Podcast
That's one takeaway from my conversation with Northern Trust Asset Servicing's Shaelyn Otikor-Miller. She joins Get Reworked to discuss the journey the company went on to vet and then launch Dynamics Copilot internally and how closely — as big a change AI may be — the rollout resembles previous technology launches.
"It does make you question what's the next phase we're about to go through? What's the next corner we're painting ourselves into? What I'm noticing is that there's a lot of passion and concerns around AI and where all the technology is going. I think we just need to slow down the mindset a bit and realize it's just another phase. It's another part of the journey," said Shaelyn.
Highlights of the conversation include:
- The hurdles Shaelyn's team tackled to prepare for the Dynamics Copilot rollout.
- How Shaelyn handles the change management side of the rollout.
- An overview of how 25+ years of technological changes have changed the workplace — and what things stay the same.
Table of Contents
- Tune-in Here
- Show Notes
- Show Transcript
- Staying Ahead of the Digital Curve in Financial Services
- The Regulatory Hurdles to a Dynamics Copilot Rollout
- The Role of C-Suite Support
- The Governance Model for Dynamics Copilot
- A Multi-Touch Approach to AI Adoption
- Where Emotions Fit in Digital Technology Adoption
- A Buddy System for AI Adoption
- The Results of the Dynamics Copilot Rollout to Date
- How Northern Trust Decided on Where to Start With AI
- A High-Level Overview of Technology Transformation
- Final Words of Advice to Digital Workplace Leaders
Tune-in Here
 |  |
Show Notes
- Shaelyn Otikor-Miller on LinkedIn
- Shaelyn's previous visit to Get Reworked
- Shaelyn's full Three Dots interview
Show Transcript
Note: This conversation, recorded in November 2024, has been edited for clarity.
Shaelyn Otikor-Miller: So we rolled out AI in the start of September. I'm an optimist. I said all thousands of users on the platform will take the required training that allows them to start using the capabilities and we'll just be jamming in a couple of months. I thought it would go so fast.
No, it is being rolled out at the same speed and with the same adoption rate as any other digital technology. I'm seeing the same bell curve.
Siobhan Fagan: You just heard from Shaelyn Otikor Miller from Northern Trust Asset Services. Shaelyn is here to share insights on how her team rolled out Dynamics Copilot and the impact it’s having on employee workflows, client relationships and day-to-day operations.
One important lesson Shaelyn shares is that as much as she thinks she’s building cool technology, it’s as much about emotions as the technology. She’s going to share change management tips and more from her Dynamics Copilot rollout so stay tuned.
Shaelyn, thank you so much for joining us.
Shaelyn: Thank you, Siobhan. It is always a pleasure to talk to you too. I'm so happy to be back and catching up on our digital workplace journey here.
Siobhan: It is a journey indeed. So to start off, could you share a little bit about your role as head of global digital workplace strategy. Give us a lay of the land for your job.
Shaelyn: Yes, of course. As head of global digital workplace strategy and asset servicing, we are focused on the digital workplace tools of the future. So these are the technologies our employees, we call them partners, use to do their job day in and day out, be it the laptop, the headsets, the emails, the Teams chats, etc, etc. So it's all about ensuring that we have a strong interconnectivity model with our clients to ensure that they're getting service quickly, reducing risk, reducing manual processing.
Staying Ahead of the Digital Curve in Financial Services
Siobhan: What I've enjoyed over these years of getting this outside view into your journey and the digital disruption at Northern Trust is that in spite of the technology pace being really kind of out of control and hard to keep up with, you have still been rolling out a lot of really innovative programs there. And I just want to hear a little bit about how you've managed to push those initiatives through. The last time we talked, it was about your low-code initiative and today we're hoping to talk a little bit about your Copilot role out.
