies aren't quite there, so at least empower and give guidance and direction for your employees with policies. So the cultural readiness, so again, your IT team and everybody can be ready, everything's safe, it's secure, ready to go. UAT tested passed successfully. Now you have to be ready, like can people actually use it when they show up to work? Do they know what to do? Are they comfortable using it? Do they want to use it? So that cultural readiness is the other branch that really needs to be evaluated and taken into consideration. So you want to be clear with the employees, your teams, about exactly what you're giving them and why and how they're expected to use it. There is room to be like, Hey, we're gonna open up some new toys for you guys. Maybe we'll run a pilot project and we can test these out, but maybe we're not rolling it out fully 'cause we don't. All of this is ambiguous. We're moving very quickly. So you don't want to have the expectation either, that you're gonna have everything perfectly mapped out. You can give these workflows and policies and teams just execute. There's gonna be discussion, there's gonna be like stops and goes and reworking. As long as you know that's gonna happen and that you can kind of isolate it within a testing. Period. A testing project, you're gonna be much more successful than if you just unleash tools. Your teams are probably gonna react with surprise and frustration, and you're not gonna see the productivity gains that you're probably hoping for with these tools. So you definitely want to avoid having any kind of mandates, no, like AI or bust mandates. And I think all of that can be achieved if you're just thoughtful, intentional, and like I said, roll it out with a testing. We're gonna try it, you know, in this. Isolated lower risk environment or project, however it is for your business first. So I think it's really important that companies not only take the technical readiness into consideration, but the cultural readiness too.

Galen Low:

I really like that you included shadow IT in your software assessment, which when we were prepping for this, we were talking about, you know, the differences between sort of a smaller, medium sized business versus like an enterprise and. We know that, you know, there are many large enterprises that should have really good governance, but they're unaware that people are just, you know, they're accidentally using their personal ChatGPT account. They are just playing with tools that are accessible on the web. Haven't been locked down yet by it, but I like that approach of like, Hey, listen, we're gonna do an evaluation, like you're not gonna get punished. We just need to know what's going on. We know it's happening and we want to build a plan based on the fact that we know it's happening. And actually that's great. That's great that everyone's sort of experimenting that they're running ahead. But yes, let's like put some guardrails on it. Let's not stifle that, but let's say, yeah, we might need to have some parameters, might need to be a limited pilot. Maybe we just don't try and like make organization wide change for, you know. Thousands and thousands of employees, but at the same time, let's not have like these little grains of sand of everyone doing a different thing. And then we'll never sort of benefit from it as a whole. And I think it ties right into that cultural readiness, right? Which is like, listen, this is about setting up some expectations both ways. The thing that you said that resonated with me is like, do people know what to do with it when they come to work in the morning? And I've been seeing a lot of organizations, they're making assumptions. You know, there's a lot of judgment around folks who aren't picking up the tools and you know, I see a lot of adoption charts that people frown at, right? They're like, we announced a thing, but adoption's still flat. Like, why don't these people get it? But I think that cultural readiness and the change management aspect, the education, and just the conversation goes hand in hand with that technical readiness of is this actually going to, are people gonna use it?

Olivia Montgomery:

Because that's exactly, you don't wanna stifle your employees, you want them to feel empowered, you want them problem solving. And so yeah, your business culture should kind of already be established of whether you are let's turn things on and we maybe, hopefully identify some power users, some people who, you know, you think will be very successful and are willing and excited to try new tools. And you let them play with them and kind of figure some stuff out while you formalize your policies and your plans, and then you roll it out to the other teams that maybe move a little slower, maybe more, a little more resistant, and you kind of try and bring everybody together through an excited curiosity and trying to move the business forward. I think your point of not taking a punitive approach or stance of any type is really key, because as soon as employees feel that there's gonna be. Something like negative stigma to them. Maybe like, oh, they're not open to new ideas. Any whiff of punitive culture is going to make employees shut down. That's where employees just hide stuff. They just don't bring up issues and you're really gonna suffer and stall out. And we are seeing, the survey does show that like adoption is. The top struggle that we're seeing, and it is often because it's a lack of clarity and a lack of a two-way conversation, a the five-way conversation. Everybody should be sharing ideas, sharing experiences. This stuff is very new and how every company is going to use it is different. One especially different and unique thing for AI is that it is impacting all departments. All industries kind of at the same time. So it doesn't really, you know, sometimes you're like, oh, well if you were an executive assistant, maybe you're gonna get really hit with an LLM. You know, the capabilities that you know, it's gonna make your meeting notes way better. Like great. But it's not just them, it's everybody. It's also your project managers, it's your engineer. Everybody now has these capabilities that they can use to augment their work and improve their work. And that's something we haven't seen quite before, other than just like moving everything digital onto a computer. And that kind of happened to everybody too. So here again, your entire company is getting exposed and seeing AI in very different ways and any kind of clarity that you can give people of here's how to safely and effectively use these tools. No punitive approach is gonna come to you. Really key.

