Workday Podcast: How Bill Gosling Outsourcing Set Itself Up for AI-Powered Growth

In the process of digital transformation at Bill Gosling Outsourcing, Joe Fanutti went from the company’s CFO to its chief integration officer—while embracing technology that provided better insights, helped improve efficiency, and enabled future growth.

Audio also available on Apple Podcasts and Spotify.

Putting innovation into practice on an organization-wide basis is easier said than done. Yet Bill Gosling Outsourcing has leaned into the concept by using artificial intelligence, generative AI, and other technologies to help predict where investments will make a positive impact on its business.

At the forefront of that innovation culture is Joe Fanutti. In a conversation following a recent Workday prompt-a-thon event in which finance practitioners were encouraged to experiment with generative AI, Fanutti spoke about his journey from CFO at Bill Gosling Outsourcing to its chief integration officer.

Fanutti spoke about the changing role of the CFO, how finance leaders can embrace technology to create value, and how his organization has already used AI to improve efficiency, predict future needs, and drive growth.

Below are a few highlights from the conversation, edited for clarity. Be sure to follow us wherever you listen to your favorite podcasts, and remember you can browse our entire podcast catalog.

  • “Traditionally, how finance has driven value is through record, reconcile, and report. Really, moving finance to the path of planning, predicting, and proposing [is] a very, very different role for finance.”

  • “The big thing is active oversight on what the AI and ML automation is doing—spending time understanding the datasets, ensuring cohesiveness and completeness of the underlying dataset. I think the oversight piece is critical. You don’t want to be blindsided by AI, so a strong governance model on data as well as AI applicability is important.”

  • “AI usage is going to proliferate throughout all functional areas. It’s going to be backed by really, really strong data governance. I think what you’re going to see is citizen data scientists really engaging with AI, creating their own models. You’re going to have prompt specialists driving their own specific use cases.”

Bruno Navarro: Ask almost any company’s executives, and they’ll say they value organizational agility and the ability to make fast, accurate, data-driven decisions. But how many of them actively seek digital transformation and invest in the technology tools to ensure their success? And what role does artificial intelligence, as well as generative AI, play? Today on the Workday Podcast, we'll be speaking with Joe Fanutti, chief integration officer at Bill Gosling Outsourcing, a company focused on outsourcing, accounts receivable management, and customer service since 1955. Welcome.

Joe Fanutti: Thanks for having me here today, Bruno.

Navarro: Tell us about Bill Gosling Outsourcing and your role there, Joe. I'm interested in your journey from CFO to chief integration officer. How did that happen?

Fanutti: Bill Gosling is a contact center solutions company, and we focus on customer care outsourcing, receivables management. We provide call center programs for various clients in various sectors that focus on the interaction between our clients and their end customers. So a little bit about the chief integration officer role. Really, the mandate of the role is to focus on combining people, processes, technologies to drive enterprise growth, optimizing efficiency, continue to evaluate the business model, and really support inorganic growth as well. So that’s really the chief integration officer role.

On my move from chief financial officer to chief integration officer, we really focused on building out a strategic capabilities group whose pillars of the business were to focus on the business model, the technology surrounding it, process transformation, building a cross-functional team whose mandate really was to futureproof the business for growth and protect it from threats. So the thinking was you can’t effectively do your day-to-day job of running the business and be an effective strategic partner, and really, to separate those two elements, unless you're Elon Musk, which I’m not. So that’s really where that came from.

Navarro: It sounds like you have your hands full.

Fanutti: It’s been a really interesting journey for sure, migrating from finance to, really, a job that touches all elements of the business, whether it be operations, HR, IT, Finance. It’s been pretty interesting deep-diving into those areas and pulling out initiatives that can drive real value.

Navarro: I’ve been reading about CFOs and their changing mandates and how they’re moving more from scorekeepers to value creators. It sounds like you’re definitely in that camp of becoming a strategic partner to the business.

Fanutti: Yes, and the way I view myself is I’m the business partner to the CEO and the president, really advising them on where does the company need to develop organizationally, what capabilities do we need to build, what technology do we need to rely on, what kind of risks do we need to watch out for? And it’s not just me. It’s a group of us that do this. So that’s the focus for the organization.

Navarro: In reading about Bill Gosling Outsourcing, it became clear to me that the company is focused on innovation. Could you talk to me through the process of digital transformation of your financial management system?

