DPM PodcastGeneralTime ManagementToolsTopics

DPM Podcast: Artificial Intelligence in Project Management (with Dennis Kayser)

By 11/09/2017 November 13th, 2018 No Comments

Want to get your work done faster?
Try Hive, the productivity platform for fast-moving teams.

Our friend and supporter

Audio Transcription

Ben Aston:

Thanks for tuning in. I’m Ben Aston and this is the Digital Project Manager podcast from thedigitalprojectmanager.com.

Today I’m joined by Dennis Kayser, and today we’re going to be discussing artificial intelligence in project management. Are our jobs as project managers going to be taken over by project management bots? We’ll find out, so keep listening.

Dennis, just first thought tell me a little bit about yourself. Where are you, and what do you do?

Dennis Kayser:

Hi. Thanks for having me, Ben. We’re a startup based out of Copenhagen, Denmark. We are a Venture funded company. We work on a bunch of different things. Just this morning we had a huge meetup with 80 people showing up just listening to kind of what we had to say about the whole future of kind of project management and AI and how that will kind of influence the kind of whole profession of project management as we see at least. That’s one of the things.

Ben Aston:

Cool. Do you help organize the PM meetup?

Dennis Kayser:

Yeah. I’m running a few meetups actually. We’re as a company running a few meetups just as a kind of giving back to the community type of thing so we can kind of share ideas between other PMs in other companies and share ideas, best practices and kind of where we see the whole profession is going.

Ben Aston:

Cool. I know that you obviously run Forecast It, but apart from Forecast It and the meetups, is there other stuff that you’re doing as well then?

Dennis Kayser:

This whole work thing takes up most of my time obviously. I try to do a lot of skiing whenever I can. I try to do a lot of running whenever I can and then I try to do a lot of cycle crossing whenever I can. Trying to keep active.

Ben Aston:

Nice. Good stuff. Tell us a bit about you’re now running Forecast It and we’ll go on to talk a bit about that first but how is it that you came about to be running this Venture funded startup to be creating project management or software? How did that happen?

Dennis Kayser:

I used to work for IBM, so I’ve been seeing a ton of big projects and smaller projects obviously both go well but also go very badly due to various reasons, typically not because the people that are working on these things are not doing a good job or not intelligent but basically because quickly these things, especially in digital, get very, very complex and that’s basically just what makes it very difficult to manage on a daily basis. Obviously learning from that and then looking at all the different products that are in the market and not really finding anything that was simple enough and that kind of made sense. Still kind of a lot of the same stuff going on in the market from our point of view. A lot of the same data entry heavy systems. We’re trying to kind of reduce that and kind of trying to take out all the boring stuff of being a PM so they can kind of focus on the more fun stuff of being creative and innovative and stuff like that.

There was a bunch of people when we started this that were saying it was a terrible idea but luckily we’ve proven them to be wrong about that and now are showing that we can actually make a big difference in the current space I think.

Ben Aston:

Tell me about the journey of the product thought. Did it start off … I mean you just said you’re beginning to prove the naysayers right but did it start off how you hoped it would? Has it all been successes from the start?

Dennis Kayser:

We started this company about three years ago building a product that was like way more enterprise like, way more similar to actually the other products in the market and we found that to be actually just a bad idea. Coming from a computer science background, both me and a few of my cofounders, we were thinking can we do this in a smarter and more intelligent way? Then kind of AI is right up the ally of this. There’s so many things that can be automated through AI in this space especially and we were just wondering why is nobody doing this?

We kind of restarted the whole company last year and built a brand new products. We raised some money from the Venture capitalists and now basically with clients globally, so pretty interesting.

Ben Aston:

Good stuff. Tell us then a bit about Forecast. Give us your elevator pitch for Forecast for those who haven’t checked it out yet or know what it’s about.