Galen Low:

I like that. The fact that it's so new that it has to be a dialogue, and that's kind of the whole point of that sort of cultural readiness. And also because I opened this whole podcast up with like the idea that maybe these people are buying software and don't know what they want from it, and maybe it's because they are either. Yeah, like falling into the hype of it all and trying to, you know, stay competitive. And sometimes it's because they feel like they have to tell people exactly what to do with it. Whereas I think what you're saying is. Give them some card rails, but tell them that it's okay to like experiment and share. We're kind of figuring this out and it's not necessarily like a mandate with no strategy. The strategy is let's you know step by step, let's go through it. Like every department, every everyone in the organization, it would be unrealistic to just overnight transform your entire business, every single aspect of it, just because of AI. It's really that, you know, having that dialogue with everybody and setting expectations.

Olivia Montgomery:

Exactly, yeah. And the more that your technical readiness is tightened up, the more that you can, you have a really great relationship specifically for project management software, whether you have an existing tool that is rolling out AI features or you're looking for a new tool because they offer more AI features. If you have your technical readiness and you're like, okay, we've worked with the vendor, we've worked with RIT team, we know the data management is where we want it to be. We understand, you know, these are, this is gonna be a safe, it's not really a sandbox, but it is kind of a sandbox. It's your production environment, but you know that it's safe and protected. And so you can then turn it on and let teams be like, okay, maybe by department, by project manager, by portfolio depending on how big or small your company is, you can be like, Hey, you know, we've turned everything on. It's all safe. You can't break anything. Roll it out with your teams as your teams can. You know, you know your teams, you know your projects. Everything on the technical side is safe. Now you guys kind of help your team with the cultural readiness and that can help a lot with adoption.

Galen Low:

I really like that. I wondered if I could pick out something you said in there.'cause we've been talking about data security and governance. We've been talking about sort of like different sizes of businesses and different needs. How does the data security and governance piece vary for different sizes of organizations and do smaller and more nimble organizations like have a leg up over cautious enterprises or maybe not?

Olivia Montgomery:

So it's my favorite response is a mixed bag. There's pluses and minuses and strengths and weaknesses for both sides of the spectrum. So your larger organizations are gonna have strengths that there's a lot more data and a lot more historical data. The data is probably maintained, hopefully a bit more. Your data hygiene policies are probably better, and you probably have a much better idea of what your ecosystem looks like and how often your teams use tools and how they use tools. So you probably have a lot more of that foundational basis that can set up AI to be a bit more effective and successful for you when you're ready to adopt. The other side of that is that smaller companies, while they might not have as much data, and their data management policies aren't as mature, they can usually switch tools a lot faster. And so they can switch their project management tool much faster than, you know, say an enterprise banking can. So you can try out new stuff, you can see what you want. You can quickly leverage the latest technologies while, you know, bringing up the other, the technical readiness and the data quality behind it. So there's definitely plus and minuses to both, and I think having a realistic understanding of those pros and cons and how to maneuver those can be the successful path forward for both sides. So it's not that only AI is great for enterprises or only for small businesses. It's definitely very helpful and useful across the board. But what it looks like and the challenges that you're going to experience are gonna be pretty different because of the size and maturity that's out there.

Galen Low:

Do you have any stories or examples from folks that you've talked to while putting together your report or elsewhere sort of in the industry?

Olivia Montgomery:

In the industry. You want my industry secrets?

Galen Low:

Can you dish spill the tea? Come on.

