In this episode of the AlchemistX Innovators Inside Podcast, Ian Bergman sits down with Celia Wanderley, Chief Innovation Officer and Head of AI at Bits In Glass, to explore the mindset, methods, and guardrails that fuel real-world innovation in an age of rapid AI advancement.

Here are the five main takeaways from their conversation:

1. Embrace Continuous, Bite-Sized Learning

Innovation moves fast—and so must you. Celia compares keeping up with AI breakthroughs to “cleaning your house”: small, consistent efforts avoid overwhelming backlogs.

• Surface vs. deep learning: Skim headlines and newsletters for trends, then dive into partner trainings or university papers to master new techniques.
  
  
• Hands-on experimentation: Build small prototypes or sandboxes to test the latest tools, rather than waiting for polished products.

2. Manage Innovation as a Portfolio of Experiments

Putting all your eggs in one “big bet” fosters fear of failure. Instead, Celia recommends treating innovation as a portfolio:

1. Generate a mix of small and large ideas.
   
   
2. Run parallel pilots to test viability.
   
   
3. Pivot quickly when one approach stalls, and circle back later to promising concepts.
   
   

This VC-style mindset not only accelerates learning but also de-risks individual projects.

3. Tie AI Projects to Business Strategy and Outcomes

A perfect predictive model is pointless if the organization can’t act on it. Before investing in an AI proof of concept, ask:

• What problem truly matters right now? (e.g., cost savings, revenue growth, talent retention)
  
  
• What operational changes are needed to leverage new insights?
  
  

Celia’s “AI primer” process aligns every use case with clear business priorities, ensuring POCs become production solutions.

4. Implement Data and AI Governance with Cutting-Edge Tools

Governance isn’t a one-time task—it’s an ongoing journey tied closely to data lineage, quality, and ethical use. New “AI-augmented governance” platforms can:

• Auto-generate metadata and data lineage
  
  
• Track model inputs, outputs, and applications
  
  
• Highlight potential biases and compliance risks
  
  

By automating mundane governance tasks, organizations are far more likely to maintain robust controls while still moving fast.

5. Cultivate a Collaborative, Risk-Aware Culture

True innovation requires both experimentation and guardrails. Celia emphasizes:

• Involve stakeholders early (security, compliance, legal, procurement) to provide “air cover” for pilots.
  
  
• Build internal champions by demonstrating small wins and sharing practical lessons.
  
  
• Normalize being wrong: Frame failures as learning opportunities and translate insights into value.
  
  

This balance of support and accountability helps organizations push the envelope without exposing themselves to unchecked risks.

Ready to accelerate your own innovation journey? Revisit this episode for deeper examples, or connect with Celia on LinkedIn to see how Bits In Glass is redefining intelligent automation in regulated industries.

Have a question for a future guest? Email us at innovators@alchemistaccelerator.com to get in touch! 

Timestamps

🎙️ Episode Introduction and Innovation Agitator Context (00:00:00)
🧹 The “clean your house” analogy for continuous learning (00:01:46)
🎯 Starting innovation with the right problem—client needs over tech hype (00:04:25)
🙋 Introduction to Celia Wanderley, Chief Innovation Officer at Bits In Glass (00:07:41)
💡 Defining innovation as implemented change, not just a think-tank idea (00:08:53)
🔄 Reimagining change management: involve stakeholders from day one (00:11:50)
🚧 Treating proof of concept as a stage in the journey, not the final goal (00:18:19)
📂 Managing innovation as a diversified portfolio of experiments (00:23:37)
⚖️ Embedding continuous data and AI governance throughout the lifecycle (00:39:09)
⚠️ Responsible AI experimentation: setting and enforcing guardrails (00:44:59)
🛬 Learning from bold experiments—Air Canada’s chatbot case study (00:46:47)

Full Transcript 

00;00;29;00 - 00;00;56;23

Ian Bergman

Welcome to season six of Alchemist x Innovators Inside the podcast, where we explore the world of corporate innovation and dive deep into the minds and stories of innovation. Thought leaders crafting the future. I am your host, Ian Bergmann, and if you're an innovation agitator like me, then this is where you want to be. You know, you say that it's never too late to pick up new languages and stuff, but I do feel like the learning curve is so steep.

00;00;56;25 - 00;01;31;28

Ian Bergman

And I was just at a meetup in Vancouver last night. Actually, it was just a casual meetup, but it was really cool with a bunch of security and IT security professionals talking about LMS and AI and how they're supporting red teaming operations. And it was super fascinating, both because it was really interesting to see how the security professionals are applying these productivity enhancement, but also discovering some very creative new exploits, you know, but also because it's eye opening on what the bad actors have in their toolkit.

00;01;32;01 - 00;01;46;16

Ian Bergman

And so, you know, I feel like I feel like all of us in the innovation space are just constantly trying to ramp up on what are the latest tools, what are the latest capabilities, what is real versus hype? Yeah. So like my brain's fall, you know.

00;01;46;18 - 00;02;05;01

Celia Wanderley

You know what what you said at the beginning, my chat is not that you have to be on the road, but for the learning. But you have to be constantly learning because it's just otherwise it creates a backlog that is too high. Right? So it's it's easier to, to keep up. It's like almost like, you know, cleaning your house.

00;02;05;01 - 00;02;20;10

Celia Wanderley

If you if you live at the complete mess, then you know, it's going to be painful. But if you are constantly, you know, taking care of it a little bit, then the chat that great, this huge mess, then it's much easier. You don't have to spend as much time keeping up.

00;02;20;12 - 00;02;34;14

Ian Bergman

Well, I'm just gonna jump right in with this. Like, how do you keep up, actually, because it's hard. Like there's newsletters, there's X feeds, there's conferences, there's hands on work and playing with the tools. Like how do you keep up?

00;02;34;17 - 00;03;03;08

Celia Wanderley

I would say it's a bit of everything. So there's of course a lot of really there's the reading that is let's say may be superficial. Right. And the things that you get on your LinkedIn feed or, you know, some newsletters and that you skim through, but there's also the reading that you have to do in depth. So for a lot of the things that I work with, I have to go into technology partner training and actually take the training, try some hands on activities.