Fanutti: Innovation is really core to Bill Gosling. It’s part of its ethos. The core to our business is doing something well that our clients don’t do well. That’s why they outsource in the first place. And as technology advances, service delivery via people diminishes. Within Bill Gosling, the approach is to figure out how to continually embed technology and service delivery before your clients come to you and tell you that service delivery via people is being removed and replaced by technology. So that innovation is really core to how we operate. Finance really needs to support this principle. Digital transformation stems from the role of finance driving operational and strategic value through business partnering. Traditionally, how finance has driven value is through record, reconcile, and report. Really, moving finance to the path of planning, predicting, and proposing [is] a very, very different role for finance.

Navarro: You presented to your peers on Workday's Financial Product Advisory Council on two use cases on how you're using Adaptive Planning to accelerate Bill Gosling's growth, using it to evaluate the financial health of M&A targets and workforce planning. What effect has Adaptive Planning had on your processes?

Fanutti: Before Adaptive, we had another planning platform integrated with Workday, not quite as extensible as Adaptive, and we combined that with reliance on Excel. We had a cohesive traditional financial planning process and methodology, but there was gaps where and what we could do with that platform. So those platform gaps really couldn't get us where we needed to go. And what we wanted to build we didn't want to build out in Excel. So we ended up going with Adaptive.

Navarro: Let’s talk about the first use case around M&A. When did you start looking at M&A as a growth driver for BGO?

Fanutti: Bill Gosling went through a change of control transaction late 2020, and we were acquired by a U.S.-based private-equity group. And one of the things the U.S.-based private-equity group wanted to do was take the investment that previous BGO ownership had done in capability and platforms and use that to accelerate growth organically, so that's service lines and geographies, and inorganically. That would be M&A, so mergers and acquisitions. So M&A as a growth pillar came to the forefront. Previous to Adaptive, financial health of a target in an M&A transaction was a combination of quality of earnings, ad hoc Excel-based modeling, and digging into the diligence room: investigating, going through contracts, looking at documents, so not exactly an efficient process, not a scalable process. And the quality of earnings process during mergers and acquisitions is really just a validation. “Are the numbers good, right?” Not necessarily, “How good are the numbers?”

So how good are the numbers? It really is indicative of what is the strength of the forward-looking business plan. Adaptive has proven planning models. We’ve built out our own Adaptive models. We’ve proven to ourselves that they work in our own environment, so they’re valid, right? When you’re looking at acquiring a same or similar business, why not leverage the models you've already built and understand about financial health, and take those models and deploy them to the target? The great thing about the process is when you walk your target’s management through it, they agree. Target management looks at the models. They’re valid because you've put yourself through them, so the biggest exercise is then mapping the data from the target to your Adaptive models and then going through the results.

The end result is you’ve got a financial health assessment of the target. And that financial health really gives insights in terms of what you’re really buying as an organization. So for us, in a lot of cases, you saw results that you wouldn’t otherwise see in a quality of earnings or just ad hoc diligence. This systemic way of approaching M&A was really, really productive and really efficient for us. You could have probably ended up at the same result—maybe, maybe not. But when you’re talking about billions, tens of millions, “maybe” really doesn’t cut it. So for us, the process of running our target through Adaptive drove some key insights in terms of our decision on whether or not to complete the acquisition.

Navarro: AI was a big topic at the FPAC meeting. How could you see AI helping streamline some manual tasks within Adaptive Planning or plan and track your workforce for purposes of revenue forecasting?

Fanutti: Where we could leverage AI, in the Adaptive world, on our M&A use case would be roster alignment. Say you wanted to understand support cost structure, and you wanted to take a look at your target and benchmark it against your own. You could use AI to do roster alignment, positions, job profiles between the target and yourself. Because you might have a manager of a help desk in one organization. Meanwhile, it’s a director of computer services in another organization. So that roster alignment helps you create the staffing model so that you can compare. That’s really basic.

On the workforce planning front, you could use AI to generate and populate attrition assumptions. That’s a key driver in determining staffing. So use AI to create those attrition assumptions and actually populate the model. And that has, again, direct impact on the staffing outlook. So that’s where you could use AI, really, really targeted use cases, and you could drive some real value there and remove a lot of the manual work.