Olivia Montgomery:

Air by air will definitely share that. The companies that, whether they're big, whether they're small, whatever size, whatever industry, wherever they are culturally, that going back to making sure that you're kind of ready and that you understand that your teams want to do well for you, and so the more clear you can be with your expectations and providing them the guardrail. The more successful you're going to be. The companies that are just saying, do AI, we're checking daily that you logged in. We're checking usage where your managers are getting reports. That is absolutely happening. I have friends, I have family. I hear like they're getting these mandates. These mandates are out there. There are leaders that are unfortunately. Taking a more punitive, like, Hey, we just bought this tool for a lot of money and every day we need 95% of everybody logged in for at least whatever, 10 minutes a day. They all have their own individual quotas and they're really sticking to those like kind of trying to measure ROI in these. KPIs like usage, and that's just, we're not ready for that yet. I don't know any company that's ready for that yet. We're still at the like, Hey, let's make it safe and help you figure out how to be effective with it. Then we'll go into the like auditing and making sure that you're logging in. I think it's a good idea to do audits. I'm not saying don't audit usage, because that's one way that you can kind of find out like, oh, this whole team hasn't logged in. Let me check in with them and see maybe why. All this team logged in and they were using it every single day for a month, and now we're definitely, the usage is dropping off. Let's check in and see why. Whatever it is you need to check in and why so. I wouldn't use those audits for usage as you're like, this is how we're gauging RO. I use it as how effective is our adoption?

Galen Low:

That's a good, clear distinction, right? Between using data to inform decisions and to guide folks versus using it to kind of punish and police. What comes to mind is like time tracking data comes up all the time in my community, especially for the folks working in agencies and consulting firms and professional services where part of it is like didn't log enough hours and you know, utilization isn't high enough. Are you even valuable as a human versus going, like, how are we spending our time and is that the right place to spend our time? And if we're not spending our time doing the important things, then what can be done? What is the reason and what can be done? It's just so funny and I'm thinking of like return to office mandates and things like that too. Like in some ways humans and organizations are very bad at like sort of measuring impact and bringing people along. There's always that sort of like quota, you know, like you get the stick if you don't do this or that. These behaviors. Even when the goal is like actually quite noble, right? It's like actually we want people to be able to like explore and use these tools and yet it comes across as, yeah, we're gonna punish you if you didn't log in today.

Olivia Montgomery:

Absolutely.

Galen Low:

What you said about that cultural readiness is like, you know, you can see how important it is and you can see how important the data you know, and the governance side of things is, and how they play together because these things need to be in place. Before things can actually move forward.

Olivia Montgomery:

Yeah, and I think to your point about time tracking, that is, we could have an entire episode on that, especially being a knowledge worker myself. How do you possibly track, I don't know, not to jump into like a neuroscience branch, but so much work, valuable work is done on walks. You have the type of work where, yes, I'm sitting and I'm typing and this is work. But the more valuable work, and I'm, for me as, especially as a knowledge worker and project managers everywhere, anybody who solves problems at work, the more that I can do the like, oh, I'm on a walk and I'm like kind of gelling the information that I have. And then that's when I have my aha moment, like, oh, okay, that's great. And then when I do sit down and I am typing and doing more traditional work, it's much faster and it's much higher quality. We tend to only capture or want to capture the time that is in front of the computer typing. And then how do we capture that, the really, truly useful time that was spent elsewhere. And I think that, yeah, RTO mandates and there's a lot of things that sometimes companies can. Be a little, maybe old school or rigid in their thinking and can limit that, and then they're not capturing. Then they're like, oh, you're not working enough. And you're like, what? I work all the time in my head. I never stop working. Right? Like that should count for something. But work you see that you like meat produce is because of all this other time that's like this, like mushy time. And I really hope that's something that AI could eventually help us with in the project management tools. There reminds me of one of my more favorite or exciting aspects that I'm seeing AI impact project management tools specifically is resource allocation. So there's been AI automation more. There's definitely, that's another, so a lot of AI powered tools are kind of more advanced automation, a lot of if then statements that we've had sent in CRM since like the nineties. So a lot of that's not necessarily new, but it's getting a lot more like turbocharged lately, which is fantastic. And so things like if you're making your project plan and you have your bucket of resources. And it's one of our main jobs as a project manager, is to figure out the right recipe for the project to work. Who's got PTO coming up? Who's got the skillset that I need? Who do I want to upskill? And I don't mind them maybe taking a little longer with the tasks because they're gonna be new, but I wanna upskill them. I need stuff done really fast. So who do I trust that can do very experienced and get it done fast? There's all that. You gotta put your recipe together. And we're seeing AI and machine learning and predictive capabilities come much better with that. We're also seeing there's some tools out there that are taking in kind of that mushy time. So taking into account the week after some, before somebody goes on vacation, maybe give them not so high cognitively, you know, required tasks. Maybe load them up with a task that are much more daily execution focused, and then they can go on their vacation and then when they come back, they're going to be pretty fresh. So you can give them those higher risk stuff that takes a little more like deep work. It takes kind of all of that stuff into account, and also some of them are getting really advanced. Where it's like, all right, this person works really well, you know, for like three weeks. They're gonna have deep individual work on my project, and then we're gonna give them a break. We're gonna give them a week of higher level tasks that aren't quite the same. Maybe they're more collaborative, whatever the case may be, of the nature of your project. But taking that human side into account, which is exceptionally complicated and difficult for a project manager. To make that perfect recipe with all those pieces. And the AI is definitely helping and we're seeing more and more of that. And I'm really excited for all of us because so many project tasks and deadlines struggle because we don't really know how to take that stuff into account. And sometimes there's not always like the safety to acknowledge that we're humans, we're not machines. Machine can do the work that you know, you tell it to and he might fry the motherboard, but you're fine. You just buy another motherboard. It's fine. We don't wanna do that with people. And so to have the most realistic expectations given to your stakeholders, given to your business owner, you need the most realistic timeline and work breakdown structure. And all of that needs to be as reflective as the capabilities and reality as possible. And I am definitely excited that AI could help us and is hopefully helping us get there.

