Episode Transcript
[00:00:22] Speaker A: Hello listeners, and thanks for tuning in to this episode of Metro G's metrocyte podcast. Beth I'm Beth Schultz. I'm the Vice President of Research and Principal Analyst at metrogy. With me today is Irmola Kukrija. Irmala is the Director of Product Management for smartsheet, which I'm sure many of you recognize as a top provider of collaborative work management software. Metrogy likes to think about collaborative work management in terms of enabling employees to accomplish and communicate with and collaborate on and manage their project and content related work within a single location. A connected workspace, if you will. AI functionality. Functionality plays an increasingly important role for facilitating and easing work within a connected workspace. And Metrogy's research in this area shows that about 62% of companies studied already are supporting or planning to support AI functionality for collaborative work management. Irmala is here today to talk with us about the pragmatic use of AI for collaborative work management. Irmala, it's nice to connect with you again and welcome to the podcast.
[00:01:29] Speaker B: Thanks for having me, Beth.
[00:01:31] Speaker A: Great. Okay, so let's start out with some basic background information just so listeners get to know a little bit about you.
So let's just start with when did you join smartsheet and what attracted you to the company?
[00:01:47] Speaker B: It's a good question. So I heard about Smartsheet a really long time ago. I used to work as a product manager in my previous job at that time and I was trying to put together this presentation for our, for our leadership team about the roadmap. And I was trying to look for a template that might fit the needs for that presentation. And as I did a search, I came across one of these smartsheet templates where you could track and manage your project and present different views for different audiences. And I remember feeling like, hey, this helped me fix my problem. And, and I will say all these years later, this is what continues to keep me energized working at smartsheet. Whether it's, you know, large organizations like Uber or Amazon managing their events on Smartsheet, or if it's a nonprofit like Operation Smile that provides care surgeries for children across the world, Smartsheet allows everyone to work better at scale.
[00:02:48] Speaker A: Okay, so you joined smartsheet when, again?
[00:02:52] Speaker B: Seven years ago now.
[00:02:53] Speaker A: Seven years ago, as in product management?
[00:02:56] Speaker B: I did.
[00:02:57] Speaker A: Okay. And so how has your role evolved since you first joined the company?
[00:03:02] Speaker B: I would say in many ways, like at the heart of it, my role continues to stay the same. I continue to listen to customers for their Feedback in terms of what they are trying to do with the platform. And then I take it back to our product team and our engineering team to see how we can best kind of meet the needs of these customers.
And I will say our customers continue to challenge us with, with what they're expecting from the platform and the technology you talked about. AI continues to evolve, which makes things that were really not possible or we had maybe an okay solution before, it's given us a fresh set of eyes to maybe look at those problems again.
[00:03:47] Speaker A: Okay, is there anything in particular that's keeping you most busy today?
[00:03:54] Speaker B: I will say keeping up with everything that is going on in the world of AI, trying to make sure we're staying on top of all of the innovation in this space so we're able to really tap into the latest technology, tap into and make sure we're solving problems for our customers with, with everything that's available out there.
[00:04:15] Speaker A: That is certainly a challenge across the industry and across many different industries, of course. Okay, so when smartsheet talks about the pragmatic use of AI for collaborative work management, when it does this, is it strictly speaking about the AI tools or is it really kind of more of philosophy, more philosophical than that?
[00:04:38] Speaker B: I think when we say our pragmatic use of AI, we're talking about our approach to implementing AI within the product and that is about having these very practical use cases that are very contextual to what people are trying to do to drive real outcomes. So the way AI is available in the smartsheet platform is very contextual to what you're trying to do. So as you're trying to build out a formula, the AI assistance is available to help you build it out.
It's, it's there to help you drive real impact, whether it's in the form of efficiency or in the form of making information available to you to drive decisions.
It's very practical trying to help you solve real world problems.
[00:05:28] Speaker A: You mentioned that building out formulas, using AI to help build out formulas, which is a practical pragmatic use of AI. Can you give a sense some other examples? And then maybe how have you seen or how do you think about, you know, kind of the non pragmatic use of AI?
