Escaping AI Pilot Purgatory with VentureFuel's Pritam Bhattarai

Fred Schonenberg
Hello everyone, and welcome to the VentureFuel Visionaries. I am your host, Fred Schonenberg, and I am so excited today to welcome Pritam Bhattarai to the show. He is my colleague. He is the Vice President of Innovation and Partnerships at VentureFuel. He spends his day helping Fortune 500 companies identify emerging technologies, partner with startups, build innovation programs, all with the idea of commercializing innovation. 

He recently was a speaker at the Innov8rs Conference in Toronto, which was a gathering of some of the leading corporate innovation leaders for some of the largest companies in the world. His subject for that talk was AI Pilot Purgatory: Why Most Experiments Fail to Scale and How To Fix It. So we wanted to unpack some of the lessons from that presentation, as well as to share a little bit of his insights. He moderated a number of panels while there.

We hosted an innovators dinner, which we do all around the world multiple times a year where we have VCs, corporate leaders, and startups get together to kind of talk about where the puck is going and how to really commercialize innovation to kind of take it from idea to impact. So really excited to unpack the themes and learnings from the presentation, as well as to dive into where things are headed. Pritam, welcome to the show.

Pritam Bhattarai
Thank you for having me, Fred. Excited.

Fred Schonenberg
So it's very nice. You and I are on multiple calls a day, so it's fun to throw a microphone on with it. But maybe for people that don't know you, can you tell us a little bit more about your role at VentureFuel and what you're focused on?

Pritam Bhattarai
Absolutely. So as you recapped for the audience that we help large organizations commercialize innovation. And so the various different services and programs we offer in service of that, that's what my team manages. So I have a series of experienced innovators that have worked from Diageo, New Balance, across the corporate innovation life that have come over to really help large companies innovate. Day-to-day, it's a lot of fun. We go from helping the blueberry industry to construction, to the cutting edge of AI. And we get to see how they all are solving unique problems in a similar way. And so it's one of the best jobs one could have.

Fred Schonenberg
I think it's really cool. You're at this sort of intersection of large companies that are trying to make a scaled impact as well as the cutting edge startups. I'm curious, are there any innovations or areas of innovation that you're particularly excited about at the moment?

Pritam Bhattarai
Yeah, I mean, I would say that I think AI is such a buzzword these days. But one thing that people haven't really fully understood or a lot of people haven't seen is that there are new markets and industries that are being built within AI to do certain parts of the AI workflow. And I'll give you a quick example. Every brand should care about how their brand shows up in AI right now, brand visibility. How does it work in Amazon Alexa? How's it showing up in ChatGPT? 

And people might not know that we work with partners that do just that and focus on delivering those insights to brands, and they're now billion dollar companies as a startup. So I think the biggest thing from my end is this unique point of view on what new jobs to be done are being created in this AI economy as people are exploring more and more.

Fred Schonenberg
I think one thing that's interesting is you've been at VentureFuel since before this sort of AI excitement began, right? Pre-ChatGPT sort of emerged on the scene. Do you see any correlations with what I would say pre-AI craze jobs to be done versus today? One of the things I may clarify about my question is often there's a problem to be solved and the idea is what is the solution to that problem and AI maybe it's just one potential tool. Do you feel that there's any correlation between pre-AI craze and today?

Pritam Bhattarai
The innovation frameworks have stayed the same, Fred. Our application of it might be more nuanced. We might have more understanding of what AI pilots usually fail, where they're good at, where they're not and a lot more experience because a lot of people are doing that. But at the end of the day, the principles around finding the right problem and need, understanding the root cause and addressing solutions for it and all the design thinking principles related, those are still evergreen. That's so far been our awesome experience there.

Fred Schonenberg
Yeah, that's absolutely what I was getting after. Thank you for saying it more succinctly than I asked for. One question I had for you is I really wanna jump into the presentation that you gave and really because I think it's such an interesting angle within AI and innovation overall, which is this idea of pilot purgatory where I remember one of our clients says to me, we're the first to try and the last to scale. And I was like, oh, that sounds terrible. Like you wanna be, you don't necessarily need to be the first, but you wanna scale and get the impact. Talk about maybe the impetus for this idea or what the problem is around this pilot purgatory.

Pritam Bhattarai
It's actually a great kind of segue to the point we were making before, which is that the innovation process is evergreen or that has stayed the same. And so are how corporates structure themselves and how they innovate and their problems within their orgs. And so what I think we're seeing is that we're seeing a lot of experimentation. I think there's some stats that anyone can, 97%, well, I asked the crowd, like raise your hand if you're experimenting or doing pilots with AI. Almost everyone's hands shot up. Then came the question that how many of you have been able to show revenue impact from it? And then all the hands suddenly went down. 

