Event Replay: Enterprise AI Adoption: Moving from Experimentation to Transformation

[00:04:58] Ronnie Chatterji: We're doing a lot of AI work, but we're not seeing the impact. We don't see the ROI. And I would posit that many of those companies are caught in the micro productivity trap. What they really need to do is figure out how to elevate and then really decide where to focus original attention. We're going to talk more, Arjun, about how to overcome that micro productivity trap. But I'll just say, when I talk to executives about our article and speak to our own enterprise customers at OpenAI, this is the evolution that a lot of leaders are trying to push for. They see AI being used by their teams to do amazing things, particularly agentic tools now, but how does it aggregate up into a business level outcome? That is a leadership question. What I find so interesting about this is we're sort of going back to basics, which is focus and prioritize. As you know, Arjun and Gene, I taught at Duke University's Fuqua School of Business for a long time. The first thing I would say to my students is focus and prioritization are the keys to strategy. Arjun's also a Fuqua grad. I'm contractually obligated to mention that. I think it's amazing, go Duke! But more seriously, when you think about strategy, a lot of this is sort of the old playbook of change management and leadership.

[00:06:08] Ronnie Chatterji: Gene, why is that prioritization difficult? Any hacks to help executives think about where AI can create the clearest business value? Because I think a lot of people hear this and they say it's easy to say, maybe harder to do. I know you've been helping your clients do it for a long time. How do you think about helping executives focus and prioritize?

[00:06:29] Gene Rapoport (Bain): It's difficult today because today you're chasing shiny objects, you're chasing adoption, you want to go deploy a bunch of tools, and you want the frontline to lead that. They naturally come up with a bunch of ideas. But in reality, what's going to be most transformational is if you think back on what is strategic for the business absent AI—what are the things that we're trying to achieve? How do we then use AI as a lever to achieve those things? Now, for some companies, AI itself will influence the strategic agenda, but for many others, you could actually use AI as a lever. That’s where I would first start: what are the three to five lighthouse areas where we need to deploy this technology? You might have a hundred use cases that support those lighthouses, but typically companies get that backwards. They start with the narrow list and then they try to work their way up to the lighthouses.

[00:07:23] Ronnie Chatterji: Yeah, I saw this with OpenAI's own work with Lowe's and had a chance to talk to their board recently. I mean, they really try to tie their AI implementation into where Lowe's as an organization creates value. I thought that was a really interesting way to just start from first principles, and it is exciting to use all these different tools, but sometimes you need to take a step back and think what is our value proposition and where can AI, using your words, Gene, be a lever? That's what I saw Lowe's doing. They were thinking, okay, let's have a clear criteria to prioritize the different parts of the business we can implement this in. Let's assess these different sites: technical readiness, how big the opportunity is, and also risk, which is a big part of this as well. What they've done in their stores to maximize key metrics, such as basket size and making sure that consumers are getting information they need about the products in the aisle, has really paid off for them. I think I saw that sort of in action, and we talk about them in the Harvard Business Review article as well.

[00:08:41] Ronnie Chatterji: I think when you think about the companies that are getting the most, it's often those who are redesigning workflows around AI—almost AI native workflows, Arjun—than it is just bolting AI onto existing processes. Again, like we see it, it's real, but it is harder to do in practice. But you are really in the weeds, in the field making this happen. What does it look like to actually build workflows from the ground up around AI rather than bolting it on? How can we help teams spend less time on the repetitive work and more time on the higher-value judgment, delegation, and discretion tasks that people are trained to do?

[00:09:40] Arjun Dutt (Bain): Yeah, it's a really important point, Ronnie. This is also part of the reason why focus matters. To really get value from AI, you have to go deep. You really have to redesign how work is done, as you said, in an AI native way, not just bolting AI onto a current process. Some of our clients call the latter, which is just bolting AI on, as the expression—the popular expressions become paving the goat trails. What is needed is to essentially start from a clean sheet of paper and say, how should this work really be done? What are the process steps that make sense? Now that I have AI as a tool, what should I really accelerate there?