00;03;03;11 - 00;03;34;03

Celia Wanderley

But in some what is what is very fascinating that I have been experience. I'd say probably from the last I know even more in the last 3 or 4 years, but there was something that I did not expect, let's say, as I was going through university was that or as I left university was that you would actually be looking at, let's say, university papers write articles being published and be able to use that technique that week in the in the industry problem.

00;03;34;05 - 00;03;40;27

Celia Wanderley

So we're seeing we're doing would be a whole lot more of that in that space. And that's that's when you think about it that's unthinkable. Right.

00;03;40;27 - 00;03;48;02

Ian Bergman

So like I mean we're talking like fundamental cutting edge research into quote unquote production in the days or weeks. That's unreal.

00;03;48;03 - 00;04;11;13

Celia Wanderley

Yeah, yeah. To solve a true crime problem. It's not that you're not going out and trying to find, oh, let me find a place to apply. This research is the other way around. You're trying to solve the client's problem, and you're reaching the limitations of what's already, you know, available in, in, in the, in the commercial platforms. And you're bringing in that technique into a commercial platform.

00;04;11;16 - 00;04;25;00

Celia Wanderley

But you using very novel technique, like a different approach that all of a sudden unlocks, you know, what you are trying to do. So that's not to I'd say that that that did not used to be very common.

00;04;25;03 - 00;04;53;27

Ian Bergman

Well, it didn't. And to your point, it is unthinkable in some ways. Right. But also to your point, great innovation, great ideas. Start with a problem, right? They start with understanding a problem and the ability now to go out and look at what's at the cutting edge rapidly and implement it for your clients. Like, I think that's an underappreciated and wildly underappreciated capability now because, you know, you're sort of talking about it like it's normal and it probably is in your day to day.

00;04;54;04 - 00;05;02;05

Ian Bergman

But I'm willing to bet that most of your clients, your partners and the people that you interact with don't even understand what's possible.

00;05;02;07 - 00;05;29;24

Celia Wanderley

They don't. Yeah. And in some ways I do. And I've had this conversation with a number of I have a number of people that I mentor that I have conversations with, and they've had that type of conversation with them. I believe that's going to be a great differentiator for the professionals of the future. Right. That is really there's so much ability to now, let's see, to have tools that will assist you in creating the solutions.

00;05;29;26 - 00;06;06;18

Celia Wanderley

But being able to understand possibilities and link those possibilities to, you know, to ways to solve them. That's not I don't think that that is that as you said, there's underappreciated type of capability. And we need to get a lot better on because some of our clients we can't it's it's an unfair expectation. I find sometimes that they would know that it is possible to solve those problems, because in some cases they have been told, and I'm speaking real about a real case here.

00;06;06;20 - 00;06;20;26

Celia Wanderley

They have been told for 20 years, you know, or more. That's not a solvable problem because this because of this set of constraints. Right. And that used to be true or it's not a solvable problem in a scalable way. Right.

00;06;20;26 - 00;06;22;01

Ian Bergman

To within constraints.

00;06;22;01 - 00;06;57;24

Celia Wanderley

Yeah, that used to be true. But things have shifted so much in this, let's say the last 2 or 3 years that that you have to be like, you know, flipping those rocks and rethinking about, the answers for those problems because it it's no longer true for everything that those problems are not solvable with technology. But I think we need to be bringing in, you know, that open mind and, and then helping, you know, educate people and, that you need an openness on the other side, of course, to, you know, to even having these discussions.

00;06;57;26 - 00;07;15;12

Celia Wanderley

But but I do see a lot more of our role is not just the, let's say, the order taker and the implement, the step by step of a solution, but it is making sure that people know that it is even possible to solve some of those longstanding, you know, challenges that, that were not possible before.

00;07;15;15 - 00;07;40;28

Ian Bergman

Well, and and isn't that what innovation is? Right. Closing the gap between the problems that exist in your world and whatever, at whatever scale and what's possible today? Right. The the art of the possible, the cutting edge. So I think that's actually a great bridge. I want to introduce you, Celia, to our audience. For folks who don't know her, I'm incredibly pleased to welcome to innovators inside, Celia Wander Lee.

00;07;41;00 - 00;08;09;07

Ian Bergman

Celia is the chief innovation officer and head of AI at Bits and Glass, and you've had an awesome career. I'm not going to enumerate it all, but, you know, most recently you were also the chief technology officer and head of invent, which is a wonderful title, an incredible Canadian startup, Alta ML. And you've got a pretty long career working at the intersection of both in-house and client business and strategic and technical challenges and innovation.

00;08;09;07 - 00;08;12;29

Ian Bergman

And so I welcome to the podcast. Let me just start there.

00;08;13;01 - 00;08;15;18

Celia Wanderley

Thank you. Yeah, it's a pleasure to be here. Yeah.

00;08;15;20 - 00;08;38;19

Ian Bergman

Yeah. I'm so intrigued to get your take on this world that we're living in, this world of a constant tension where we feel pressure to innovate, to keep up, to stitch, to remain competitive, where we feel pressure from the outside by the advancing state. Art of the possible. Like every day it feels like there's something new and I and people are telling you, I can do anything, which isn't always the case.

00;08;38;19 - 00;08;53;18

Ian Bergman

But that pressure is there. So I'm really I'm really curious to kind of get your take on this. I'm wondering if I can step back, Celia, and ask you a simple but hard question. In today's world, what is innovation?

00;08;53;21 - 00;09;30;01

Celia Wanderley

That is indeed, you know, a tough question. And and I would say I'm always I'm always inclined to practical innovation. Right. So one thing is for you to have a like a think tank and, and, and be talking about innovation and be writing about innovation. But I have been always am a lot more biased to implementing innovation. And so I take a lot of I know, I think that my energy comes from seeing the impact of innovation applied to to business problems, to people's workflows.

00;09;30;03 - 00;10;05;04

Celia Wanderley

So I would say that, that's that's what's somewhat fascinating about what we have be living at the accelerated pace of innovation that we have all been experienced in, in and especially in the, in the last, let's say, five, 5 to 10 years has allowed for that to become more, more visible. So you're not you're not seeing organizations, or even in our consumer life when we're no longer waiting several years for everybody to try it out and for us to be convinced that, okay, I should be really looking into this.

00;10;05;07 - 00;10;09;05

Celia Wanderley

There's there's more of that, that early adoption type of mindset.