Shaelyn: Yes! AI, AI, AI, everyone's talking about AI. I think the last time we talked last year, was the fear AI was going to take over and the robots were coming, right? Well, AI's here, the robots have not taken over. Everyone still has their job. We've learned a lot about AI in this past year, about the development, the training of the models, the accuracy of the AI models, what they can and can't do. So I think the impact of them is much more honed and specific than people thought it was going to be last year, where it was just, "Hey, it's going to take over everything," specifically in financial services.
We're so heavily regulated, to roll out AI at the rate that it's been rolled out in the public, across the world to various individuals or companies — we're just not moving at that same speed. There's risk inherent in any digital technology, any digital platform. And when you're rolling out these capabilities, in the past two years, we've rolled out a brand new client relationship management system. We've rolled out a brand new workflow system to help organize millions, millions of emails we're getting every single day from internal and external parties.
Now we're using AI to say, OK, now that we have this data, we've organized the chaos, we've brought in unstructured data. We've organized it. We're tracking it. We have metrics. Let's use AI to start getting data and insights and analytics out of how our work is going, how we're servicing our clients, kind of our entire global operating model. We're in 20 countries, so it's really hard at times to keep track of every single query, every single email that's going around the world. So the new platform we've been building has now enabled us to do that and have much better data for proactive business decision-making and strategy planning.
The Regulatory Hurdles to a Dynamics Copilot Rollout
Siobhan: There's a lot to unpack in that answer, and there's a couple different directions that I want to go in. But one of the questions that just came to mind, because you are a global company. So are there extra challenges about using the AI across these different borders and aligning with the regulations within each country? How do you manage that?
Shaelyn: Absolutely. So we went live on our first generative AI, Dynamics Copilot. I'm excited to say that.
And it was one of the first formally fully approved generative AIs at the company. So everyone in financial services were playing around with different MVPs and proof of concepts in development environments. We're testing things. We're not using real data. So this was the first one that went through every single risk and review in every 20 country. You could imagine, it was a 12-month process. It was not easy.
We put it through the ringer at our company, especially as a heavily regulated FSI. So we did have to go through all 20 countries to 20 different risk reviews, 20 different compliance reviews, talking to 20 different cloud governance teams about AI regulation in every single country, how they feel about data being used, sensitive data being used, the accuracy of the results, how the results are used, how we're training partners.
It reminds me of that meme of that Greek God who's pushing the rock up the hill. And it keeps rolling down and he keeps pushing it up and it keeps rolling down. That's how the process felt. It felt like the rock was going up the hill and then another risk review, another meeting, another audit, another approval. But through it all, we tried to keep our head, we tried to be patient, just knowing that we were doing the right thing. We had to dot all the I's, cross all the T's, so that when we did turn it on in production and enabled it for thousands of users in the platform, we felt comfortable. Our data is protected. The chat capability is functioning properly and that the way we're using it is not going to increase risk at the company.
The Role of C-Suite Support
Siobhan: I want to touch on the fact that you have a good deal of support from the C-suite for these initiatives that you're doing. With that, you've already got a leg up on a lot of other people. What kind of advice would you give anybody else?
Shaelyn: I will say the C-suite helps. Our CEO, presidents, they were on top of AI, just like some of our junior employees. So it was really remarkable and inspiring for them to come to some of us and say, "Hey, I run into my CEO, I run into the president of asset servicing, and they always ask me, how are we doing with AI? What are we using it for? How are we getting it live?"
And they understand the concept that we have to move fast, we have to start leveraging it where possible to drive results while also considering all the risks and regulations and making sure we're not increasing that. But the support that they gave, driving it, asking questions, they even participated in some of the testing to see what kind of use cases and how they would use it in their own day-to-day lives. So it's been really cool to have them that deep in the weeds at our company and really knowing what's going on. So that makes a difference, hands down for any company.
Overall, I always tell people, when you start rolling out digital workplace technology — the digital technology that is used by people on a day-to-day basis — get out of the PowerPoints. You have to get in the weave. My mentor used to tell me all the power and none of the point. And I talk to my peers in the industry and they're like, "I have to keep creating PowerPoint after PowerPoint."