[00:10:35] Arjun Dutt (Bain): Frequently, we find that we are working at the front lines with our client teams, working with the people who actually do this work. They have to be sort of change-oriented and willing to partner with us to relook at the way work is done and say, what are the steps that we no longer need to do? What are the steps that we've been doing just because that's how the work has evolved over time?

[00:09:56] Arjun Dutt (Bain): has evolved over time, may not make sense to continue doing that way. Then we sort of look and see, okay, all the things that we still need to do, what can be taken by an AI tool or an AI agent? You know, when we started talking about this article about a year ago, I don't even really think agents were on most people's radars. But that is a reality today. And being able to re-look at your process and say, let's cut out some things which we no longer need to do, it's just corporate overhead or bureaucracy that sort of developed over time. Let's add AI agents to be able to take some of this work on. And then let's just figure out end to end, how do we maintain quality while we're driving sort of velocity? Is the key here? That work, it's hard work, it's deep work, and it requires really partnering with the front line.

Arjun Dutt (Bain): So we talk about the fabrication co-example in our article. The redesign of their work processes to do faster quotes, respond to customer inquiries, was really done by completely re-examining how the work was done, what steps could be accelerated with AI and what steps really no longer needed to be done. So that's typically the approach that we take. And it's very consistent, whether you're talking about coding process or software development, it's shocking how similar some of the core underlying tenets really are of doing this work.

Ronnie Chatterji: I love that, paving the goat trails. That's what it is, Arjun. I'm not successful in incorporating Gen Z lingo, despite my kids' best efforts into these events. I don't have the riz for it or something like that, but I will try to take this paving the goat trails piece because I think that could be helpful. I don't know if that's going to resonate for the Gen Z, but yeah, that is definitely the lingo that we do quite well.

[00:11:39] Ronnie Chatterji: I love it. I mean, Gene, one thing I was thinking about this when Arjun talked about Fabrication Co and how the employees are moving faster with AI. I know you've got the frontline employees, you have management, you have the board. I think about your work with the PE practice in particular, there's lots of different stakeholders who are going to need to sort of move fast with AI. How do you train folks across the company, get them in the right mindset to adopt this? Particularly when it's changing so fast. And is there a fatigue that's setting in with all the AI trainings, with all the implementation? How do you avoid that when you're trying to get the best performance out of a company? And even from the investor point of view with a PE owner who might be coming into a new organization?

[00:12:21] Gene Rapoport (Bain): So typically we find that the biggest gap that companies have is talent and their understanding of AI, the potential of AI, how it applies to their workflows. We typically find that the average employee will think about AI as a chatbot plus. In reality, we've had reasoning models come out. We've had agenda capabilities come out. There's so much more that you can use the technology for. And organizations are really struggling with how do we get everyone using the technology in the intended manner? How do we get the executive team to be aware of what's possible? Because in turn, their familiarity of what's possible and where it's headed is actually gonna inform what they should do with the technology. This is a major obstacle. We've seen folks develop champions networks, bring in third parties to actually help enable people, celebrate folks who are actually experimenting and share those learnings out. But it is a multi-front war, frankly. There's not one single thing that you could go do to drive understanding within an organization about how to leverage this technology.

[00:13:26] Ronnie Chatterji: Yeah, I see the same thing in a lot of our work where there'd be a set of power users who are all in on AI, they're doing amazing things. You look at the data inside the organization and they're the power users of CodeX, of ChachiBT, what have you. But there's a set of users who aren't really using it as frequently. And the question for corporate leaders is how do I do enablements? How do I do sort of trainings to help those folks learn use cases from the others? A couple things I found is like role specific use cases are really important. I think we've all sat through lots of AI trainings. We've given a couple probably on this call, but when it's actually tailored to your role, what you're doing every day, it's much, much more powerful.

Ronnie Chatterji: And the second thing is when the tools change and new opportunities open up, sometimes you have to go back and you can make things you're doing even better. I have lots of automations that have improved as the technology's improved, but if no one was ever telling me about those things, I kind of keep doing things the old way. I think the other thing is sort of having some support to be asked the question of. I find it really helpful. My team is really expert on a lot of AI applications and being able to ask them is really key. So that cohort effect, it really matters.