00;10;09;06 - 00;10;12;11

Ian Bergman

Yeah. This experimental. Let's try it out. Let's see what happens.

00;10;12;16 - 00;10;46;17

Celia Wanderley

Yeah. And and part of it because of the fear of missing out, which, you know, okay, what what's going to happen? How much disruption am I going to be experiencing? Am I going to be on the let's the corporate side is like, no, are my competitors going to outpace me? And if I don't start right away and in our personal lives and so am I going to be, you know, no longer be able to advance in my career or not know, not be able to compete with my peers if I'm not adopting some of this practical innovation.

00;10;46;17 - 00;11;02;06

Celia Wanderley

Right? Yeah, I'd say that that that's, that's this is where I think innovation is about transforming something that that you do and the ways that you do things, and that is to, you know, to not doing things always the same way that they have been done before.

00;11;02;08 - 00;11;24;10

Ian Bergman

And that's a tough mindset, right? Like that's, that's a mindset that you have to be willing to take some risk. And I think, you know, that's, you know, the counter side of innovation. The flip side is always this kind of risk assessment. But I think it's a really interesting point. And you said something else early on that I want to maybe put a point on and talk about a little.

00;11;24;12 - 00;11;44;08

Ian Bergman

Right. You talked about has to be something that is implemented. It can't just be an idea, can't just be a think tank concept. And, you know, at least in my mind, innovation is not innovation. If it doesn't get people to change behavior, right, like people change behavior when there's a better way of doing things. So that's that's innovation.

00;11;44;14 - 00;11;50;17

Ian Bergman

But let me start there. Do you agree or is can innovation can change be forced?

00;11;50;19 - 00;12;16;26

Celia Wanderley

I think we have so many examples like, you know, day in and day out that it is really hard for change to be forced. Right. And that's partly why so many, you know, projects in this area fail is because I think we have had for a very long time, we have had this idea that change management, training, communications happen at the end after we have tried everything, everything has gone well.

00;12;17;00 - 00;12;18;24

Ian Bergman

What what a interesting point.

00;12;18;24 - 00;12;50;16

Celia Wanderley

Yeah, yeah, that's kind of backwards because if you are not evolving, those who will experience change to be your greatest champions right from the beginning, how can you expect that, there's going to be sense of ownership, right? And an enthusiasm to to embrace change and to do things differently. So I think part of it is just these are like in some ways a bit of that waterfall type of mindset that we have operated in, probably because of the risk, you know, being being very risk averse.

00;12;50;16 - 00;13;00;21

Celia Wanderley

We don't want to disappoint those stakeholders if something doesn't work right, but then you're we're missing out on involving them and having that sense of joint ownership.

00;13;00;22 - 00;13;22;01

Ian Bergman

Yeah. And it and it's really interesting. There's some sort of cultural and process innovation or change required to implement innovation. What we're going to come back to that. Let me let me ask you, tell us a bit about bits and glass. Tell us a bit about the work that you do and your responsibilities as chief Innovation officer and head of AI.

00;13;22;04 - 00;13;43;03

Celia Wanderley

Okay. Yeah. And of course, so Bitchen Glasses is is a company that was founded here in Western Canada, and I'm business administration from Brazil, but based in Edmonton and that is a company that's now a little over 20 years old. But that grew to, to be a global companies. We're still in a small medium business around 250 people.

00;13;43;05 - 00;14;11;09

Celia Wanderley

So globally, with our team spread across Canada, USA, the UK and India. Okay. And now we what we how we define what we do is like intelligent automation. So we work with data, we work with processes and we work with I know I know, which is very much embedded in the internal got data and processes to help businesses re-envision their workflows.

00;14;11;11 - 00;14;25;08

Celia Wanderley

Sure. Well, if you think, you know, 20 years ago, it's people moving from paper to digital. Right now, it's I'd say it's in some ways it's like digital to digital. What it's like it's like another level of automation.

00;14;25;10 - 00;14;37;12

Ian Bergman

Now that's a good point. It's like the it's not this business process workflow automation of the past where we're replacing stamping, you know, stamping insurance forms with automation where completely rethinking digital processes.

00;14;37;12 - 00;15;04;23

Celia Wanderley

Completely rethinking digital processes. Right. So somewhat we were quite a bit of insurance, for example. And and it is fascinating because of course highly regulated industry. And we know for good reasons, risk management and and compliance and all of those things need to be in place. And be a big part of of any decision. But there's we're seeing our clients start to be open to, okay, how do we do smart ingestion of documents.

00;15;04;23 - 00;15;29;13

Celia Wanderley

Right. How do we get help with, you know, document comparison and how do we tap into our knowledge bases of compliance or governance documents in a way that doesn't require the, you know, the human to be and going and reading every single, you know, line of those documents. Those are not just going to be these discrete features in an application.

00;15;29;18 - 00;15;40;29

Celia Wanderley

It's going to mean a significant change into these workflows. But we were starting to see I know people ask for it and and willing be willing to embrace it.

00;15;41;06 - 00;16;13;12

Ian Bergman

And so let's talk about some of these projects because, you know, you you've got, a career of working in and around data and the foundations of what has become modern AI. But let's let's talk about more recent projects. What are some of the indicators of success or the patterns that lead to success as your clients and your partners are evaluating AI as they're thinking about, you know, transforming processes that may have been part of a digital transformation process a decade ago.

00;16;13;14 - 00;16;24;17

Ian Bergman

What what do you look for and what do you need to make sure exists in the end, for a successful AI project? You know, a nice a nice small question for you.

00;16;24;22 - 00;16;48;03

Celia Wanderley

I love the question. Yeah, I was going to say that is that is the million dollar question. Right. So, you know, I'll start with an example. And it's not maybe a recent example, but I think it illustrates the point. You know, quite well, one of the things that, let's say in the, in the pre generative AI era, one of the use cases that seems to be very popular is customer churn rate.

00;16;48;06 - 00;17;13;23

Celia Wanderley

And and I would hear that over and over from clients. And and at some point I started to put the question back to them and say, let's imagine that the results, of the model are just going to be perfect. I mean, just you're gonna you're gonna know exactly. You know, based on, let's say, the metrics by which we, we measure, you know, we measure, you know, successful modeling, let's say it's going to be the best we get.