That was one thing I put my foot down and said, "We will not create PowerPoint presentations to get approvals for digital technology that the people approving will eventually want to use, too." The best way to get the approvals and to assess the full risk is to show them what it does.
And then by showing them, giving them live demos, even letting them participate in some of the testing, it gets their wheels turning a bit more. They want to help because they see the benefit for them and their space as well. It's not an abstract idea that, "Shaelyn's just trying to get this for asset servicing, but legal will be left out." No legal, you'll benefit from it too. It also helps them see what the legal risks are that we might not have thought about if we just saw it on paper.
The Governance Model for Dynamics Copilot
Siobhan: What kind of governance model do you have for Copilot? You spoke about the cloud computing governance models in each of these 20 different countries. Do you have sort of an overarching governance model for Copilot? Can you talk about it a bit?
Shaelyn: Well Northern Trust, like most FSI, we have multi-layers of risk and governance models. So in getting this live, at the ground level, what we would call level one is business unit risk and control and data privacy, enterprise level. And then you have to go per country because you have to go to where the users are. There's employees in Switzerland, in Luxembourg, in Australia that are going to use AI capabilities or even cloud computing and going to send or store sensitive data locally or maybe internationally. There are a lot of rules and laws that are in place now to protect this.
Conference
Sep
16
HR Tech Conference Las Vegas 2025
Register
Conference
Oct
27
Gartner HR Symposium/Xpo Orlando 2025
Register
Just sitting in United States, we're not going to know all of these. So we will go to Australia and first talk to the risk committee that will help us understand what risks we need to mitigate. Talking to the compliance committees to see what type of compliance concerns. We even talked to audit, because after we got all the approvals, we said, let's audit the approvals now and make sure we didn't miss a beat with how we were doing our due diligence, how we were giving the reviews, how we were getting the approvals done, let's make sure we tied the bow around that as well.
It was really a very intricate process. I call it my global world tours where everyone gets to talk to Shaelyn. But like I said, it's long and it can be tedious, but there is a light at the end of the tunnel. And when you finally get it live, my team, everyone's just so excited. Because it also lays the groundwork for the future.
So now all of these individuals, all of these committees have been educated. They've been upskilled on what the newest, coolest AI technology that's coming out, what people want to use, how it works. So now when we come with new use cases, how we want to expand the features and we want to implement more, at least now we have a foundation and we've pulled together various policies and specifics to generative AI that'll help drive growth in the future.
Siobhan: You just said when you get it live, how long has it been live?
Shaelyn: It went live officially in the beginning of September. So almost 12 months after it went into general availability.
A Multi-Touch Approach to AI Adoption
Siobhan: We had talked before and you had talked about how, because of the speed that technology is coming out, that it is the responsibility of individuals to teach themselves. So I'm wondering how you're balancing that this time around with AI.
Shaelyn: Yeah, this time around, I'm doing a mixture of Hi-Fi and Lo-Fi is what I'm calling it.
And I'm going a little Lo-Fi. I am recognizing the emotional toll all of this new technology is starting to take on people. There are individuals who have just flat out said, "Hey, I don't have it in me to learn any new digital capabilities. I'm just not interested. My job doesn't require it. I don't have it in me."
I have people who do need to learn it and they're just so constrained by their daily work that they just haven't plugged in. And again, people just don't have the heart sometimes. So what I'm doing this time around — I'm still a big fan of personal upskilling. It doesn't matter your profession, your region or whatever. If you're not staying relevant in your industry, whatever your industry is, be it financial technology, the tools you're using, then you're just not keeping your career up to date. And I think that's important for most people to just understand.
When it comes to the digital technology, when you have, your Gen Xs, your millennials, Gen Zs who are coming out, they're adapting to it right away. This is the future workforce. They come into the office, they have these AI capabilities on their mobile phone. So not approving them in the office does not stop your employees from using them.