[00:14:29] Ronnie Chatterji: I think, Gene, the one thing I want to sort of delve into a little bit is you mentioned talking to people closest to the work, the frontier teams, and working with Lowe's. That was another thing I discovered that they really did a good job of that. And I'd say this is one of the things for people watching that everyone will tell you in business school or in a business book, talk to the customers or talk to people closest to the problem. The challenge, everybody knows that, is actually doing it. And so if you're listening to this and saying, oh yep, I'm gonna do that, yep, I've done that.

[00:14:54] Ronnie Chatterji: I'm going to do that. Yup. I've done that. At least with Lowe's, what they did is they piloted the AI assistance with the store associates first to gather feedback for scaling it. That was a really important step. So it wasn't just saying they were going to talk to the front line. They actually did it and incorporated that feedback. They have real examples of how the associate feedback helped improve prompts or the guardrails or the UX before they deployed it across like the 1700 stores. And so I think that if you can not just tell me that you're talking to the front line, but show me that you're talking to the front line, that's going to be a big difference. Maybe Arjun will ask you about this. Once you do implement this stuff, the question I get all the time is how do you measure the value? Very difficult question. You work across a lot of different businesses. How should we think about measurement of real value for business, for employees, for customers, and if you have a framework for how to think about it or some advice for people out there who are dealing with this in their own organization?

[00:15:49] Arjun Dutt (Bain): Great question, Ronnie. We talk about this in the article. You know, I think usually when folks like Gene and I get involved, most of our clients are talking about productivity and efficiency writ large. And I think the problem with that is that while those are right in a macro sense, those can be very difficult to measure because most companies aren't really instrumented to figure out like, okay, well, what does productivity mean? What exactly does efficiency mean in our context? The way, and I think the approach that you all took at Lowe's is probably very consistent here. We typically focus on three main things. One is just the business value. What is the business outcome that we are looking for? Is it a profitability measure? Is it a revenue growth measure? We ultimately want to boil everything down to specific metrics that we can sort of actually observe over time. The second is around employee experience. One could simply be, you know, do the people using these tools, do they actually like the tools? And then, of course, you can measure, you know, whether they report, whether they get more time back. Are they able to take on other tasks? And the third, and, you know, some might argue, most important metric is the customer outcomes. So are you able to respond to your customers faster? Are you winning, you know, a higher proportion of the deals than you would have historically? And I think these all need to tie back to what the business case was in the first place. We talked about focus, and I think a big part of the focus is, what are the sources of value? And typically our successful clients will then write a business case and say, okay, well, how am I going to go capture that value?

[00:17:31] Arjun Dutt (Bain): At the highest level, you know, we like taking the productivity efficiency kind of framing and really think about it in terms of business, employee, and like customer-facing metrics, and ideally tie these to things that the company already measures today. So you also don't have a huge burden of new telemetry in the organization and use the language that people are comfortable with, used to today, to really get a sense of like, well, are things getting better? And we think that that's a crucial part of one, showing progress and then bringing others along.

Ronnie Chatterji: So what you're saying, I'll go to Gene on this, but like three things I took from what you said. One is like AI is not an excuse not to think carefully about your business and your value proposition. In fact, I might require you to think even more deeply about it because it's a general-purpose technology that can do almost anything, but you still have to figure out what you're trying to maximize so you can run, measure it. And that's something I think people can skip that step sometimes. The second thing is, yeah, building new telemetry is really hard. I love the way you put that. And so people are trying to build an entirely new measurement system. It's both, it adds a lot more bureaucracy and process to the challenge. It might take you away from core metrics that were working well before. And that third piece you mentioned is familiarity in the organization. For a lot of folks, AI is a very new phenomenon. They’re getting used to it. There’s uncertainty about it. And if you’re going to measure a bunch of new things that don’t necessarily make sense, at the same time we’re introducing a new technology, that can create friction. So I just love the trio that you’re talking about here. I think it's really practical.

[00:18:53] Ronnie Chatterji: Gene on Fabrication Co., if you want to talk about that here, how did you choose those metrics? How did you figure out whether AI was creating business value in those cases? I’m happy for Arjun to bump in on these too, but I’d go to Gene on Fabrication Co.

[00:19:06] Gene Rapoport (Bain): Yeah, Arjun should jump in. He was closer to it. No worries, Arjun, go ahead. We'll come back to Gene.