00;17;13;25 - 00;17;18;02

Celia Wanderley

What are you going to do with this new insight?

00;17;18;06 - 00;17;19;06

Ian Bergman

Sure.

00;17;19;08 - 00;17;20;02

Celia Wanderley

You can never be.

00;17;20;02 - 00;17;24;15

Ian Bergman

Sure your your perfect ability to predict customer churn. Wonderful. What are you actually going to do with it?

00;17;24;16 - 00;17;49;16

Celia Wanderley

What are you going to do with it? Because arguably you you're going to need to maybe invest in into, let's say, more resources to be giving those customers a call or are you going to have, you know, some reward, some incentives, some and you would be surprised with how many people actually would do nothing who are prepared to not in they just could, you know, make additional investments.

00;17;49;18 - 00;18;19;12

Celia Wanderley

So that was a signal that if you were not, I think what to put that back up a level. What that means is you're not prepared to make changes in your workflows, potentially re-engineer them. Change how you were doing things today, then it's, it probably means that, you are not ready for, you know, even for a very successful, result of AI, which is not always the case.

00;18;19;12 - 00;18;47;26

Celia Wanderley

Right? So I think, the other thing that you would probably hear quite a bit is that, while 80% of, the, of the use cases never go into production at this stage, proof of concept and, and new in those early conversations, you start to see and hear those who are looking the proof of concept as the end goal or proof of concept as a stage in the process.

00;18;47;26 - 00;18;49;12

Celia Wanderley

So yes.

00;18;49;15 - 00;19;04;21

Ian Bergman

Ma'am. What what a kind of a depressing world to be in. So if you end up in a world where you realize that the is the end goal, like that's the metric, right? Like, that can be fun, that can be creative and inspiring, but you're not going to make any change.

00;19;04;23 - 00;19;05;25

Celia Wanderley

Exactly.

00;19;05;28 - 00;19;25;28

Ian Bergman

You know, something something like one of the things, in my experience, that is almost critical upfront is, you know, you're sort of setting aside, you're set, you're assessing your problems. Okay. You know, customer churn is a problem, whatever. But you're also need to assess the ability of the organization to do anything about it, which I think is what you're saying.

00;19;26;01 - 00;19;42;07

Ian Bergman

And the relative priority to the organization. Right. If you ever if if it's ever going to move beyond POC into production, this actually has to really matter. When do you do things like that in the process, in your work? Is that upfront work? Is that is that kind of along the way?

00;19;42;09 - 00;20;04;09

Celia Wanderley

It is. It is. It tends to be upfront work. Anytime we can influence it. So we have something that's called the, you know, I call it the AI primer. And, it is about understanding what is your business strategy, what are your business priorities, and how do we tie how do we help clients help themselves? Right.

00;20;04;10 - 00;20;26;17

Celia Wanderley

So in terms of identifying those priorities, it's very easy to fall in love with, with an idea just because it's cool. Right. Or but it's, but you need to go back to the basics and say, how does this align, to my business strategy and my business priorities and, what am I trying to get from it?

00;20;26;17 - 00;20;50;03

Celia Wanderley

So I'll give you an example. Some, you know, for some. And of course, that changes, you know, depending on, on where business is that. But, for some companies the most important thing is short term savings rates. And they can eliminate cost. That could be at a point in their history. What, the most important thing is competitive advantage.

00;20;50;06 - 00;21;19;11

Celia Wanderley

There could be others where the most important thing is really talent attraction. Or in either case, it is about efficiencies. And it's those long term low savings. But if you were if you were looking for, let's say if you have, if you've promised to your shareholders that you're going to be, you know, so that that the return is going to be immediate return or term return is going to be higher, you're going to be looking for certain types of use cases where you can cut costs.

00;21;19;11 - 00;21;49;24

Celia Wanderley

So that that doesn't tend to be necessarily people, but it tends to be maybe where there are contracts, for example, even supply chain. Right. So what that we tracks where we could immediately optimize those costs. Well, where are maybe route optimization for our vehicle fleet management, where we can maybe optimize some, I know some of those routes and then reduce fuel consumption or things like those that that can lead to immediate savings.

00;21;50;01 - 00;22;13;08

Celia Wanderley

So I think it's there's, there's a bit of a, of a method to the madness to be able to align the specific things you're going to be working on back to, to the business strategy and business outcomes. Because if you do not, even if those are very successful, they may never see the light of day, in terms of operationalization because, you know, just it's it's a nice feeling, good results.

00;22;13;09 - 00;22;33;03

Ian Bergman

It's nice to have. Right. It's the painkiller, but or the, the vitamin, not the pain killer, but. Okay, so let me I want to challenge you on one thing though, because I think I fundamentally agree with you. And I actually see the biggest mistake and for a lot of folks make that I talked to is falling in love with the idea to use your words, or falling in love with the potential solution and not really evaluating.

00;22;33;03 - 00;22;56;26

Ian Bergman

Does the problem matter and falling in love with the problem? I think I've got the book behind me falling in love with the problem, but there is something to being exploratory, to immersing yourself in what's possible, to experimenting. And then, you know, realizing that you might be able to do something that dramatically changes the priority of the problem.

00;22;57;02 - 00;23;15;16

Ian Bergman

Right? Like, you know, okay, so we're trying to manage customer churn. We don't actually have a way to intervene even if we know someone's churning. But, oh, it turns out that there is an AI driven voice contact center that's incredibly friendly, that at zero incremental cost can, you know, call people at risk of churn and provide them an offer.

00;23;15;19 - 00;23;36;29

Ian Bergman

Right. And that wasn't possible before. So how do you balance the kind of being sober upfront about the importance of the problem with being exploratory, experimental, optimistic, and just and nerdy? Frankly, to try and see if if maybe there's something out there that will change the priority.

00;23;37;01 - 00;24;02;17

Celia Wanderley

That in that I love that, you know that question too, because one of the preferable and then I think what I have seen the most success in, in, you know, in this space is when companies treat that innovation capability not as a one project, right, but as, as a capability and as a roadmap of projects and ideas.

00;24;02;19 - 00;24;28;14

Celia Wanderley

So ideas can be big and small. And, the life cycle in how you going to implement them always needs to have that experimentation. But it could be that what you learn in that experimentation phase is that, we're not quite ready for it now, but we're going to bring it back, you know, let's say it's time from now because there's there's a lot of potential here.