And it could be that people aren't fully there yet. But I think most of us have seen this story play out where there is a lot of activity that's happening, great activity around AI, but that connection to commercialization and commercial impact that show up in a P&L or show up in your customers view of your brand and things like that, that hasn't fully materialized yet. And we believe that's because it's the same innovation principles. We just need to apply it here.

Fred Schonenberg
I'd love to ask you why you think so many AI pilots fail. And I'm gonna put air quotes around fail. I think that the first piece that you shared reminded me of something my old boss used to say to me, my mentor, friend, shout out to Dennis, if you're listening. We talked about the rocking horse, right? Lots of activity, but you don't go anywhere. That's what it feels like right now is that almost people has gotten permission to try things, but maybe they're not using those frameworks. So they do the pilot. 

Maybe there's some technical promise in it, but there's no commercial impact. So they do not scale it up. And we've seen so many reports from MIT and other people that X percent, high 90s fail. These AI experiments fail. Whereas when we audited our internal work with our clients, we were seeing that around 40% were being scaled up across the organization. So I would love maybe the first question is like, why do you think so many are not scaling? And then let's unpack how to get them to be successful.

Pritam Bhattarai
Yeah, what's interesting too, is that whenever there's an entire conference of innovators, emerges a certain kind of villain or anti-heroes in the story. And in this case, it's usually things like legal procurement and all the different folks that are rightfully responsible for managing risks at big companies. And what we try to do with our presentation and our communication with our clients is, you do need to change how the organization is structured to be better AI ready. And those are true problems that have hurt a lot of AI projects. 

But when it comes to AI pilot purgatory, we believe that if you're picking the right problem that shows actual value, if you're picking the right partner and choosing to build with a partner or establish the results of the partner, and if you're setting up the pilot correctly, you're going to get those wins that allow you to go against the organizational green. Because a lot of times these conversations can just be, okay, well, if only legal or skip IT or things like that. And that's not how we view it. If you're designing the right pilot, you can make the leap.

So Fred, I would say that to reverse that, there are a lot of organizational challenges that exist within different companies that are trying to navigate this new space and that's going to put some barriers. But where I think the needle moving action for innovators is realizing how they can set up their pilots in a way that can better bridge that gap. 

Fred Schonenberg
You know how much I like the letter P? We've got a rubric internally at VentureFuel, the eight P's. I just heard three that you just shared, right? Which is like the right problem to solve, the right partner to solve it, and the right path to that going to scale. And I think what's interesting is that often people blame technology or somebody like procurement. And the truth of the matter is if the problem is important to be solved, everyone gets on board, right? And then it's picking the right solution and the right way to get it done.

We have a podcast that we just recorded that has not been published yet with a Wharton professor who was saying that he believes that procurement will be the center of innovation excellence moving forward. And I think there is an opportunity to really think through legal, cyber, and procurement as not defensive skill sets, but really value creators as they start to think and move differently. 

Pritam Bhattarai
100%, that is the only way moving forward. And so once you start putting yourself in their shoes, you start to kind of change what actually needs to be demonstrated. That's what we're seeing. And picking the right problem, we did a short exercise with the folks there. I think it actually, there's a podcast episode for folks to go on the Kellogg professor that you had talked with two by two. The idea is that when it comes to use cases of AI, on the bottom axis you have, are you doing it for productivity or are you doing it for growth? On the top axis you have, does your customer see it? Is it external or is it mostly internal? 

And the issue is, when AI teams or innovators are asked by their C-suite to AI transform their innovation, they rightfully start from the bottom left where it's all internal productivity. And in this exercise, we asked everyone to put where their use cases are. And what becomes really kind of obvious is the big wins are those that are starting to be more customer facing and can be a new product. And so we saw some innovators share some interesting examples of what they're looking at in the top right, but that is what we need to also pick the right problem and get a nice portfolio approach of different problems you're trying to solve.

Fred Schonenberg
Yeah, I love the portfolio approach too, because there's some value in starting in the bottom left, which is the internal efficiency type of problems to solve, because they tend to be quicker wins and very tangible. So I do advocate for that portfolio approach where you are taking the larger swings with the bigger return, but also stacking a couple of wins that are quick and very visible internally. 