[00:19:09] Arjun Dutt (Bain): With Fabrication Co., there were lots of different use cases that they could have picked, and part of the work was really understanding where the value was going to be. And with their management team, with the front line folks that actually work on serving their customers on a daily basis, we essentially landed on two different processes. One is just sort of responding to customers' RFIs, and the second is generating a quote. And then the metrics really came from there to say, well, what are the most relevant metrics that the business is used to today that should be the ones to inflect here?

[00:19:52] Arjun Dutt (Bain): That should be the ones to inflect here. And so the types of metrics that shook out of that, one would be something like a quote turnaround time. So how much time does it take you to actually get back to a customer from the time that they initially reached out? And this company would measure that. That was a metric that they measured today. And so that was a metric that we wanted to see if you added AI, if you kind of redesigned your process and added AI, how much could you impact that?

The second would be something like a win rate. So for the types of quotes or RFI responses that you're generating with AI, what does your win rate look like relative to the ones where you don't use AI? That's got to be an important measure, right? Because that'll help you sort of figure out, are you responding fast, or are you responding fast with quality? You know, there were other kind of efficiency measures, just how long does it take you to generate a quote. So for FabricationCo, you know, before we used AI as part of the process, it would take an engineer anywhere from four to six hours to generate a quote. Once we redesigned the process, that shrank to 20 minutes. So you know, pretty significant impact there.

As part of the co-creation process around these kind of efforts of redesigning work with AI, injecting AI into how work gets done, you always measure employee NPS. So do people actually like it? So I would say that that's always pretty critical that you should measure no matter what. And then finally, profitability was an important measure here. So what was the gross kind of profit impact? And for FabricationCo, and I think we mentioned this in the article, within about three months, they were on track to a $30 million increase in annual gross profit. Just from tackling, you know, one major function, which was like responding to customer requests. And you know, honestly, tackling two important use cases within that.

So it also just shows you the power of focus. Because when you focus on fewer things, you can obviously dig deeper and do them better. But you know, the other part is you can deliver real results quickly.

[00:22:00] Ronnie Chatterji: That's amazing. That $30 million is a great outcome measure. And when people talk about ROI, these are the kinds of things they should look at. I love the employee NPS as well. Gene, I wanna ask you one more question before we get to the Q&A. We're already getting a lot of audience Q&A, which is awesome. You guys are generating a lot of buzz. But Gene, you have a lot of experience with private equity and investors. And I was thinking about sort of for PE sort of leaders and those who are investing across wide portfolio companies, often across very different sectors. Is the AI playbook sort of similar across multiple sectors that need to be customized? What are some insights you can add from the PE work that you're doing?

Gene Rapoport (Bain): I think it's a really important layer for people to understand here, given the role of private capital in the economy and their role in terms of implementation of AI. So in 23 and 24, a lot of private equity funds thought that they could develop a playbook, roll it out throughout the portfolio, and it would just get executed on. I think for a lot of the reasons that Arjun mentioned earlier, that this is a people transformation, a process transformation that the devil's in the details. You know, those funds started to learn that those playbooks didn't actually work very well at individual companies, that they needed to be tailored. They needed to allow those companies to focus on the few things that matter, and not on deploying a series of playbooks.

Gene Rapoport (Bain): That said, playbooks could be helpful now. They could point companies to the right types of solutions or the right way to architect an agentic system. So I don't mean to say that they're not helpful, but they themselves are not gonna drive a transformation for you.

[00:23:35] Ronnie Chatterji: That makes a lot of sense. For both of you guys, and then I'll go to leave some great questions. Just maybe end us in this part of it with just one or two thoughts—what excites you the most about AI? I think we can work on this stuff so much. And sometimes we get caught in the nitty gritty and the workflows and implementation. I mean, we love this stuff. What's exciting? I mean, you're seeing this interesting arc at such a key time in all of our careers. You're living it in your organization. Maybe Arjun, then Gene, what excites you most about kind of enterprises adopting AI now where this is ongoing?