00;24;28;17 - 00;24;50;09

Celia Wanderley

We know that it doesn't. I've been very busy with this idea for the next year. It's not going to really, you know, make it work. And we want to get back to this because there's there's a lot of potential here. So you have to nurture innovation capability and organize your, your innovation processes in a way that, that it allows for that.

00;24;50;14 - 00;25;12;14

Celia Wanderley

It's, it's a portfolio of ideas versus the, the one discrete idea that you take on as, as most people, you know, I'd say most companies still do that, that they say this discrete project, the challenge with tackling it as a discrete project is that there's always there's the perception. The fear of failure is is a big thing of course.

00;25;12;14 - 00;25;13;06

Ian Bergman

Because.

00;25;13;09 - 00;25;19;06

Celia Wanderley

If you have that one project and you have the fear of failure, which is, you know, very, very all.

00;25;19;06 - 00;25;21;28

Ian Bergman

Your eggs are in one basket, you're never going to take any risks.

00;25;22;00 - 00;25;40;25

Celia Wanderley

You're going to try to make it work, you know, any way possible because that's all you have. But if you treat that as this portfolio of ideas, you're going to know that, okay, we need to now it's time to quickly no shift to this one because we want to provide value. And these other things will we'll come back to them, you know, at the right time, right.

00;25;40;28 - 00;26;00;13

Ian Bergman

Yeah. I love that. You know, there's, you may have seen this, I don't know. There is an HBR Harvard Business Review article is probably about a year ago now. Actually, that was, you know, basically talking about how in times of uncertainty and disruption in the world, right, corporations need to think like VCs and they need to have a portfolio of bets.

00;26;00;13 - 00;26;20;17

Ian Bergman

And it was it was kind of operating at the strategic level. But the reason that article is stuck with me is one, I don't know if we're ever going to be in a world that doesn't feel like a time of constant uncertainty and disruption again, right? And things are just changing too quickly, you know, until what does that mean?

00;26;20;17 - 00;26;49;18

Ian Bergman

That means for all of us, whether we're at the CEO, whether we're an innovation manager, or whether we're a business line owner trying to figure out how to solve some KPI, like manage a portfolio of ideas and solutions and let let you know, let the let the flowers bloom. And I think that's really hard in most enterprises. And I think it's even harder in enterprises or organizations where there is a lot of outside accountability and transparency.

00;26;49;18 - 00;27;11;06

Ian Bergman

So like the public sector as an extreme, like really hard to take risks and be experimental. I think in the public sector, because you're beholden to so many stakeholders, everything you do gets audited. Nobody wants to, you know, nobody wants to have their emails hauled up on the CBC. So, you know, maybe we can use this like, how do we help?

00;27;11;09 - 00;27;35;02

Ian Bergman

How do we help decision makers? And also just kind of like regular folks trying to get their jobs done. Think about this. How do we help them be experimental, be innovative, build a portfolio of ideas in a world where, as you've said, it can be really hard to be judged for failure.

00;27;35;04 - 00;28;13;28

Celia Wanderley

And, I do think, you know, it's we almost need to make a bit of, you know, empathy. Right? I we've, sometimes we'll say, well, you know, it's just it's easy to blame, right? And say, well, you just said the sectors are so close to experimenting and to try new things, but as if you go back to what it takes and, you know, the processes and what they're bound to, as you said, what the, let's say the reputation of the image or in the case of the financial services, some my, let's say, the licenses and the, the regulatory compliance that they're they're bound to for good reasons.

00;28;14;00 - 00;28;43;08

Celia Wanderley

So I think it is it is it has to be about not using that in, in your favor and not as, not as, as a way, a reason for not to do it. Right. I think with, with public sector, for example, part of the, the challenge in this space is that, as you know, there's a lot of concern over time, data needing to be used, for the purposes that it was collected for.

00;28;43;10 - 00;28;43;26

Ian Bergman

Right?

00;28;43;28 - 00;29;16;22

Celia Wanderley

Sometimes, you see stakeholders taking that interpretation almost to AI to a limit. That's not it's an unrealistic. Right. So I think you have to you have to challenge some of the interpretations that maybe are, out of what, what is really required and, and encourage change. But within the this within, let's see a pattern that wasn't, just spoke people and then moved them completely away from trying to do it.

00;29;16;25 - 00;29;42;17

Celia Wanderley

I think, you know, I've, I've been I've had pretty good experience as with within the public sector to create, let's say, a predictable model where innovation can happen within. And they know AI innovation can happen within. And it's great to go through those processes of being, you know, the experimentation to pilot, to operationalization, but also where you involve the right people.

00;29;42;17 - 00;30;08;11

Celia Wanderley

So you you involve security, right? Do you involve compliance? You involve legal, you involve a procurement that is on site so that, so that, you know, that, that, that those folks, can have the, the certainty that they're doing things right. So you don't want to be, I think, to create the perception that that you're doing something that, that would get people in trouble.

00;30;08;11 - 00;30;29;14

Celia Wanderley

Right? So it's like, how do we how do we innovate in a way that people are willing to innovate with you? And because they do feel that that that you have their backs, right, that that you understand that they are bound by this set of, of, of good reasons or good compliance processes that that exist for a reason.

00;30;29;16 - 00;30;51;01

Ian Bergman

Yeah. You give them, you give them the air cover of a methodology, a process of strategy, you name it. But the air cover of something that is well thought out, that's a I think that's actually a wonderful insight. Are there examples, whether in Canada or abroad, that you've seen of public sector innovation resulting or an innovation process resulting in something really cool?

00;30;51;03 - 00;31;20;17

Celia Wanderley

Yeah. So I think in my prior, my prior role, one of the things that I was involved with and help, you know, envision and launch was an initiative called gov lab dot I. Right, which is still very much, you know, out there and and being successful. What they tried to tackle there were a few different problems was one bring this this notion that you don't necessarily it's easier if you if you shift the procurement mindset from that, you know, project.

00;31;20;17 - 00;31;41;20

Celia Wanderley

My project to this idea of a portfolio right of of innovation initiatives and, and as long as it's under that that methodology and that approach that you can control the pace in how you do it. And then and that that gets gets people a bit more comfortable going through it. The other thing that was very I would say that that was very interesting.