Galen Low:

I love that. I love resource allocation as an example because it is very squarely human. And it shines a light on exactly what you're saying, right? Like humans are not machines. We were almost like. Hopefully all right. At this apex of like more traditional industrialization where we kind of did make humans into machines and treated humans like machines because we needed that uniformity, right? Did that person punch in at 9:00 AM and did they

punch out at 5:

00 PM That's how I can measure, because that data I can understand. Everyone's working style and their mushy time and all that stuff, and like, I don't have enough capacity to deal with that. Let's not even think about that. So we've I was gonna say that AI has, there's an opportunity to challenge our assumptions about work, but even more than that, I think it gives us an opportunity to challenge like what fallacies we've actually created around work. Not just assumptions, but we've been like. That's too hard. It's too difficult. Let's just make everyone show up at the office at a certain time and clock their hours. You know, we will do annual performance reviews and we'll ask him. We'll ask him once a year how they're doing, and you know, what could be better. That's all we have capacity to do. That resource allocation use case is just like, such a good example of like how machines frankly can help us recognize our own humanity as we do our work, especially, you know, knowledge workers and, but I just like, I'm so excited about that because. It does, it plugs all of this together, right? Where the sort of like the data piece, the cultural awareness, the sort of use cases and like, you know, what exactly does this look like? How, what am I supposed to do with these tools day in and day out when I show up at work? But also like how is it actually transforming the work for us? Not just the task, but like the notion of work and like how we attack it. That's a really good one. Yeah. I wonder if it leans us directly into the future because as we prepared for this. We were talking a little bit about the sort of like leveling of the playing field, and you mentioned it earlier, right, that like it's changing and impacting everyone in every department within an organization. And one of the things that you and I talked about is that sort of, that like AI becomes like bit of an equalizer, something that kind of makes technology and or business like a little bit more inclusive and a little bit more accessible. What are maybe some of the near term and longer term impacts that AI features, like let's say integrated vibe coding or like automation orent workflows, what sort of impacts will that have on the way work gets done?

Olivia Montgomery:

Super, super exciting. Let me go in the near term first. So near term we kind of been seeing with the ChatGPT coming out and these LLMs coming out. The immediate thing is that project managers and non-techies can talk to computers in ways that computers understand better now. And it's so like this much more frictionless than it used to be. And that is so exciting. Like to me, I'm blown away that I can see, like my parents now can like talk to an LLM and get some, you know, work done that they've never been able to do with a computer before because they've fumbled around like with their smartphone and all kinds of things. But now they're like, these LMS can even tolerate incomplete sentences for prompts they can tolerate. Misspellings in your prompt, they can tolerate a lot of incorrectness. Where before, if you're having to write lines of code and you're having to do injection queries, you had to be precise. Exactly precise. Now there's a lot more leeway with the natural language processing NLP capabilities kind of coming together with these large language models that are trained on vast amounts of data that is normal. People speak and we've not really had that before. So this is the first time we have computers that they know how to build and develop a lot of things. Machine learning on their own, and now a human can talk to them a much easier. So yeah, a pm a non-technical PM can go into their project management tool today. There's a lot of tools that today and you can go and type in, like, Hey, take project A and give me a work breakdown structure that runs through the end of the year. Thanks, appreciate it. And it's like, okay, and it responds like a teammate and it's like, here you go. Here's your information. You probably have to check it. You know, we've all used these LLMs and they make a mistake, and then you're like, Hey, you made a mistake there. And it's like, oh, you're so right. Thanks for pointing that out. Like we've all been there. They have their own issues, but the fact that we can talk. To a computer and get really effective outputs. We can build workflows. It's taking that no code, low code trend that we've been seeing for a very long time and just like injecting it with ease that we've not ever seen before. And that is definitely exciting. And I would say that's the near term that I definitely wanna bring up. It brings up red flags that I'm seeing for this. Just what I'm saying that you, this. Colloquial talking to computers like a teammate. There's a lot of issues that I'm seeing that we might be relying on emergent capabilities of these tools, and we're not really, the marketing isn't clear that these are emergent capabilities. There's definitely a bit of like muddling of what these tools specifically. I'm gonna narrow in on LLMs because they are the biggest near term impact. We'll get to the other ones for the longer term in a minute, but the near term is these LLMs. We're all using them. Every project manager I know is using them, especially to summarize information and gather data points, analyze data points, and all of that is emergent capabilities. And that is not clear right now. The LLMs are intended to generate text or images in a human-like form. They're statistically predicting word order based on the vast amounts of information that it has. It doesn't know what summarize means. It doesn't even know what the words summarize means. You can really break it down like I'm no data scientists and I'm no. So don't come at me. Anybody who's out there like building their own LLMs. But depending on your token, like it can break down your word. Like we'll take disconnect. Disconnect. You are like, oh, I know what that means. I got it. That's an obvious one. There's a disconnect. There isn't a disconnect. I got it. An LLM doesn't know what disconnect is. It's very likely broken that down into three tokens, dis con and neck, and it's using those three tokens. So taking the first one, dis, it's using that for discount. So it's like it doesn't know what dis means in con, like it doesn't know it's using that in any other words that have con, it's that same token getting, if that makes sense. There's a huge disconnect between how even an LLM understands the word disconnect. That's where I see like my background in like English literature and language study language. That's where I see this come in of like, I see where it gets that all mixed up and that has compounding effects. When, let's say, okay, I've brought this up very theoretical. Let me ground this down, like into an example. So you're a project manager and you are getting ready to do a big status update with the business owner. So you need to gather all of the statuses from everywhere that they come at you from. So you're gonna pull them from your PM tool. You probably, I don't know if you're a fan of voice notes. I send a lot of voice notes. I like them. There are companies out there that are giving people tools to like text the voice, dictate updates, emails, whatnot. So let's say you've got a couple voice memos, the updates from your tool, and you're gonna put these all together into an LLM and be like, help me understand what's going on so that I can give the full status to the business owner that I'm meeting with. That all sounds great and we are being told that it's fantastic at that. By other people, and I am telling you it is not fantastic that there are a lot of really big issues that can happen. In general, the LLMs are transforming the information that you're giving, and they tend to take out a lot of emotional words. They tend to take out a lot of sense of urgency. They take out a lot of nuance. Sometimes they just take out entire sections of what you've asked it to look at. And so it will almost always soften the language, remove those emotional cues. So an example, let's say you've got your voice memo, is your boss saying that the new vendor contract that you're waiting on is stuck in legal? He is had trouble getting a hold of them. You can't move forward until a legal response. And he is not feeling good about it. That'll be the voice note and then your AI summary, and you'll feed that into the LLM, but then in the status report that it gives you, it'll say vendor contract is in final stages. And that's not incorrect. It is in the final stages, but it's a tension filled, frustrating, final stage. And you are gonna miss that. You might not even actually listen to the voice memo because you are giving the transcript to the LLM, and you're expecting it to manage that. So you might even miss all of that frustration. So then you are not gonna share that with the business owner, and you're not gonna follow up with your boss, and you're not gonna try and unblock, you're just gonna be like, oh, it's in the final stages. Cool. And that happens a lot. And so you really gotta watch out for that. It's going to always soften your language. It's going to hallucinate. Hallucination rates are still very high, even in proprietary systems that companies are doing a fantastic job monitoring and making sure that, you know, they're adjusting for model drift and all the technical capabilities, right? But PMs need to know that the odds of that happening are very high, and it's only gonna come back to them. So these tools don't actually know what summarize means. It's just trying to give you what it thinks it wants. So, yeah, we need to know that. Again, I can't like stress this enough. It's softening language. It's making it seem like it's misrepresenting a lot of stuff and that could bite you in the butt. I know it's biting some of us in the butt. I hear about it.