I do have a question for you. So one of the things we talked about is we see a lot failing, but with our deployments, we've seen a lot succeeding. I wonder if there's anything from your session, from the playbook that you shared with everybody that you would point to that's helping to move from experimentation to this corporate impact.

Pritam Bhattarai
Yeah, I think one thing that'd be interesting for folks is just how we're thinking about the pilots. This doesn't only limit itself to AI. Every venture field person kind of learns through how to structure pilots by testing the most critical thing that needs to be tested. Oftentimes we would think that if a startup and a big company meet, they would know exactly what they need to do. And we find that in many cases, that's not the case. And where within AI I would say was interesting for many of our projects is that a lot of times the value comes from a change in workflow. 

Something changed about how the process fundamentally ran, not just that a new AI tool kind of showed up and made something a little bit better. So a lot of companies are testing features of AI like, oh, can this make me move X percent faster, et cetera. Instead of saying, if I move X percent faster, this is the quality time throughput dollars that I could get versus my old, can we make sure that that workflow works? When you do that small change in thinking, you're starting to input and need to do some back of the unfold math of, is this the right problem? Did we set the right bar of what a home run really looks like? Is it 5% gain, 10%, 20%? So taking those principles of pilot scoping in AI has been what has really kind of helped teams think about the commercialization piece. 

Fred Schonenberg
Are there any common mistakes you see or that were shared in this presentation or the discussion around it really, that you see are preventing large organizations from doing this successfully? 

Pritam Bhattarai
No, one small thing I haven't hit on yet is, AI can obviously be very scary because it does a lot of exciting things, but you need to be able to trust it. And a lot of times when corporates are setting up these types of experiments, that piece is usually set as like, is it accurate? And the moment things are not accurate, it's like, it doesn't work, it'll never work, et cetera.

And what was an interesting point of discussion that we had with our clients and also in this conference is what does it actually mean to trust AI? And how do you design a pilot around it? Because what innovator stakeholders are really nervous about is not just accuracy, but it's also sometimes about explainability. Can we actually understand what's happening? It's about oversight. Have we identified who actually, which human still is accountable for this AI to work? Can we audit it? And can we actually recover from our mistakes? 

In AI drafting an email is very different from AI approving a chemical process. And so when companies apply a right level of this is what needs to be true about how we need to trust it and test according to that, that's how we would go about it. So many companies put too high of a threshold right off the bat on what needs to be true about the accuracy and miss that. 

Fred Schonenberg
Is there anything else about your presentation that you think we should share with our audience? I know we could go a bunch of different ways, but I also wanna save a little bit of time to talk about what you learned at the conference. But from that pilot purgatory presentation, was there anything else that came up that you think people listening should know about?

Pritam Bhattarai
Yeah, and I think the last piece that we talked about was some kind of organizational adjustments that need to eventually take place. How one has to be careful about treating AI as purely software because it operates very differently. And so we need to bring in IT and also expand beyond. And also learning about how to actually get AI deployed across the organization. 

We talked about some case studies of how our clients do it, where they have a capability, a team that is looking at a centralized AI space, collecting use cases from all their BUs and then getting the BUs to raise their hands on pilots and creating that pull effect versus a push. So I would say that the key thing is within your organization, are you pushing AI or is it being pulled? And are you able to create centralized standards but with decentralized value creation? 

Fred Schonenberg
I love that. Yeah, I just came off another call with one of our clients. And what's interesting is this push-pull dynamic. We advocate very strongly for there has to be a problem we're solving. There has to be a business unit, somebody in the organization that is saying, hey, I need this thing solved in a better way than the status quo solution. 

Like that is critical because you have a buyer, you have a reason for being there. This is not then the startup petting zoo and you've got something. What's interesting is this conversation we just had was, and you have the opportunity if you're in innovation to also inform the larger organization about what's possible. And oftentimes everybody that is in their business unit doing their job has laser focus on how to do that thing really well. And so they may even find an AI tool that does their thing really well. But that ability to say, hey, by the way, I know you're not seeing the cutting edge of all these different things. Here's what is possible. And then does that solve a problem that you didn't even think you would be able to solve, right? But it's just as important. 

So it's very interesting, the push and pull. You've heard me talk about it a million times. I very much advocate for you to need the business clamoring for it. But I also think there's this moment of provocation that you can't lose, right? And it's why whenever we do a project, we try to include the wildcard case study to be like, hey, by the way, here's one you never would have thought about, maybe from an adjacent industry or something very creative and a new way to solve a problem.