Dennis Kayser:

Sure. Forecast is an AI driven resource and project management solution. We try to keep things very, very simple. Not being too complex so people actually want to use it. We want to basically automate as much of the boring stuff as possible and then obviously let you integrate to whatever best practice or whatever favorite products you have. You get like a fully synchronized solution; this being like invoicing your Slack, your Google Drive, whatever you want to connect to, we make that very, very easy. You can get like a full perspective of what’s going on.

Ben Aston:

Cool. Let’s talk about AI for a second then. For people who don’t know what AI is, what is AI?

Dennis Kayser:

AI is a very broad topic. AI basically in the sense that we’re using it is basically learning from all the stuff that is going on. In this case, in project. Learning from if things usually take longer or shorter for a certain client, if your budgets are way off and we kind of realize that overtime that things they usually eat slower for instance. Then the AI kind of picks up on that and actually helps you improve. Whenever we kind of learn … that might be Ben might be slow at Java but he might be super fast at html or something like that. The AI would pick that up and then will actually help you prioritize the work that you need to do so you don’t have to kind of manually sit and shuffle all this stuff around.

Ben Aston:

Cool. The system is automatically or auto-magically increasing people’s allocations for certain things and in the background trying to do some kind of resource leveling thing to … out the … Is there a way that you’re setting priorities in terms of maximized revenue or maximize the amount of billable hours people are doing? How’s the system know what’s a priority?

Dennis Kayser:

First of all, actually we’re not doing anything automatically because that often confuses people more than it helps them. What we do is we actually help and recommend things that you could do to improve your plan. For instance, our kind of core audience for our product is typical an agency or like an advertising consulting company, the type of business where they want to kind of maximize the hours they have. They want to have high utilization but they also want to make sure that they’re actually doing the projects for the clients that are kind of the good clients. You’d actually be so surprised at how many web based agencies that are building all these cool projects for clients but that actually end up not making any money off them because they don’t have the control of what’s actually going on and what they’re delivering.

Ben Aston:

Cool. It’s not making decisions for you but it’s just making suggestions. Hey, if you’ve got Dan working on your html, he’ll probably be all right but as soon as you put me on something else, I might be a bit slower. Is it comparing estimates versus actuals on certain allocations, so you have to get quite granular in terms of the resource allocations and what people are exactly working on?

Dennis Kayser:

That was actually one thing that was important for us that you can actually use it at any level you want. You don’t have to use it very granular. You can if you want to but for a lot of companies they don’t want to kind of bother with all that kind of detail stuff. It learns from basically a lot of aspects. It will learn which clients are the ones that are the best for you, which people are most efficient at what stuff. How do we pack the schedule for people in an optimal way so they don’t get either too little to do but also don’t get too much to do because there’s actually nothing worse than just having … Everybody probably knows that type of project manager that just kind of overloads you with stuff and then you’re just super stressed for two weeks and then you have to go home sick for a week. That doesn’t really work either. It’s kind of all about balancing the whole thing.

Ben Aston:

Clever stuff. In the process of creating this artificial intelligence and creating I guess all these rules, were you basing it on your own internal projects or how did you kind of create these algorithms to decide what’s a smart thing to suggest and what’s not?

Dennis Kayser:

Basically what we did is we built this whole AI thing from the ground up. I have some very clever guys with some fancy degrees in both the machine learning AI and computer science. What we did is we actually got a whole bunch of project data from a lot of different companies and we kind of pooled that into a big database and then we trained the machine on that. What actually happens is that once you create a new account in Forecast, you basically get going very, very quickly because you’re starting off of baseline dataset of I think right now it’s about 5,000 or 6,000 projects and there’s like more than three million hours worth of stuff in there or something like that. It’s actually a pretty substantial dataset already.

As you start working in it, then it will kind of tweak itself to work or kind of figure out how you’re working. It actually will learn and pick up stuff from your company and that will kind of just be unique just for you. In that sense, the more you put into it basically, the better it gets.

Ben Aston:

Nice. Kind of where do you see this going then? Right now you’re using artificial intelligence to improve estimates, budgets, work plans, resourcing. What’s the end goal here? Is it to make project management redundant or how does project management fit into artificial intelligence? If a machine can make all the right decisions for us, where do you see the role of PM? Kind of what do you see it evolving into?