[00:24:02] Arjun Dutt (Bain): Sure. Well, so I'll tell you what I find inspiring kind of on a human level, and then kind of more, I guess, in an enterprise-oriented way. At a human level, you know, when I see my clients, you know, get measurable results from these, from the use of AI, I find that their level of energy and enthusiasm just skyrockets where they can just see that it's working. And that kind of positive feedback loop to say it’s working, I need to figure out how else I can use this. I find that to be incredibly inspiring and sort of keeps me—it's hard work. And, but, you know, that's really what is the day-to-day motivator on like continuing to do this.

At more of an enterprise, you know, value creation layer, I'm finding that this type of work really...

[00:24:50] Arjun Dutt (Bain): That this type of work really, you know, it doesn't just make people or clients faster or, you know, slightly better. It's entirely redefining how they think about competing in the markets that they work in today. So for Fabrication Company, it was truly like, we can be customer first in a way that no one else in our industry has really ever been. For some of my tech clients who are applying this to software R&D, for example, just the level of customer responsiveness, being able to deliver a feature that works in a record amount of time and, you know, drive that customer delight is just, you know, next level. So I find both of those things pretty inspiring. You know, just helping our clients kind of redefine how they play in their industries and honestly helping many of my client partners just redefine how they do their job on a daily basis. It's pretty inspiring for me.

[00:25:46] Ronnie Chatterji: That's awesome, Arjun. What do you think, Gene? How about for you?

[00:25:48] Gene Rapoport (Bain): Yeah, I'll mention two things. So one of them is I think a lot of organizations have just organizational muck—processes, you know, you have to—there are decision rights, there are ways of doing things. AI deployed successfully can actually begin to break some of that in a very positive way. We're seeing even, you know, at Bain or some of our clients, we can take product development, we can take what used to take six months, collapse that down to two weeks, right? For some stages of developing an initial idea, building a prototype, getting it connected to real data, testing it with the end user before would take a long time and you would have to go through a long process, engage a lot of different people, and now you can actually get the person closest to the frontline building the first versions, and this was unimaginable years ago. So that is just, that's not the, can we go find a 12% productivity gain initiative, it is, can we take something that used to take six months and collapse into two weeks, which is a lot more inspiring for people and organizations. And then the other point is, it's allowing companies who might be the number four in their market to actually say, you know what, we're gonna apply this technology and we're gonna try to leapfrog number three and number two. We're gonna try to go toe to toe with number one because we're gonna apply AI to how do we go to market, how do we respond to RFPs, how we develop products and that is just creating an enthusiasm within those companies that's very real. I love it, the idea of just inventing an entirely new way to work is, and I'm seeing that on my team with Codex right now in terms of we're just working differently than we were even six months ago, which is amazing.

[00:27:20] Ronnie Chatterji: Bunch of audience questions, I'll hit a couple and then thank you guys for the time. Henry Smith in our audience today, Henry asks, you know, look, we want companies to use and practice AI, but it has to be safe and, you know, especially in highly-regulated environments. So how do you guys think about this idea of, you know, advancing AI but doing it safely, and you often run into pitfalls when you move too fast in the other direction? What's your advice for Henry and other people who are thinking like him?

[00:27:50] Gene Rapoport (Bain): So there's an organizational answer and a people answer to that question. So from an organizational standpoint, you must secure the technology, enable folks to use the right technology, and you have to prevent them from using solutions that are not approved. So there's just setting up the right tech infrastructure, ensuring that your cyber team is actually aware of the new vulnerabilities that are introduced by AI or AI-enabled applications, e.g., prompt injection attacks is a very simple one to talk about. So there's an organizational readiness, you know, folks need to step up from that standpoint. And then the average employee must be aware of how their use of AI can introduce risks. You know, how they, what is a markdown file and what is a skill and what if I embed, you know, proprietary company information into the skills that I'm producing, and then is it okay for me to download a skill from the internet? The answer generally is no. So there's just another training and upskilling burden, frankly, on the organization because you need to get people understanding what these vulnerabilities are.

[00:28:56] Ronnie Chatterji: Love it. Let me ask you a question, Arjun, from another one of our questioners, unless you wanna add there. Shipali Bala has a really good question on this. Shipali says, look, I'm all for AI, but we need DRIs, that's how I think about it in my world too, directly responsive individuals. How should companies assign ownership and governance when AI changes an entire workflow end to end, not just individual tasks? And this is, you know, this is really important as we move from the individual to the business level, hard to do. How do you think about that piece? Any thoughts to help Shipali and other people who are thinking like her?