Pritam Bhattarai
Yeah, and in many cases, those are actually the ones that ended up being the competitive advantage for our clients. It's the AI, the digital tool that meshes with the product that made it more sticky. And so, yeah, I agree completely, Fred. 

Fred Schonenberg
So I think one of the things that we don't talk about a lot on this show, but is like my favorite, one of my favorite parts about VentureFuel is we do these innovators dinners. And it usually tends to be around a conference when we're in a different city or around an industry where we pull in venture capitalists, the corporate innovation leaders for that industry, and maybe a startup founder or two. 

And we have like a very sort of just, have a steak or something, have a dinner and figure out where the pop is going. And I wasn't able to go to Toronto, so I had a complete FOMO moment here if I did not get to hear what people were excited about. Was there anything from that dinner that was thought provoking or even outside the dinner at the conference where it made you think differently or get excited about something new?

Pritam Bhattarai
Yeah, I mean, I think during the day there were a lot of things like venting or talking about the problems. And I think by night, it was all that had been safety valved out. And it was actually an exciting moment of possibility. So what I think that's always exciting about these dinners is the fact that we bring in different industries together and the ability to see that one industry insight applies to the other, that's what's always so exciting.

So in this case, one particular example was how in the communication space, there's a company with a lot of different assets that we're trying to connect together. Like how do we make the sports tickets show up on this app and help this consumer experience go really well? Well, that was actually very similar to how a pet food company took a look at the ecosystem of pet care and the different hospitals they worked with and veterinarian people they worked with and channels they worked with. And those insights being able to be like, oh, have you thought about this? And have you considered that actually landing? That's what's really exciting.

Cause we just wanna have a good dinner. That's all we're there to do, to talk and get excited about innovation. But being able to actually see creative sparks fly and those collisions take place, that's always the big highlight.

Fred Schonenberg
I always think it's so interesting too because while every problem is different, they're all similar, right? There are challenges that people are solving across multiple industries, but they tend to see the solutions that are for their industry. And so I love that moment of dot connection, right? Where it's like, hey, wait a second, like pet food, you should be talking to this media and tech company and you should be talking to this construction company cause really you're solving the same problem with different regulations and stipulations and it's gotta be customized. But it's really, it's cool when you can put them in the same room and watch them fly about.

I'm curious if there was any movement within the conference when you're meeting with so many different groups and hearing people speak with how corporates are partnering with startups. Was there a trend change or something that you noticed different in that world from the last innovators conference that you and I attended, I believe in Phoenix, like about a year ago. Was there anything that jumped to you like, oh, hey, this is, that relationship is evolving because I saw this?

Pritam Bhattarai
Yeah, I think it was an interesting little bit of a pendulum swing moment too. I think the interesting piece that I noticed was a lot of innovators because AI tools were trying to build something themselves. And so there was a little bit of a moment that everyone had was like, oh, we started building and then we realized, wait, we could never keep it up. We can't have the team that fully, be able to scale this, et cetera. How do we jump back into partnering? And so everyone seemed to be coming into that little bit of like an aha moment of that build partner position there. Because yeah, you can maybe build it, but is that what you should do? And I think that that was what was interesting. 

Fred Schonenberg
The buy build partner framework, it's been so interesting over the life of VentureFuel watching that evolve because big companies became big because they're great at building. And then all of a sudden they hit maybe a ceiling on that core thing that they did. And so they had to start buying, right? They sort of the evolution of R&D to M&A and not many have gotten that in between stages, right? Of partnering. And I think AI companies, startups like Repl.it that enable you to build something really interesting without having to know coding has enabled the idea of, oh, we can build all sorts of things, which is awesome. And I love it. 

There's a second step there, that is, you have to operate that thing and you have to operate it at the cutting edge of where you need to be as an overall organization. And I think that's what ends up happening. I think it's really easy to build 75% of their model, which is really cool for presentations. And then when it gets into the intricacies of plugging it into all your different APIs and all the different users that you have in your organization, it very quickly becomes a different beast.

So just for those listening, like we always think of the buy build partner framework is like, if it's core to your business, like it is the core thing, you absolutely need to build it because it touches IP or it's proprietary or it's competitive advantage. And if it's not, the best people that are out there building that thing, you should partner with them because A, you can start on third base, they're already doing it. You don't have any sort of catch up time. And then also while you're catching up, they're moving forward. And so there's this, you're racing against the clock, in a way that is really difficult. That is super cool.

All right, Pritam, I'm gonna get you out of here. You know this, a couple episodes ago, we decided we're gonna end each episode with a rapid fire. I'm gonna ask you a bunch of questions and just give you a one sentence reply and just from your gut. So you're ready to go?