Dennis Kayser:

I definitely don’t think that the PM will ever not have a job. I see it more as an aid to actually take away all the boring stuff for a PM. Instead of having to spend the few hours every day or every week fiddling with a resource plan, why not have a system that actually helps you plan this so you can spend your time on actually maybe talking to the client or helping develop the project or whatever you kind of want to do that’s kind of more on the creative human side I guess? We basically want to take out all the boring stuff and make the work more fun.

If you’re looking into the future, I think you asked that, some of the things we’re working on right now actually is we’re working on basically what we call an AI assistance. Basically be an assistant that you can ask questions about your projects or ask it to do stuff for you and it will basically then do that. The same as having a real human project assistance but just like a super-powered one because it knows everything that’s going on and has kind of full visibility into everything; all your clients, all your projects, all your people.

Ben Aston:

That’s really cool. I saw something. I read an article just a couple of weeks ago about something similar because I think one of the interesting challenges for PMs is the fact that so much of our time is spent answering questions where people just don’t know where to find information. For example, the classic one, what job number is this? Or where should I be putting my hours? How much did we spend on that project last year where we did some banner campaign for whoever? Being able to have some kind of …

Dennis Kayser:

I totally agree. Just imagine just being able to ask an assistance saying currently like today what are my five most high risk projects? Which ones should I be worried about? That would just be like an immense help in your daily work and you don’t have to kind of sit and shuffle through all these lists of different stuff. You basically just ask AI and they will tell you exactly what you need to know in an instant.

Ben Aston:

How are you working on that as an integration through Slack? I think the one I saw was like an integration through Slack so you could ask questions through Slack. Are you building it out within in the system to be able to quire other tools as well?

Dennis Kayser:

We’re actually building it out into other tools. You could use Slack if you wanted to. That’s not a problem. You can just connect it to Slack and you can use that as the interface if that’s what you want to do. Also, we’re actually providing our own interface within the application so you can ask that as well. Basically, what the Slack bot does is it actually just quires our service. In the end, everything is known because we’re connecting your systems. That’s why you’re able to ask a lot of advanced questions that a lot of other … Normally, you wouldn’t be able to do this unless you had like a person that actually knew this stuff.

Ben Aston:

I think improving the lives of the resource manager who’s often the one at the brunt of all the stupid questions that people ask or as a PM, if you keep on shifting your resourcing around and actually it doesn’t need someone to do that when it could be done through just artificial intelligence and just being able to tell the system hey, my project’s been delayed by a week. Shift out all my resources by a week and let’s start it next Monday rather than this Monday. It shouldn’t need to go through your resource manager for that to happen. So much of this is we have so many manual processes and even things like opening a new job number or creating a weekly status report. If you could just write a command and get that generated for you, that would be awesome.

Dennis Kayser:

Yeah, I totally agree and that’s what we want to do. We want to automate the stuff that’s boring and then you can focus on the fun stuff. I just had a client meeting yesterday with one of our clients and she was like wow, you guys, you’re saving me like two or three hours a day of just like sitting and shifting resources around. Just that proves the case that it’s nice that you can help people in that way so she could spend the two hours doing something more fun than sitting and kind of dragging and dropping people on a timeline.

Ben Aston:

That’s great. Have you had any other kind of interesting feedback of how Forecast It has helped people in ways that you weren’t expecting?

Dennis Kayser:

Actually, this is going to sound a little boring but actually not that unexpected because we focus a lot on talking to our clients. We have very deep interaction with all our clients to make sure that we’re actually building stuff that helps them on a daily basis. We take that feedback very seriously. Our UXer, she usually has around five to 10 sessions with clients a week; one hour sessions where she just kind of gathers feedback on how is it to use and how nice it is and what doesn’t work and how should we fix it?