Galen Low:

I hadn't really thought about it too hard in terms of like removing the emotional content and what impact that might have. And I mean, you know, just to nerd out on language and linguistics a bit, there's a difference between being taught. Language and training on language data and these LLMs are so convincing, right? We were talking about leveling the playing field, right? Everyone's kinda like, oh, okay, so it's as smart as a human 'cause it responds like a human would. And yeah, it makes mistakes like a human would. But basically it's a human, but it's not. And you can see where the hallucinations. Come from token by token, right? Like process by process, easily, you know, sent down the wrong path based on maybe limited information, limited training, and just context, but then also that emotional content, right? We haven't tracked the nut on that and what you're saying. Is a lot of a project manager's role is picking up on nuance or funneling emotion, you know, redirecting emotion and sure, we do summarize right as in like maybe shorten some things down, but we also transform some of that emotion along the way to say, okay, well listen, I need to shield my team from a bit of this person's angst. But also they need to know that it's urgent and that there's tension and that's gonna help us sort of get to the goal. And it's just so interesting that yeah, you know, you're like, okay, project management software that I bought that has all the AI features in it, I guess I'll just dump all these files and you tell me what the status is. And that whole like emotional bit could be completely missing. That is terrifying. Fascinating. And my gosh, what an interesting challenge ahead of us, I guess. I like your point that this is emergent, right? It's not perfect yet. This isn't the end, it's the beginning. And we do need to kind of like factor that in to the way we work.

Olivia Montgomery:

Absolutely. I think one way you can kind of, maybe account or limit that specifically for project managers because like you said, we're held accountable. If you miss the nuance, if you miss the frustration, when you don't address it, you're held accountable for that. And so we all know like make sure you're actually checking all the inputs that you're doing, all of that. We know it. I don't think it's happening quite enough because there's so much marketing messaging coming at us that these tools are good at this. And there's not this clarity that these are emergent capabilities. If you stick to the generating aspect, you're probably gonna be a bit safer. Or if you're in a company that has a very advanced, dedicated team that has not only the LLM trained on a huge database and trained very effectively, but they also are putting in a lot of if then statements, there are definitely ways that you can improve the reliability. But there's also the fact that these are still black box systems, and even the people that are designing them have come out being like, yeah, we're not sure why it actually can summarize text. We're not sure why it can do that. And you're like, oh gosh, that's terrifying to me. I know it's not terrifying to everybody, and that's great, but that's scary to me. And so I think when you see features, you know, in your PM tool of like the smart suggestions, those aren't so scary. If you're gathering user requirements and you're like, okay, yeah, I'm in my work ticket, and it has a little prompt of like, hey, kind of maybe getting rid of that like blank page anxiety that we often can have. I think it's fantastic for that, like generate the first draft. That's kind of my thinking with the LLMs is I try to remember that they're generative AI, so I'm just gonna have it generate a first draft, whatever I'm working on, and they're very effective and pretty good for that. Anything beyond that capability do remember that those are emergent capabilities. They're not tested and proven. We don't really know even why they're working like that. And so, yeah, to be a little careful, but definitely don't be scared. Blank page anxiety is way worse than a hallucination of an AI. At least that gets the ball going. At least that helps you move forward. So definitely don't be scared of that, but I think people do need to be aware of, and I think it, like I said in my, I think I have a kind of unique perspective. With my language and linguistics background, and I use AI tools all day, every day, and we study and research this and I've worked on a lot of IT teams and I know kind of the technical side of it, so it's. Definitely kind of a unique perspective that I hope, and I'm, I hope people kind of don't be scared, but know that we don't fully know how these things are working. I would not rely on it for your full status report. I really liked your example of like. You can write out maybe your status report and be like, all right, help me tailor this to the business owner and then tailor this information to my team. Because yeah, both those people, both those groups need different information, different types of information, different aspects of the information, and it can help you with that. But don't rely on it fully and don't trust it fully. Absolutely do not trust it. The other thing to do is they're so convincing because they were designed to be convincing. If you showed up and you use a LLM and it, you know, had an attitude or you knew that it was wrong or it admitted, oh, I don't know, you're probably not gonna use it again. And that's the opposite of what is desired. So it sounds convincing on purpose, but know that's just kind of, it's like getting you to use it tactic not because it actually knows. And can do what you're asking it to do.