That's one thing people have to understand. People are going to find ways to use the new tools, so would you rather have them do it under approved, risk-reviewed, mitigated circumstances or a kind of cowboy in the wild west letting the horses roam and hoping we get it right? So it's really about working with people to say, who are the super users? Giving them a little bit of guardrails. Telling them go, but here's where you stop. And then pulling some of the slower adopters into smaller group settings. Since the emotional toll is really high, people's defense mechanisms are really high is what I've noticed. I definitely hear a lot of, Well at my age or I'm tired or I'm not that tech savvy. And I always tell people that that's not true. You don't need to be all of that to learn some of these new capabilities, you just have to have an open mind.
It's interesting, the more senior tenured people are usually, and some of the least tech savvy, are usually the ones that start to adopt faster is what I'm starting to see. I think it might be related to my own personal opinion is that when you're more senior, more tenured, you're used to having new things thrown at you, new regulations. We all went through the Dodd-Frank and AIFMD in Europe, new regulations come out, new risk reviews, thick 100-page papers and reports that they have to read through to learn how to remodel their whole business. And because they have that mentality of upskilling naturally to stay relevant and to stay ahead of their business trends, technology is pretty easy for them to adopt to.
What I'm also noticing is when we think about some of the younger generation, who do have it on their mobile phones, who do have it on their apps. They're so app savvy that when they come into the office, it's a little bit different for them because we're more desktop computers and laptops and the configurations are different. The settings are different from just going to your mobile tap, tap, tap, and it's done, right? Siri does it for me. I never have to organize my pictures.
Siobhan: I'm going to have to send you this article based on a study that confirms what you are speculating about, where it's the most senior executives and the most inexperienced are more likely to adopt this thing. Part of it was — to speak to the emotional toll you were talking about — that there's less risk when you're senior because your position is firm, you're allowed to take risks, you're expected to take risks. And then for the junior people, there's nothing to lose. And it's those people in that in-between sandwich part that are really hesitant.
Where Emotions Fit in Digital Technology Adoption
Siobhan: So let's talk emotions.
Shaelyn: Yeah, let's talk. Because that's what this space is about, Siobhan.
It's taken me six years, and the last year or two really drove it home for me. I started this journey thinking I'm just building cool new capabilities, I'm going to roll it out, and people are going to adopt to it, and they're going to love it. And you do have that.
But what I've noticed, no matter what we build, what we roll out, even AI, it's the same bell curve. You still have the super users, the late users and everyone else in the middle. So we rolled out AI in the start of September. I'm an optimist. I said all thousands of users on the platform will take the required training that allows them to start using the capabilities and we'll just be jamming in a couple of months. I thought it would go so fast.
No, it is being rolled out at the same speed and with the same adoption rate as any other digital technology.
I'm seeing the same bell curve. I'm seeing the same pace of change, which are top 25 % adopting right away. Everyone else like, "I'll take the training, eventually I'll get to it." And we're going to sit down with them now and get into the weeds and we're going to handhold and we're going to have small group sessions and office hours to say, this is what this can do to make your day better. This is how this tool can improve your productivity. This is how you can get the answers you need faster.
So we're going to have to do that because like anything else, if you blast it out and communicate, it's really when it comes to sitting down and getting into their safe space, bringing them into a safe space to say, you're not asking dumb questions. There are no dumb questions in digital technology. Ask all the questions you need, raise your hand, and then being in a setting where you realize I'm not the only one. And then everyone feels a little bit better there.
A Buddy System for AI Adoption
Siobhan: When you're doing this handholding, when you're going through these handholding exercises, and you sort of alluded to this earlier, are you doing a certain amount of segmentation where you are breaking people down as to comfort level or what other factors are you looking at when you're breaking people down in these groups?
Shaelyn: Exactly. We've learned that lesson too. We've learned that you have to segment the population. So office hours we've held on various components of the platform. We recognize that they would quickly get dominated by the super users who are like, "Hey, I'm using it and have level 201 questions, 301 more advanced questions." So the people that were at the 101 or double O one, they would sit there quietly just feeling out of place, out of speed, like this is not the place for me and completely checked out.