00;31;41;20 - 00;32;05;17

Celia Wanderley

There was this public, private academia partnership. So one of the challenges that you would see in the public sector is a fear that while Big Tech pays big money, right, that to some of those, you know, the technical resources. So how can public sector compete and attract, you know, the talent that has the training and the aptitude to to do this, this type of work.

00;32;05;22 - 00;32;30;20

Celia Wanderley

So, so I think what we you know, what we applied there was this idea of partnering with, you know, universities that were forming these young and very, you know, highly qualified talent, bring them in early on and expose them to the problems that public sector, solves, which are fundamentally unique. And, and that's a.

00;32;30;21 - 00;32;45;11

Ian Bergman

Fascinating and really fascinating. And you can bring in and you can bring in people who are unconstrained by like that, by what's not possible. Right? They're just they're not constrained by all the rules, all the history, all the legacy. What a great idea I got.

00;32;45;12 - 00;33;20;29

Celia Wanderley

I think we we proved that we can change people's mindset because sometimes maybe public sector work is associated with not interesting outdated. Right. On the technology side. So because if you are applying that, let's say state of the art methodologies, state of the art techniques to those very interesting problems, then that young talent starts to see that as a possible career path, and then you start to solve the talent problem, because as things move to production, you were able to actually, you know, Steph, a team that can maintain it, that they can continue to grow that capability.

00;33;21;01 - 00;33;52;14

Ian Bergman

That's a it's such a great model. And it's such an interesting example of process innovation that, that doesn't serve it. Well, I, you know, that that makes it sound more boring than it is. But but it is, it's a, it's a, you know, innovative approach to, to rethink how organizations do business. And I and I think that there's so many opportunities that most organizations can find right to rethink how they do business to inject energy, inject creativity into their workflows.

00;33;52;16 - 00;34;18;09

Ian Bergman

But there's also always it feels like almost a mental hurdle. There's always a mental block somewhere. Might be a very senior executive, might be some kind of stuck middle management along the way. Do you ever find yourself having to like, I don't know, use a stick, not a carrot, like kind of like push, shove, pull people through some kind of mental block because, you know, as an outside observer, an expert, that there's opportunity for them.

00;34;18;09 - 00;34;19;25

Ian Bergman

And how do you do it?

00;34;19;27 - 00;34;22;01

Celia Wanderley

Yeah. And I think the answer is yes.

00;34;22;01 - 00;34;26;11

Ian Bergman

It's okay. It's okay to say I hit them with real sticks. By the way, I won't tell anyone.

00;34;26;13 - 00;34;54;15

Celia Wanderley

Yeah. No, I do not hit them with real sticks. But what I would say is that that happens internally and externally, right? So let's say even though I'm I'd say right now I'm in on the consulting side and helping new clients. But in order to do that work, we also need our internal teams to be disrupted. Right? People will have, if you think about, you know, the technologies we've been implementing, it's easy to get very comfortable with what that looks like and has worked.

00;34;54;15 - 00;35;17;01

Celia Wanderley

And we've been successful and we have grown as a company. And then by doing things a certain way, and if I if I give you just a very, a very concrete example, even in the area of software engineering like the what generative the level of disruption that generative AI is bringing, right. With the system code or code called generation.

00;35;17;03 - 00;35;31;20

Celia Wanderley

And can we actually ignore that? And the kids say, no, I'm, I'm very good. And I'm going to, you know, write every line of code because I can and that's it is true. You can. But when I think about velocity.

00;35;31;23 - 00;35;34;29

Ian Bergman

You're not competitive. There's no way there's no way you're competitive at this point.

00;35;35;02 - 00;35;54;03

Celia Wanderley

There's no way. So over time, you know, it's going to last for a little bit, but you're not going to be competitive in the long run. Or let's say in our case, as we we what we what we do is, you know, we implement this, you know, this as digital transformation in practice, right? This like you be building digital workflows for our clients.

00;35;54;05 - 00;36;26;04

Celia Wanderley

Can we really ignore that? You can have a gigantic workflows that will now automate the big portion of those processes versus the the step by step that we always did? No. So that requires really a lot of that internal, you know, change management for those teams that deliver. And if you if I then isolate that into the client relationships, there's a reason why clients hiring a new technology consulting companies and hire help.

00;36;26;07 - 00;36;49;08

Celia Wanderley

And it's not to just say yes, yes, yes this dear. As for for our professional experience, for our inside. So it is I think it is fundamentally our duty to, you know, to be able to advise. And that sometimes is not what people want to hear, but it is about doing that respectfully. It is about establishing credibility. Right.

00;36;49;11 - 00;37;10;15

Celia Wanderley

And in some cases being willing to be wrong because could very well be that the specific, constraints or situation are business setting, just not ready for that level of disruption that that time. So there's it's a I think it's assumption that true collaboration where I'm, you know, I'm there to do the best work that I can.

00;37;10;15 - 00;37;16;26

Celia Wanderley

And that may not be just always saying yes to how you think, but that's why we're there.

00;37;16;28 - 00;37;42;16

Ian Bergman

And that might be a hard conversation. But boy, if more leaders of sort of services and consulting organizations thought like that, the world would be a better place because it is. It's incredibly important, especially and again, in a world that is just fast changing, right? I mean, the the challenge with the world that's fast changing is that you have to keep trying new things and you have to push yourself to keep trying new things.

00;37;42;16 - 00;37;51;07

Ian Bergman

And if you don't, you're not actually going to know what the right project, the right technological approaches. And so I love that you're able to embrace like, hey, and sometimes we're wrong.

00;37;51;09 - 00;37;51;29

Celia Wanderley

Yes.

00;37;52;01 - 00;38;21;19

Ian Bergman

But that's part of our job now for every innovator or want wannabe innovator who's listening to this podcast, I want to put a point on that. Part of your job as a, innovator is to be wrong and to learn from it and translate that that learning into value. But I think that's so important as we, as we sort of get to the end of the conversation here, though, I want to bridge into kind of a, I mean, a tough area to think about in the world of innovation and AI.

00;38;21;20 - 00;38;50;02

Ian Bergman

Right? And it's a it's tough because sometimes the consequences of being wrong or of making a mistake or of, you know, or just not predicting an outcome in AI especially can be problematic, right? They can run the gamut from annoying and embarrassing through to financially impactful in a negative way, through too deeply, deeply problematic from a safety or or a regulatory perspective.