Galen Low:

You know, as we go through this, it's becoming obvious why adoption was one of the top challenges in your report. Because we're coming down to like these mandates of just do AI, there's pressure everywhere. People aren't sure you know what they're expected to do with it. When they're at work, that little voice in their head is going like, well, you probably shouldn't have to do that. Just dump all those files into the LLM and like, that's probably what people expect of you. We're not necessarily having that dialogue all the time. The education training ourselves about how the technology works and what we do and don't know about it, and how we guide how we use it today. I wondered if we could look a bit deeper into the future. We're talking about some near term stuff. We're talking about generative AI, but you know, I think still that looming elephant in the room over the past few months has been like agentic AI, agentic features are making their way into project management software. Where does that take us and like what does it mean for us in the future? Future us.

Olivia Montgomery:

Future us. Yeah, so I am definitely very excited about future us. Current dust, it's a little chaotic. Future us, I'm really excited for. I think agentic AI to kind of level set what that means. It means a lot of things to a lot of different people. So within your organization, your IT team hopefully is defining that for you and the vendor, if they're offering, if they're selling Agentic AI, that they're defining what that actually means for you in general. We're not quite there. So if you are getting marketed that, hey, we've got genic AI fully ready to go, I would be very critical of that offering and really dive into the technical aspects of that. We're not there yet, so this Agentic AI fully sits in the like longer, near term. It's not here yet anyway. Agentic AI is going to be where the system can kind of make decisions and execute tasks at a pretty complicated flow. Ideally, we would like to be able to be like, Hey, book me a family vacation to Greece in October, and then it can go, it can check your calendars, the calendars of everybody that you wanna invite and negotiate the best rates for a car rental. Find the best flights. Go and do all of that. Act as like your personal agent and go and make all those decisions for you, and then come back and be like, all right, here's your trip. It's done. And that is exciting and I hope we get there someday. And that is kind of the goal that we're looking at. I think things like the NLP, the natural language processing, this ability of non-techies to be able to talk to computers. Do no code or low code automations. Even now we can see that happening. We can build out our workflows now and that's going to hopefully cross systems. So right now, maybe your PM tool, you can build your workflow, but it doesn't carry across, likely, it doesn't carry across your CRM and across your email, your calendar, et cetera, et cetera. But it will continue to increase those capabilities and that's where we're gonna get closer to Agentic AI. So you're going to be able to say, Hey, like I said, that example book my Family Vacation Degrees in October. Now that prompt that I just said is also kind of what Vibe coding is and vibe coding is this like new-ish thing. Like if you're not a developer, you might not have heard this. If you are on a like really new team or a, I should say, a new team using these tools, you might be vibe, coding and not knowing, but anyway, vibe coding, just a level set is using plain language as your prompt, and then the computer is the one that's doing the actual building of what the logic needs to be. It's not vibe coding like, oh I'm gonna make a cool vibe. Or like, oh, I'm in a good mood. This is gonna be so fun. Or, I want the output to be, you know, really like hipster cool. It's not that kind of vibe. Vibe coding is definitely more just using your plain language, Hey, book me a trip. But then the amount of logic that is behind something so simple as that is insanely complex. Most everybody here has booked a family vacation. We know it is insane. And so to think that we can just like pass that off to somebody else, to a computer is exciting, but also, you know, it's gonna be fraught with issues and it's really complicated stuff. So we'll see that it's gonna have the same risks because again, if you're vibe coating, even now we're seeing this issue with vibe coating. The same words don't always mean the same to everybody. And so you can be like, oh, book me a family vacation. And even if you do specify the four people you want invited, the AI can't go and actually like confirm that those four people want to go, that they want to do that. There's so much connective tissue that gets lost and nuance. It gets lost. And emotions that get lost. It's gonna feel good that it performed the task and gave you the output at the immediacy. But when you go and you go on that family vacation, that AI booked for you, whew. You might have an interesting time. You might not know what's coming up.

Galen Low:

It's like getting in a Waymo today, right? It's like a bit of a gamble. It's like kind of proven, but it's like a bit of a gamble.