[00:05:46] Speaker B: Yeah, so building out formulas is, is, you know, our, our feature that's really loved and used by lots of people. The other capabilities that are in the product that people find really valuable and helpful. Actually my favorite is what we call analyze data. And what that lets you do is we will have lots and lots and lots of content in Smartsheet and you need to get like insights from this data. But the kind of insights you can get will depend on, you know, the skills of the person if they had time to get these insights. And I feel like our engineering team has built out this very clever way of and we're the only ones in our category who'll allow visualization of this data that's available in a work management platform. So we're using AI where someone can come in and ask a question.
Sometimes you don't even know what question to ask and you could really just come and say what might be trends that might be interesting in this data. Tony, I will support you and help you come up with the kinds of questions to ask, the kind of insights to get, you know, do ad hoc analysis on that data and then if you like what you see, you can build it out as a dashboard and share it with your team from there on.
And I think you asked about maybe what might be non pragmatic and I think some examples there might be the more academic application of AI or the more academic use of AI where you know, there's lots of organizations and companies doing work, kind of pushing the technology to the boundaries, trying to, it's not necessarily ground in an application or you know, to solve a real world problem, it's to extend the use of the technology.
And I will say all of that is, is really important work because you don't know what the next breakthrough is going to be. We've seen self driving cars evolve from some of this research and development that has gone on. That's not necessarily where smartsheet is right now. We're building on the great work that has been done and finding practical applications of IT in work management.
[00:07:56] Speaker A: Okay, well certainly building out formulas and then being able to tap into that data and pull out insights and visualize insights so easily is got to help be a tremendous help for companies.
So Armila, what are some of the key AI technologies that you see driving innovation and trends that are shaping up around collaborative work management?
[00:08:24] Speaker B: The biggest one I'd start with is just natural language processing. I think we've all, all had our minds blown with like wait, I can just talk to my computer in, in a way that I would talk to another person. And so I think that's, that's definitely going to have, we're already seeing the impact of IT in collaborative work management in a way that you don't, you don't really need to know the details of how the technology is going to work. You don't you know, we talked about formulas, but you don't need to know how to build an automation necessarily. You don't really need to know. This is how you plug in the data and build this report and build a dashboard. You can just talk about your real world problem to your software like you would to your coworker or someone assisting you to do work. And it will really be there to support you and build it out.
There's advancements in machine learning that we're tapping into. We've had for a very long time suggestions in smartsheet where you can come into the platform and it will recommend what you should be working on, what's the most important work for you. We're building on that to kind of build out this knowledge graph for you. And because it's got all these connections of, you know, who might be the person with the right skills for the job, what is their availability, what is their pay rate, it can come in and really help you come up with recommendations for who might be the best person that you want to assign this job to.
And then I think there's advancements in deep learning. You know, we talk about smartsheet as the platform for work people, content. People upload a lot of content within smartsheet, whether it's in Smartsheet or in brand folder. And I think advancements of being able to really look through this content, whether it's in the form of documents or images, is really going to help people get deep insights from data. That's in some ways like it's been in these systems and now you can do deep searches through AI within them and get insights from that data.
[00:10:31] Speaker A: So natural language processing, machine learning, deep learning, none of these are new concepts, but they're all just coming really into practical use these days.
[00:10:41] Speaker B: Absolutely. And I will say the power of them coming together using AI on top of the knowledge graph we've had is just helping us solve problems that we didn't think were possible or they were really, really hard in the past and now they're not.
[00:10:57] Speaker A: You know, one term that we hear more and more often is agentic AI.
So is smartsheet talking about agentic AI at all? And if so, what role do you think it'll play in collaborative work management?