[00:29:26] Arjun Dutt (Bain): Yeah, it’s a really good question. You know, in all of the work that I’ve done on AI to date, there’s always a human in the loop. And, you know, the way people get comfortable with really scaling up their AI initiatives is to actually start in a very methodical way.

[00:29:48] Arjun Dutt (Bain): In a very methodical way and build trust in the AI system. And make sure that there's a human layer of governance. They're all of the things that Gene was talking about, the security reviews. So most of my clients are putting real guardrails in place with like clear human in the loop requirements to say, okay, how do we make sure that we're providing intelligent answers to our front line or providing intelligent responses to our customers? It doesn't matter if it's artificial or not, but there's usually a human layer that's always checking.

I will also say it's often surprising to people how well AI systems perform at scale. And so we've literally done side-by-side experiments where AI generated responses to customer questions at scale over time really start to outperform humans. And then what that allows human workers to do is to really focus on the job which would be something like customer care. How do we make sure that this customer is really, all of their needs are met? I can rely on AI to answer some of the nitty gritty questions based on kind of our knowledge repository and knowledge base, but my job is to make sure that this customer is happy and satisfied.

And so I also think, I find that that's a nice sort of partnership between the algorithms and the people, where the algorithms can speed up a task, but the people are still doing the very important work. Their jobs on a day-to-day basis. So probably standing a bit beyond kind of what the question was, but I think that that's an important distinction for us all to keep in mind.

[00:31:26] Ronnie Chatterji: No, I think it's great. I mean, we need to think about these bigger pictures in the organization. All of these questions reflect people who are doing it in their organizations. And so these questions are sort of like, okay, I agree with what you all are saying. How do I implement? So this is really helpful.

I'll give you one more, Gene, here, from Lana Romanova. And great question, very simple. Should companies appoint dedicated AI owners? And I'll maybe make it a little slicier. Should every company have a chief AI officer? We know you should definitely have a chief economist. That's just table stakes nowadays. You have to have a chief economist. But what about chief AI or a dedicated owner of AI, pros and cons? I'm sure you've thought a lot about this, both of you in the Bain context. Happy to hear from Gene and Arjun if he wants to jump in. But Gene, what do you think?

[00:32:06] Gene Rapoport (Bain): As the chief AI officer of our Private Equity Group, I have strong opinions here. I think it's important to appoint someone, but I will tell you it's also important to give people authority. If they just have a title, then they end up being a mascot, and that's not very helpful. So sometimes you do not necessarily need a chief AI officer type of title, but you need someone in the organization who is pushing this forward.

I will tell you, the companies that we were most impressed by are companies where the CEO is playing that role as well. It needs to be the executive team. This can't be a side project. This is strategic for most organizations, and it needs to be top-down.

[00:32:50] Ronnie Chatterji: That's right. Arjun, what do you think?

[00:32:51] Arjun Dutt (Bain): I’ve forgotten about the Chief AI Officer title, so it’s like asking me about chief economists. What do you think?

[00:32:57] Arjun Dutt (Bain): Yeah, I'll sort of off-point because I fully agree with him. So, injection work. We see that it's critical that the business needs to be a general manager of who's in the line business, who's business unit leader. Those are the ones that are going to be impacted.

[00:33:15] Ronnie Chatterji: We might have lost Arjun real quick, but Gene, you can hear me okay?

[00:33:17] Gene Rapoport (Bain): Yeah, I think I'll just pick up where he was leaving that off. On the thought partnership, it is important for an organization to have some folks internally, senior enough, who can actually inspire, where is this going? What does it mean? How to get companies unstuck from the thinking of AI as a chatbot, and what are the true use cases?

How is this gonna impact our industry? How can we deploy this technology in a more sophisticated manner? Folks who can look around the corner. So that thought partnership must be present. You know, I’m not necessarily saying that needs to be the CEO. There needs to be folks in the organization who are capable and willing to carry that.