And I got that feedback, because feedback is key too. You have to talk to people and just say, what's holding you back? Just tell me, is it the features, the functionality, is it time? Let me know so we can work on the best way to fix this. And a lot of people did say, well, I feel out of place. I feel like it's well above my head. So then we started to have office hours where — hey, this one is specifically for the beginners. Super users, we love you, but this isn't the place for you unless you want to teach it.
And then we started Buddy Systems as well, asking the super users, since you guys are so great, pull your group aside, have small group sessions two or three at a time, because I look at it as just like a pyramid scheme. You have to two people, tap two people, tap two people, and over time it spreads, but it does take a while.
The Results of the Dynamics Copilot Rollout to Date
Siobhan: With the results that you've seen so far with the people who are using it, what kind of feedback have you gotten in terms of the people who are actively using it? What kind of results have you seen? Do you have any kind of points there?
Shaelyn: Yeah, so Dynamics Copilot has a lot of features. We're only live with two of them.
One of them is case summarization and the other is knowledge article or ask a question. So with case summarization, it's the concept of an email comes in, it creates a case, it pulls into our workflow, but it could then generate eight, nine other emails if they're working on the case and resolving it. If you're a senior leader, a client exec, and you need to get up to speed quickly, you don't have time to read those nine emails. That was their life before though, right?
You get it, I'm sure you've been there, Siobhan, where you get a three-week-long email chain. You're trying to decipher everything within 10 minutes. So the case summarization feature works great for them. They love it. They absolutely love to be able to look into a case, get up to speed quickly, go and ask more detailed questions or read parts of the emails if they want more answers. So they love that.
The knowledge articles and the ask-go-questions piece, that is such a productivity saver. It is a time saver. It is essentially our own internal self-service, self-support chatbot. We only use it internally, but it is great for peer assist. We were losing time to peer assist, that is, I'm training you and I have to retrain you and remind you and you're confused because you just started and can't find materials and forget what someone said.
Or, hey, the company is so big, I don't know who to send this email to in 20 countries, or I don't know where to go for this, and I spend an hour chasing that down. So now we've loaded all of those FAQs, those trainings and those procedures into the database, and we're training people how to get to those questions quicker. You no longer have to go ask Shaelyn or five people on Shaelyn's team, how do I do this? How do I get this done? Now you can get all that information at your fingertips, and as people are starting to use it, they're blown away. They love it. They said, this is saving me up to an hour or two a day, not searching, chasing and just the wild goose chase of where something is.
How Northern Trust Decided on Where to Start With AI
Siobhan: Copilot has so many different features and all these different bells and whistles. How did you choose to narrow it down to those two use cases to start? Why did you pinpoint those?
Shaelyn: They were the two easiest ones to get through all the risk reviews.
Siobhan: There you go. That is a very, very legitimate response.
Shaelyn: They were the most benign, they were the least risky. We're not using AI to draft anything that goes externally from the company. They're 100% internal and the database is secure. So when it comes to AI, generative AI is all about your data. Your data has to be secured, clean and in very dedicated places. And with Dynamics Copilot, the data sits within Dynamics. We can point the AI at SharePoint sites and intranet sites, but then the results get a little bit more tainted as well. So we love having the clean data set where we control every data that feeds the generative AI prompt and that we can track our data, what's uploaded, how long ago it was uploaded. So the control of the data and then the security, right?
We have concepts where you can upload your training material just for your team, or you can upload a FAQ document for everyone who's servicing clients. Having those security levels with the AI is helpful too because one of the big things with generative AI that slows down lot of regulated companies is the controls over the underlying data. So, some of the broader generative AI, they have access points that would require some massive data cleanup, revisioning of security permissions. So, we started with these because they were the most benign, the easiest.