Ben Aston:

Cool. Apart from the kind of PM assistant, what else is in store for Forecast It? Where do you see it in five years time? You’ve built out your PM bot system, what’s next? Where do you think you’ll be in five years time?

Dennis Kayser:

Actually, it’s pretty important that we don’t kind of branch out too broadly. We don’t want to be like a new big enterprise system that’s very difficult to use. It’s very important for us to kind of keep things simple and easy to get going for new people. Also, I think sometimes you end up in this feature and we are very aware of not doing that because we can’t suit everybody and we want to make sure that we actually suit the people that is kind of our ideal. I think in the product side, it’s definitely making the AI really, really amazing. It’s not easy to do. It’s actually a pretty big task. Then basically connecting to as many systems as possible so you can use whatever you like without any problems and then basically giving you more insights that are AI driven so you can better understand kind of how your business and how you can improve your business in a more automated way.

Ben Aston:

Cool. Apart from Forecast It, I’ll be interested to understand from you what else is in your tool kit as someone who’s developing products? What are you using to manage your product development and the projects within that?

Dennis Kayser:

We’re definitely “dog fooding” because otherwise it wouldn’t make much sense. It’s a little bit made up but we’re using our own product to develop our product right now and that’s very important for us also because if you’re building a product and you’re not using it yourself, something is wrong. You should hopefully build a product that you’d want to use yourself. On top of that, we’re using all the Google stuff. We’re using Slack a lot. That’s really big for us. We do a bunch of stuff with slack. What else are we using? I think the stack is not that deep actually on the PM side. We’re trying to keep it to as few tools as possible.

Ben Aston:

You’re not using Gera or Trello or any for task management or your kind of …?

Dennis Kayser:

No, we’re using our own system to do that. We’re fully capable. Actually, we just won a client over from Gera last week. The thing with complicated products is that you will end up with a lot of people that never use it because it’s just difficult and annoying to use. We’re basically trying to not do that.

Ben Aston:

Yeah. I think the big challenge with … One of the things you’ve talked about that is enabling people to do their best work and not stifling creativity. Whilst Gera might be a great … Gera is kind of the defacto project management standard I’d say now because it is very robust and for development you can get really granular. You get lots of control but whilst developers might like it, everyone else hates it.

Dennis Kayser:

Exactly.

Ben Aston:

There’s this balance between do you use Trello up to development and then when do you start using Gera but it’s having a system that everybody enjoys and can agree that does everything that you need it to do.

Dennis Kayser:

That’s our aim.

Ben Aston:

Yeah. I can switch between Kanban and Scrum-

Dennis Kayser:

Show the client the Gant Chart is that’s what you need and all that stuff but without it being too difficult. That’s actually funny because I had a talk with a friend the other day and he said now his wife got a job as a digital PM and he was now learning Gera. It was really not fun and it required a lot of gin. That was kind of his take on that.

Ben Aston:

Yeah. It can be a bit complicated. Cool. Thanks so much for joining us, Dennis. It’s been great having you with us.

Dennis Kayser:

Thank you. Thanks for having me. Always fun to talk about stuff that we’re passionate about.

Ben Aston:

Great stuff. If you’d like to contribute to the conversation about artificial intelligent project management, check out the article on thedigitalprojectmanager.com and join us on Slack as well. We’ve got a community of digital project managers discussing all kinds of interesting things about project management. Head to the community section of thedigitalprojectmanager.com and join the conversation there. Until next time, thanks for listening.

Ben Aston

Ben Aston

I’m Ben Aston, a digital project manager. I've been in the industry for more than 10 years working in the UK at London’s top digital agencies including Dare, Wunderman, Lowe and DDB. I’ve delivered everything from video virals to CMS’, flash games to banner ads and eCRM to eCommerce sites. I’ve been fortunate enough to work across a wide range of great clients; automotive brands including Land Rover, Volkswagen and Honda; Utility brands including BT, British Gas and Exxon, FMCG brands such as Unilever, and consumer electronics brands including Sony.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.