Pritam Bhattarai
Yep.

Fred Schonenberg
All right, so what is one AI tool you can't imagine working without?

Pritam Bhattarai
Granola AI. 

Fred Schonenberg
No doubt. What is one emerging technology that you're watching closely?

Pritam Bhattarai
A lot of the agentic commerce space overall, all the technologies being developed there.

Fred Schonenberg
Is there a startup trend that more large companies should be paying attention to?

Pritam Bhattarai
If you're in anything that touches on lifestyle and food and beverage, understanding how GLP-1 impacts your brand and the different technologies in food tech, I'd say.

Fred Schonenberg
What is an innovation myth that you wish would disappear?

Pritam Bhattarai
So many. That we could build it here. I think it's just a non-invented here syndrome, if that could go away. I think so many problems can be solved in the world if that mindset happened to shift with the one. Sorry, that was more than a sentence.

Fred Schonenberg
No, it's a good one. And the non-invented hair syndrome, and boy, is it just, it's human. It's just the way we are, right? Like people believe they can build it. And there's an amount of ego that is healthy when you're at a successful company and you're a leader. But boy, I agree with you. If there was a willingness to say, hey, maybe someone else might have a better mousetrap or let's work with somebody on creating one. 

Okay, this is one of my new favorite questions. We've called it startup partnerships for years. It's now being called corporate venture clienting, the idea that the corporate and the startup are, it's a business arrangement. So this is the question, corporate venture capital or corporate venture clienting?

Pritam Bhattarai
Oh, corporate venture clienting.

Fred Schonenberg
How come?

Pritam Bhattarai
Drive commercial outcomes with more certainty and then you can still invest, right? I prioritize the corporate to hit those big problems fast.

Fred Schonenberg
I love it. What is one book every innovation leader should read?

Pritam Bhattarai
Let's see. I would say the, man, it's hard not to always start with the Innovator's Dilemma, but I'll say a little bit different. I think that in my career, 10 Types of Innovation by Keeley has been one of the more influential works for me because it opens up the mindset of what innovation is and all the different types of innovation that are out there. So I definitely would recommend that for anyone.

Fred Schonenberg
I love it. What is one prediction for where enterprise AI will be, let's say, three years out?

Pritam Bhattarai
I think that we're going to see a lot of AI usage, but a focus on, we're going to go from, right now we're thinking, what are all the things that AI can do, et cetera, et cetera. I think that we're going to pivot to how we keep the human soul in our work, in our brands, and everything. And so I actually think that the trend of where enterprise AI goes is how to determine what the human soul part of your brand and company is and how it's in the organization. And so it's kind of a yin and yang there.

Fred Schonenberg
All right. What is one word to describe the future of innovation?

Pritam Bhattarai
Exciting. Oh, man, that seems so cheesy, but we'll go with it.

Fred Schonenberg
Yeah, go with it. I like it. I mean, I think it is exciting and that's why you and I both love the work that we're doing, right, is that you get to solve hard problems in exciting new ways. So that's perfect. So let me get you out of here on this. Thank you for joining us today, giving us the behind the scenes.

Pritam Bhattarai
Those were hard questions. That was good.

Fred Schonenberg
Yeah, we've got to keep you under your toes. So one of the things that I think is really interesting is trying to share some of these learnings. Obviously, we always protect our clients and their information. But what we wanted to do is do a complimentary AI opportunity prioritization session. And this is kind of taking the learnings from what you presented to innovators, some of the work we do with some of the largest companies in the world, and help innovation, R&D, and digital transformation strategy leaders identify AI opportunities that are most likely to deliver business value. Could you tell us a little bit about what participants can expect from that session?

Pritam Bhattarai
Yeah, and I think we were talking earlier about a portfolio approach. And the main thing that we want to make sure participants don't leave addressing is any blind spots. If they have a blind spot in the way that they're thinking about AI, how AI should be used in their org and their approach to it, a prioritization session like that can help get a secondary perspective to make sure you're not missing anything. So if there's a lot of activity you're doing in AI, and you want to make sure that that is the right activity without any blind spots, I would definitely recommend it.

Fred Schonenberg
Great. And we will include a link so that anybody listening can schedule that complimentary session, again, around sort of the prioritization of AI within your organization, how to figure out what to attack first. And Pritam, thank you again for taking the time to share some of this with us. Obviously, we'll have you back on the VentureFuel Visionaries very soon.

Pritam Bhattarai
Awesome. Thank you so much, Fred. I really appreciate it.