[00:11:16] Speaker B: Let me start by maybe saying what we think about when we're talking about agentic AI. I think we've seen lots of application in lots of systems where AI is your copilot, where it's assisting you. You know, you can either use it for early brainstorming to bounce ideas off of. Or you can be like, I wrote this draft, help me summarize it. But it's really there to like assist you and support you. And as this technology is getting better and better, I think we're starting to see that we're able to trusted to make autonomous decisions. And so you can have like this AI agent that's very, very skilled and very trained in a very specific task. And it's so good at it that you can almost delegate a that task to it. It can work in with autonomy, make decisions and move work forward for that particular area. So the way we see it in smartsheet, you know, we talk about automation in smartsheet a lot and you can imagine that you have a sentiment analyst agent in smartsheet. So a very common workflow people have in smartsheet already today is they'll put out forms to collect feedback. We actually use that for our internal feedback process in smartsheet too. And so we get, we get feedback or feature requests from customers through a form. And you can imagine the data from that form feeds into a sheet. And now you have an agent that's really good at analyzing the sentiment in this as it's coming. And then you have another agent that's really good at all right. I think that this kind of feedback probably belongs to, you know, the enterprise team to go resolve or it belongs to the dashboard team to go resolve. And so it's making this decision about categorizing what, what type of feedbacks come in. Then there's another agent that probably is really good at task assignment we talked about, because AI has access to information around whom has the skills for the job and who might have availability. There's another agent that's really good at assigning this task. You can see this entire workflow that used to take lots of steps was very manual. Now with an AI agent or actually multiple AI agents agents in here, you're, you're able to kind of really streamline the process that was, that, that was already even more powerful with automation and smartsheet. And now we'll see a way to, to really get dividends on our investments in, in automation over here.
[00:14:06] Speaker A: Okay.
It's going to be so powerful when it's like really fully engaged, I think. And, and, but you really have to be able to trust, right? You have to put a lot of trust, trust in your data, trust, trust in the AI, etc.
So just, you know, be what I'll be watching kind of closely in terms of when Companies sort of are willing, you know, really willing to sort of put together that string of, of AI assistance and really streamlining those work processes.
[00:14:42] Speaker B: I think that's true. I think there's, you know, we'll have to put thought into how there's a human in the loop with these agents working. And sometimes it could be, all right, I'm going to check this box and then this process can run through or I'm going to make sure that AI is going to do the work to categorize the sentiment and then I'm going to approve it for the next flow. So I think we'll have to still until we're able to build a trust and until the technology is able to, to do it in a way that we trust it. I think there's really clever ways within smartsheet automation that you can still have a human decision maker and then delegate the tasks that are low risk, low impact for AI to take on.
[00:15:26] Speaker A: Yeah, for sure. And making that interplay between human and virtual agent assistant rather just very seamless. So, Armila, how are teams benefiting today from AI enabled work management? You gave a few examples, do you have any others? And then how do you see this evolving as kind of use and available capabilities expand?
[00:15:48] Speaker B: Yeah.
So AI in Smart, you know, we talked about your ability to not need to be a power user anymore, that the tools available, that AI is going to help you use the tool to the fullest without really needing to know the details of how it works. So you can maybe focus on your, your business problem and AI can help you kind of build it out, build out your solution for you, track it.
The next kind of dimension is around removing the boring and the repetitive and the mundane work. I know, I told you I've worked in Smart Chain for a really long time. I haven't heard a single person, whether from our customers or on our team, say, yay, it's time to write that status report. So, so, you know, I can really help with this repetitive work. The information's already in smartsheet. You can, you can point it to it and it will help you kind of build out that, that report for you. The way it's, it's available today already in the platform is to give you assistance through help. You know, we talked about sentiment analysis, we talked about translations as a really good use case. So all of this like repetitive work that today requires a lot of copying and pasting and manual data, AI, we're seeing really good application of AI in it. And then the next dimension was around Getting insights from data. You know, you have all of this data, whether it's in smartsheet or other systems in your enterprise.
We're seeing really, really, really promising and good use cases for ad hoc analysis of this data using AI.
[00:17:34] Speaker A: So we talked about the benefits and the cool stuff coming down the pike. What are some of the biggest challenges and limitations of using AI in work management?
[00:17:45] Speaker B: The challenges and the limitations.