00;38;50;04 - 00;39;09;01

Ian Bergman

And so how do you think about responsible and governance around AI experimentation? How do you think about balancing? We have to try. We have to take the risk with the fact that there are very real risks. I'm going to leave that wide open question, and then maybe we'll come to some specific examples.

00;39;09;04 - 00;39;32;28

Celia Wanderley

Yeah. And and that there is like, even when I'm not asked the question, I try to insert that type of, you know, of entry into, into the conversation that I have in a publicly because that is that is really important. What I do, I tend to approach this with, with clients is that, you know, governance is not is not a one and done governance is a journey.

00;39;33;03 - 00;40;03;16

Celia Wanderley

It never ends. Right. And and that as much as we have now and starting to see really and especially after the generative AI rise, I started to see AI governance become more top of mind. But the AI governance also structure with data governance, which we've always known we had to do. But but kind of the no, no, that's not you know, it's not so exciting to invest in in in data governance, good data quality, good data lineage.

00;40;03;18 - 00;40;29;01

Celia Wanderley

So if I, if I bring those, let's say, guiding principles back into how we do it. So you know, in the in the more practical way it goes all the way from education at the beginning to understand, okay, what are you doing for data governance? And how do you what's your understanding of what I governance looks like? And as it relates to to data governance.

00;40;29;03 - 00;40;52;00

Celia Wanderley

And so that's, you know, that that say that that education, what's the organization's understanding and practices around the who wants data? I know what, what are the implications of change? Who approves what the use of data. How do you audit things back? So those are very practical concerns. And with technology that is very good at doing those things.

00;40;52;02 - 00;41;04;14

Celia Wanderley

And I would say that the the nice thing about it as well is that we're starting to see improvements in that technology that make the work, let's say, more easier and a little bit easier to be accomplished.

00;41;04;20 - 00;41;06;18

Ian Bergman

Like, what's it like, what's the yeah. What sort of improvements?

00;41;06;19 - 00;41;25;02

Celia Wanderley

I like to say no, there isn't a good name for it, but let's say it's generative AI, augmented, you know, governance. So if you think about, you know, in the past you would have this and I I'm guilty of it going to a client doing like a very large data governance project, leaving them with 500 page decks. Right?

00;41;25;04 - 00;41;45;16

Celia Wanderley

No. Page like it's like that for them to go implement that is so hard that people don't. It's just collecting dust, you know, in the shelf technology that exists today. In many cases, it's going to be understanding based on the data that you have, let's see in the table or in the data structure, what's the definition.

00;41;45;16 - 00;42;13;29

Celia Wanderley

You know, of that data. So creating that metadata can be automatically generated lineage for how data flows and what data it is use can also be automatically, you know generated. So things that you would have to have somebody, you know, doing, you know, can get of tools and therefore make your journey of implementing, you know, data governance. And therefore it can also involve high governance a lot simpler because there's this less manual.

00;42;14;02 - 00;42;37;28

Celia Wanderley

So I think that's one thing that you had mentioned before is that sometimes people do not know what exists. So that's what that's I find that that's one of the areas that there's actually less awareness that we can now use, you know, state of the art technology to help those tasks that maybe were not very, you know, very exciting to move and therefore people wouldn't do it.

00;42;38;00 - 00;43;04;00

Celia Wanderley

And if you think about that in this lifecycle of as you have good understanding of how your data flows, what your data looks like, you know, privacy, anything that you know, so personally identifiable information, how do you match that? Who has access. So I say all of those those activities that may not be so exciting. If you have tools that are helping with that process, it's going to make organizations a lot more likely to do it.

00;43;04;02 - 00;43;32;06

Celia Wanderley

And that's going to feed into, okay, now I have this data. It comes from here. It's transforming this particular way. Which machine learning models are using this data and then how which applications are then using this machine learning models, what types of decisions are we making? Which you know, which stakeholders, internal or external, are seeing, you know, how are being impacted by this decision?

00;43;32;09 - 00;43;32;17

Ian Bergman

Yeah.

00;43;32;24 - 00;43;57;28

Celia Wanderley

And allowing you to go all the way back to what are some responsible uses, of that data. Right. So for example, if you think about the decisions about loans and am I going to be creating perpetuating bias because there's some pieces of data here that are making this model the bias towards certain decisions about a certain type of population.

00;43;58;04 - 00;44;17;19

Celia Wanderley

And what am I going to do, you know about that? So I think it's there's, there's, there's all the way from education guiding principles, but there is the use of technology to move all the way from, you know, the data governance to the good uses of the AI that will allow you to map all or all of that impact.

00;44;17;19 - 00;44;23;08

Celia Wanderley

Because if you leave all of that to be done manually, chances are it's going to get done.

00;44;23;11 - 00;44;41;08

Ian Bergman

It's just not going to get done. And I think that's such a that's such an interesting insight is that there, you know, there is this sort of governance and, and risk management that can be done effectively while taking risk. Right. But you've got to do it. You have to do it. You actually have to do it. And people don't want to.

00;44;41;12 - 00;44;59;04

Ian Bergman

So use tools. I love that thought that we can use AI. We can use innovation to help us do the things that we as people or organizations don't want to do. And so put to the bottom of the priority stack, let me ask, I'm going to put you on the spot with a yes or no question. This is a really tough one, right?

00;44;59;05 - 00;45;21;03

Ian Bergman

The context always matters, but I'm going to do it anyway. Should organizations, if, you know, try something that they think might be extremely risky, should they just put it. Should they, should they try? Should they throw it into production, even knowing that it could backfire in a serious way?

00;45;21;05 - 00;45;43;28

Celia Wanderley

That's that's a hard one for, you know, I if I just go by the words you used, I would probably say no because it seems that there's no guardrails around it. Right? Right. Should organizations try, you know, risky things? I say the the answer is yes, but I would say there's always guardrails around that. I think you have to measure.

00;45;44;00 - 00;46;06;29

Celia Wanderley

Do we have are we going to use this in a pilot setting first and understand what are understand and measure. You know what what the outcomes of of those risks were taken are and, and know that we're able to almost ensure for those risks because, that could the consequences can be very high if you have no guardrails.