It also helps with the learning curve. If I would have sat in front of legal and risk and compliance and said, "Hey, I'm using this to build a whole chatbot to respond to every single client query," we would still be talking about it. So it's building blocks. It starts small and now we're able to bring more use cases to the table and on the same model. So we explain the model, we explain the training and the data, and now we just need to go back and we explain the new feature and the new feature and we keep building on it.
Siobhan: I love it. So you're starting off with the training wheels. It absolutely makes sense.
A High-Level Overview of Technology Transformation
Siobhan: I want to ask you a final question before I hand it over to you to cover anything we didn't talk about. You've been with Northern Trust for over 25 years now.
Shaelyn: Mm-hmm, 28 and a half years. I started when I was five.
Siobhan: Exactly! And congratulations! From your perspective now, how have you seen technology transform the workplace there?
Shaelyn: When I started, we were just migrating off of black desktop phones to the phones with the caller ID. We were just going live with the first email.
That was back when email was just the black screen with the purple text. I remember being in high school because I started as a high school intern and my friend was a high school intern at another company. We both got the first email — I forget the name of it — and I remember me and her calling each other to say, "Did you get this at your company? I did too! How do we email each other?" And it took us days. "Did you get it?" "No, I didn't get anything. Did you get it?"
When we finally sent our first email at work, we were just like, my God, this is so cool. We sent emails. and so that was cool. And then we went to Lotus Notes. Then the email became colorful with texts and graphics and GUI. So that was cool and took it to the next level.
I will say this has come a long way. But then I'm thinking to myself how hilarious it is that no matter how long the technology has been around, the same training curves, the emotions are still there. And I sit there with AI sometimes and laugh. Like, I bet you this is how people felt when they were first rolling out email. Those were the people that were like, I'm not using email, I'm just gonna call people all the time and just refused to use it. And now look at this, we've got too much email.
Siobhan: That's such a valuable perspective though, because when I asked about AI, you were like, AI, AI, AI. And it kind of feels like you cannot get away from the topic. And yet at the same time, putting it in this place where it's just the latest iteration. Yeah, it can do some pretty cool things. But we still need to get over these humps. We still need to figure out where to use it.
Shaelyn: You cannot avoid it.
Email came out and it was like, God, I don't want to use it. Then I'll use it. This is great. Then they created Outlook Shared Mailboxes. That was even better. Now we can all share. And now we're a company with thousands of Outlook Shared Mailboxes and millions of emails. And know, my colleague and I, we were joking one time and he brought this up too. He said, "Hey, when Reply All came out, everyone loved the ability to reply all on an email."
So it does make you question what's the next phase we're about to go through? What's the next corner we're painting ourselves into? What I'm noticing is that there's a lot of passion and concerns around AI and where all the technology is going. I think we just need to slow down the mindset a bit and realize it's just another phase. It's another part of the journey.
Final Words of Advice to Digital Workplace Leaders
Siobhan: Is there anything that we didn't cover in this wide-ranging conversation that you wanted to say?
Shaelyn: You know, two years ago, cloud was a thing. Last year AI is what's coming. Next it's visual AI, chat AI, they say a keyboard-less world, where people go totally keyboard-less and all chat.
Siobhan: Final words of advice?
Shaelyn: I just say for everyone, good luck on your journey. Stay patient. Don't worry. There's a community of us out there and it's going to be OK. And every little win is a win and you stack them up more and more and you'll finally succeed in your battles.
Siobhan: Shaelyn, thank you so much for joining me. Thank you for being a voice of sanity in the AI world. I really appreciate it. We're going to have to make this an annual event.
Shaelyn: Thank you for the invite. It is great to catch up with you again.
Siobhan: If you have a suggestion or a topic for a future conversation, I'm all ears. Please drop me a line at [email protected]. Additionally, if you liked what you heard, please share Get Reworked with anyone you think might benefit from these types of conversations — word of mouth marketing is the best marketing anyone could ask for. You can find more coverage of related topics on reworked.co. Thank you again for exploring the revolution of work with me, and I'll see you next time.