You know, I think when I, when I started learning computer science, when I was like, you know, the first time I was introduced to computers at the age of six, I feel like one of the first things we learned was like, it's a machine, it's garbage in, garbage out. And I feel that's still so true of AI where the kind of results it's going to give you are going to be limited by the data that's available to the platform. So if you're going to point to it two projects and one is going to say the start date is the 10th and the other one says the start date's the 22nd and there's really no indication of which was updated later or who updated. There's really no way for the system to come up with the right results. So I think it continues to be limited. One of the limitations is going to be data hygiene and thinking about how are we providing this system the right data. And I actually think the way we've approached our Q Business connector, if I may say so myself, I think it was very clever the way the team approached this. Where Amazon Q Business is, is like your enterprise search tool where you can go in and ask it questions and instead of saying, you know, here's my smartsheet, go search everything, index everything and come up with results. We gave power to the end users to say, what do you think are your workspaces that have the most important information, that have the information that is, you know, kept up to date. And I think we're seeing really, really good use cases and application from it. Our legal team uses it, it's got, it's got the latest non standard agreements attached to it, so they're able to come in and ask questions about those agreements. We track our roadmap in smartsheet so our product and engineering teams are able to go into Q Business and ask about when something's going to launch or what teams are working towards that launch. As long as you're able to point it to the data that is applicable, I think you can start to come over these limitations.
[00:19:56] Speaker A: That goes back to your point earlier about being Able to speak to the data, being able to speak to the software and the language that you use every day?
[00:20:05] Speaker B: Absolutely.
[00:20:06] Speaker A: Armila, do you have any best practices advice to share on implementing AI powered work management tools?
[00:20:16] Speaker B: I would say like you build out anything else, continue to start with the problem instead of being like here's a technology and let me see where I can apply it. Really start to think about what problems and what use cases you're trying to solve.
Whether it's about, you know, bringing efficiency, making decisions, what are the things that, that people are, are having a hard time collaborating or getting work done. So I'd say, I'd say continue to start with the problem and then something to be mindful of is it's, it's easy to see early success with these models.
It like I think we say like to get 80% of the way is probably going to take you 5% of the time and it's after that that the tweaking and making it really enterprise grade and the hard work starts. And, and so I'd say don't give up when you start to see the first kind of incorrect results.
Share it with trusted users, iterate on it, make sure they understand the state of the software. Collect a lot of feedback. Is the way I ask questions may be so different from the way you ask questions. So really enlist your, your user group in improving the product even from there.
[00:21:36] Speaker A: Give it time in other words. Right. I think we've heard that, we've heard that across different disciplines.
You know, sometimes those proof of concepts really need to extend out, if not by months, by given a year or more.
[00:21:51] Speaker B: Agree. And one of the things you mentioned really was around trust. I think people are still kind of building trust in AI technology. Since you're building this out if, if you can make sure you're being really transparent about where AI is being used and if you're providing a result kind of linking back to the source so people understand why AI came up with, with the results that came up was, will go a really long way in really allowing people to continue to experiment and try these tools.
[00:22:22] Speaker A: Absolutely agree. Okay, one last question for you and this is AI or not or related to AI or not? What are you most excited about for the year ahead in collaborative work management?
[00:22:38] Speaker B: Actually I think it is AI related and I do think, you know, smartsheets promise has always been around enabling the business user and empowering the business user and you don't really need very special skills to either learn to write code or really work with someone in it. To power your process. And I really continue to feel very excited about the impact that AI can have for these business users. Really, all they need to do is be curious about the problem they're solving and go in and give it a try and not be worried about breaking stuff. And so kind of the impact that we can see that AI may have on these business users who are trying to solve these problems that are either very hard or people have even told them are impossible in the past, I think, is we're going to start to see these users have have new tools and new capabilities to solve them.
[00:23:42] Speaker A: So lots to look forward to. And I do think that's a good place for us to leave off or Mila, I want to thank you again for joining us today and sharing all your insight with us. Really appreciate it. Thank you, Beth and listeners, till next time. Take care, everybody.