[00:34:04] Ronnie Chatterji: So interesting, both Gene and Arjun, your comments. Arjun's back with us too, and Arjun, the only thing you missed was just Gene basically kind of agreeing with you on that point, but also saying, look, the CEO, that C-suite has a really important role. I'm interested in terms of the CEO delegating it to someone who's the AI person versus, you know, Gene, your other point earlier of like the CEO needs to be really vested in this and more than just a mascot when it comes to AI, that's a really provocative question, right? Because CEOs are busy, they have a lot of stuff going on, but this is fundamental in creating business value. It's not necessarily someone else's job either. So that's something I think we should definitely delve into.

Maybe we have a future article in HBR up our sleeve, which is, is your CEO also your chief AI officer? I will ask you to, I guess, have two more, time for two more. There's good questions here, but I will do two more.

[00:34:46] Ronnie Chatterji: There's good questions here, but I will do two more, and then I'll wrap us up here on time. And I really appreciate you guys' time. It's been fantastic. Maria Patapuchek has a great question about how you measure AI readiness. And I take this question from Maria to mean, you know, we talked about implementation, but how do you know if your organization is actually ready for it? Do you think about it in that way? Are there KPIs that you should track when you're thinking about AI readiness? And good question from Maria. Arjun, go ahead, and if we lose you, I'll let you know, but I think you look back in, saddle here.

[00:35:18] Arjun Dutt (Bain): Yeah, AI readiness. I mean, the truth is, like, you know, when you talk about AI at scale, no one feels ready. But the truth is most companies are ready. And I think if you take this approach, and like, you know, with a very focused sort of lens on like where can I create value and dig deep, you essentially figure out where the organization is ready or not ready, and you solve those problems one at a time. So that's kind of the very pragmatic, hands-on approach that we've taken. We do see that, you know, people will run into challenges around data or governance or not having enough guardrails. But, you know, this is an important enough priority where you can't wait to let all of those pieces fall into place before you say that we're ready. The only way to be ready is to get started and then figure out what is not ready? How do we make it ready? And so that's my very sort of action-oriented way of approaching this with clients. And, you know, quite frankly, this is the approach that we see working. Otherwise there's a million reasons not to do something.

[00:36:17] Ronnie Chatterji: I love it, I'm gonna use that. I love the reason to do it. I'm gonna use that again. The only way to get ready is to get started. I love it. Arjun is very quotable here. I'm using a bunch of his things from now on. Gene, real quick, one last question from Andra Jermunsen and then we'll wrap up. Andra has a good question. This is something I didn't think about that really matters. Most organizations are organized in silos. What we're all talking about in our article is something that's fundamentally cross-functional. So in your work, have you seen particular organizational structures or do you need someone who's connected the dots to make this happen? Because for a lot of us, our organizations don't look like it's very easy to always work cross-functionally on these AI projects. How do you think about that?

[00:37:51] Gene Rapoport (Bain): That's one of the benefits of having that point person who's gonna oversee the implementation of AI is they can bring these different forums together, these different functions together. And ideally, you wanna be celebrating within each function what they're doing, but then linking that to what other folks are doing throughout the organization as well. I'll tell you a lot of the common obstacles with AI as it relates to the technology itself, they are not exclusive to finance or sales or software development. They can carry across those functions and you don't need to relearn those lessons painfully in every single function. That said, it's actually helpful to have some folks experiment in places and then pick from the best.

[00:38:30] Ronnie Chatterji: Oh, this is amazing. First, I just wanna thank Gene Rapoport, Arjun Dutt from Bain & Company. This has been amazing. What I love about my job is the ability to work with people like you, but also to learn from you. And both the article in Harvard Business Review and this experience have been fantastic. I wanna thank you guys for joining us and also thank you to everyone who tuned in today, the folks who I got to ask their questions and the folks who I didn't. Thank you so much for your contributions to OpenAI Forum. Before we wrap, we'd love to see you at another OpenAI Forum event next Tuesday, June 9th. We're gonna keep going, building an AI-powered workforce at Brooklyn Sports & Entertainment. Be an exciting one, I'll be there. The team is dropping a link into the chat now. Until next Tuesday, thanks again everyone. See you soon.