Olivia Montgomery:

Yep. Yeah, absolutely. And there's a lot of bias. You know, there's things that can be unexpected. Let's say I use my AI that kind of knows. It knows me, it knows my preferences, it knows who I am, and I am the one that does the family vacation. And I'm like, all right, book it. And then it's probably gonna send us, it's gonna like book a lot of like history museums. It's gonna book modern art museums. It's going to be a tinge to me. Like, Hey guys, we're gonna go to the train museum now. I'm the only one that wants to do those things. The family doesn't wanna do that, but the AI wants to make me happy. So there's so much of that goes in and we're just gonna have to keep a communication going. And again, it comes all back to the like questioning challenging, knowing these things aren't perfect, knowing that all the marketing messaging you're hearing, take it with a grain of salt, ask deeper questions, and really know that these tools. I think like said, like I said earlier, the shine is kind of starting to like come off and everybody's like, oh, it's great at telling my kid bedtime stories. But yeah, it kind of messed up my status report and it did make me look like I was out of touch with my project because I over relied and I some of my thinking or outsource some of. My problem solving to it, and now I'm being held accountable for that. So yeah, it's definitely needs to be a continuous challenging dialogue.

Galen Low:

It's a really interesting point about, you know, yeah, like the agentic stuff is, you know, we're not there yet. And I think, you know, a lot of folks might say actually we are, but I think everyone would agree it's early days, and I like that it's not necessarily like. The progression of the technology. Like I think the other thing you're saying is it's also how we train it and how we give it context to begin with. Because like that in itself is the art. And you know, right now it's hard for us to trust it, to make decisions about the meaning of a word, not because. It doesn't. Well, I mean, yes, it doesn't understand, but we also haven't necessarily trained it with all of the nuance that comes along with human communication. And to go from that decision to deciding how to negotiate rates on Travago is like, you know, there are leaps, but the potential is there. I think it's doing a great job of painting a picture of how work will be, you know, in five, 10 years. It's messy right now, but it is making space for, you know, some of the mushy ways that we work. And at the end of the day to bring it all back, right? There's pressure. There's pressure because there's economic investment happening, you know, at the infrastructure level, at the government level, and that's trickling down to pretty intense investments at a corporate level, at a team level. And that is, you know, the pressure is trickling down for the users to use it, to figure out how to use it to, you know, log in every day and try something and share your knowledge. But in order for all of that to work, it's gotta have the right. Guardrails. It's gotta have the data and privacy and security in place and it also needs, you know, the human parameters. Like, what is expected of us, what are we supposed to do with this, and how can we support one another as a dialogue to kind of come out the other side. That's probably bigger than project management software. We started project management software, but I actually loved where we went. It is a little bit of a microcosm for work, right? Projects and the software that we use. This has been incredible.

Olivia Montgomery:

I think project managers are one of those roles that attracts people who tend to think very dynamically. They like to think big picture and small picture and that they often are, have a lot of that connective tissue of like, yeah, I problem solved this personal issue this way, and I'm gonna come and take this into my work. PMs are exceptionally good at that and not a lot of roles in business work that well. So hopefully they also appreciate that. Yeah. Like we can talk about, you know, the linguistics and the macroeconomics, but that does all impact how you use these tools. Hopefully it deepens your understanding of what you're seeing in these tools. Like your day-to-day, you show up and you're like, why did it mess that up? Like, why did it totally skip what I asked it to do? Why did it do that? Hopefully conversations like this and information like this. Help everybody understand a little bit maybe why they're seeing that, and why they're right to question it and what to do about it.

Galen Low:

Probably call out that project managers are good at that. Amazing. Olivia, thanks so much for spending the time with me today. I had a blast. Always great having you on the show. I mentioned at the top that you've just published your report, Capterra's 2025 Project Management Software Trends Survey. Where can people go to find out about it?

Olivia Montgomery:

Absolutely. It is posted on capterra.com. You can check it out there. Also, you can follow me on LinkedIn, Olivia Montgomery. I try to post my research regularly, my thoughts and ideas, insights there regularly. So yeah, check it out.

Galen Low:

Awesome. Love that. I'll include the links in the show notes as well. And Olivia, thank you again. Always a pleasure.

Olivia Montgomery:

Thank you so much.

Galen Low:

That's it for today's episode of The Digital Project Manager podcast. If you enjoyed this conversation, make sure to subscribe wherever you're listening. And if you want even more tactical insights, case studies, and playbooks, head on over to thedigitalprojectmanager.com. Until next time, thanks for listening.