00;46;07;01 - 00;46;21;13

Celia Wanderley

One human impact. So let's say that's been health care, right. So would you do that and risk someone's life. So what kind of guardrails I'm going to put so that that's not the case so that you're helping. You're still moving the needle.

00;46;21;15 - 00;46;47;20

Ian Bergman

All of it I gave you an impossible I gave you an impossible question and scenario. Right. Because of course, but I but but this point that you can take risk with understanding certain guardrails and you can you can, you know, if a pilot or a C at small scale, whatever you can do. I think I've talked about it on this podcast before, but I mean, I still remember the press around Air Canada chat bot from 18 months ago or something.

00;46;47;20 - 00;46;48;15

Ian Bergman

Do you remember this?

00;46;48;16 - 00;46;49;12

Celia Wanderley

Yes I do, yeah.

00;46;49;16 - 00;47;07;29

Ian Bergman

Yeah. I mean, and for for folks that maybe haven't heard me on a previous episode talk about this. I mean, the, the super short version is in Air Canada. I believe LM driven chat bot. Yeah. You know, promised I it promised a fair or something. Air Canada tried not to honor it. They got sued. Air Canada lost.

00;47;08;01 - 00;47;43;11

Ian Bergman

Rightly so, because an agent in this case and an AI agent of the company had had made a commitment. And it's really funny because there's this there's a press narrative that emerged around that where Air Canada's the the bad guy. And let's be super clear, I actually agree with the fact that they were the bad guy for not honoring the fair to begin with, but they are such the good guy in this innovation experiment in deploying something that can solve wait time issues and you know that can solve quality of communication issues.

00;47;43;11 - 00;48;04;25

Ian Bergman

Trying it out with some guardrails, realizing a guardrail got broken. I'm, you know, and then they're going to fix it. But, you know, for me, that's the sort of thing that maybe has a little bit of a black eye for a short period of time, but shows me that there's an organization that is actually it really deliberately and intelligently pushing the envelope of how they can serve their customers over time.

00;48;04;25 - 00;48;15;28

Ian Bergman

And like, I don't know, I, I look around and I just I don't see enough examples of that. So I love these stories of these little embarrassments where, you know, nobody got hurt or killed, but it shows that people are pushing the envelope.

00;48;15;29 - 00;48;38;12

Celia Wanderley

Yeah. No, I, I, I, I agree with you. And then, you know, in most cases I would tell you what the journey has been for that type of, of experiment is that is use the internal knowledge basis, you know, be very comfortable with them, see how they go, fine tune it there and then scale it to the public, you know, to the externally facing, chat box to a knowledge base.

00;48;38;15 - 00;49;01;26

Celia Wanderley

So I think different clients have used different approaches in their journey to, to how they do it. But yeah, and, and and you will see also a lot of I think we've been, have been shown by what, you know, even OpenAI has done with, you know, ChatGPT out there and it's hallucinating and, and all of us been comfortable, which we kind of learned before, is either you are absolutely right or don't tell me anything.

00;49;01;26 - 00;49;08;11

Celia Wanderley

And all of a sudden I, you know, we we kind of found out that, you know, what? We are kind of comfortable doing some critical thinking or.

00;49;08;11 - 00;49;35;27

Ian Bergman

Totally comfortable like that. I'm I'm in such love with perplexity, like perplexity is my entry point to the internet. And it's really fascinating how it's changed, how I think about research and summarization. Right. Because I have a built in check. Right? Like basically what happens is now my, my research or knowledge gathering workflow is ask perplexity something, get the answer.

00;49;35;29 - 00;50;00;20

Ian Bergman

And then if something seems off, go through and check the primary sources and I stuff. I find that much more. I find it much faster and much more complete than what used to be the case where I'd have to go, you know, find a bunch of primary sources, integrate them into some thesis, get myself an answer. But but I find myself just naturally adapting to the fact that, like, that first thing could be wrong.

00;50;00;20 - 00;50;26;11

Ian Bergman

It may have just completely made something up, although, you know, more likely it interpreted something a little off. But yeah, we're getting we're getting comfortable with that. And I actually think that's great because that's how humans work. Humans exaggerate humans, you know, misinterpret information all the time. And I and humans have guardrails around it. You you don't expect a police officer to do that in the same way that you might expect, you know, a casual a friend in a casual conversation.

00;50;26;13 - 00;50;27;24

Celia Wanderley

Yeah. Very much.

00;50;27;27 - 00;50;48;13

Ian Bergman

Well, so silly. I want to say thank you so much for coming on innovators inside. This has been a great conversation and I really appreciate your time. You're doing fascinating work for people who are interested in following your work, in following bits in glass. What's the best way to do that? It's website, LinkedIn. Where where should folks go to, to follow your work?

00;50;48;19 - 00;51;08;29

Celia Wanderley

Yeah, I, I love to connect with people on LinkedIn less bits in glass. We also have a presence on LinkedIn. And I try to to share quite a bit of leadership and what you see. Right. So there's always I find that it's it's interesting for people to hear directly from clients and see what they've tried, what, you know, what has been successful.

00;51;09;01 - 00;51;36;03

Celia Wanderley

We don't we all have a lot of information coming our way. But sometimes when it's distilled with practical experiences, it's, it really helps and then resonates, brings, helps differentiate hype from reality. Right in the piece that actual adoption of innovation is happening in different industries. So yes, I would love to, you know, for anyone to that's interested to, to just reach out the only theme and then follow me or in the follow up, it's amazing.

00;51;36;10 - 00;51;43;06

Ian Bergman

Well, we've been in conversation with Sally. I wonder, Lee, thank you so much for joining us and innovators inside and have an amazing weekend.

00;51;43;08 - 00;51;45;27

Celia Wanderley

Thank you.

00;51;45;29 - 00;52;06;20

Ian Bergman

And that's a wrap for today's episode of Alchemists Dax Innovators inside. Thanks for listening. If you found value in today's discussion, be sure to subscribe to our podcast and check out our segments on YouTube. Links and follow ups are in the show notes, and if you have questions you want us to feature in future episodes, email innovators at Alchemist accelerator.com.

00;52;06;22 - 00;52;11;13

Ian Bergman

Stay tuned for more insider stories and practical insights from leaders crafting our future.