Navarro: How do you balance the advantage of using AI and ML automation with keeping humans as the final decision-makers?

Fanutti: The big thing is active oversight on what the AI and ML automation is doing—spending time understanding the data sets, ensuring cohesiveness and completeness of the underlying data set. I think the oversight piece is critical. You don’t want to be blindsided by AI, so a strong governance model on data as well as AI applicability is important. Constantly testing the validity of that model, periodically evaluating its relevance, its applicability, its robustness. AI models evolve. And just as humans, we go through performance reviews, you want to put your AI models through performance reviews. It’s no different, right? And I guess the final thing is, from a risk-management perspective, putting in circuit breakers. Depending on the use case, you want to make sure you’ve got a circuit breaker in place so that when it trips, the AI model is not doing more harm than good.

Navarro: That makes sense. It sounds like you’ve thought about responsible AI usage. Going forward, what does that look like in your organization, or what do you imagine that’s going to look like?

Fanutti: AI usage is going to proliferate throughout all functional areas. It’s going to be backed by really, really strong data governance. I think what you’re going to see is citizen data scientists really engaging with AI, creating their own models. You’re going to have prompt specialists driving their own specific use cases. So I think it’s going to be grassroots. But you need to have a governance structure in place to make sure it doesn’t go rogue, and that’s always the risk. And the other thing is to make sure it’s scalable, right? Scalable, that it doesn’t reside with individuals, becomes a corporate strategy so that you can manage it as a strategy.

Navarro: You participated today in the prompt-a-thon on how to curate the right AI use cases and prompts to deliver better business outcomes. How did today’s session impact your comfort level with generative AI?

Fanutti: Today was pretty interesting. I think we generated a pretty robust prompt set, but we didn’t layer in the level of creativity that was expected. So I think it drove the right outcome, so I was happy with it. But the thing about prompts, the thing about AI is understanding exactly what you’re looking for and iterating, right? So what I find for me personally is that the more you iterate, the more you learn, the more you understand how the models work, and then it just feeds on itself. So as a human, the more you engage with the end technology, the better, because you begin to understand it and interpret it a lot better. For finance teams, I really think what you could do is not just introduce them to AI but have them come up with their own use cases, have them say, “Hey, put together a use case, no matter how simple, and deliver it to the business.” Say, “OK, I’m going to do a very simple task using AI.” That gets them on the journey, right? The key thing is getting people on the journey because if they’re not on the journey, then it becomes irrelevant, right? They become irrelevant, so you want to make sure they’re on the journey.

Navarro: It sounds like you’re already thinking about skills for finance teams and sort of how they’re going to align with future uses of AI that haven't even been implemented or maybe even conceptualized yet. How do you think about hiring the right people with the right skills or training people with the right skills for that future that hasn’t been written yet?

Fanutti: So, two different things: When you’re hiring someone, you’re looking at technical aptitude, ability to think creatively, ability to think about how to drive business value, so there’s the hard knowledge. There’s the accounting knowledge, but those other elements are just as important, if not more important, because if you don't have that technological aptitude to say, “Hey, I’m going to go out, explore the technology, embrace it, and then drive it forward,” then you won’t be able to drive value. The development side is then creating an environment where you can disseminate knowledge, have champions within the organization who can drive home the understanding of the technology so that the people can embrace it, right? Deliver knowledge to your team members in a way they can understand and they can digest. A lot of times, you’re sitting there learning about AI, but you can’t digest it because it’s either too much, it’s too complex. It’s really taking those building blocks and building them out in a way that someone who can walk in off the street who’s technically savvy, can digest all the different pieces of AI and ML, and then enter into simple use cases, and then start delivering. So I think that's the environment you want to create.

Navarro: What sort of effect did today’s session have on your idea behind prompt generation or the way you think about prompts? How did today’s session affect it, if at all?

Fanutti: Traditionally, how I’ve done prompt generation is iterative. You start with the basic prompt and you expand it out. Today’s session really made you think about all the different dimensions of a prompt that you need to have top of mind, right? So prompt generation is more of a, I guess-- traditionally, I guess it's been more of an art, but it's really a science that, once you've got the formula down, you can really start to expand what you can do with prompts if you understand that formula. So today, I think we got a good introduction on how you put together the formula to generate a good prompt.

Navarro: What are some of the best practices you learned today about how to curate those prompts and how to deliver better outputs and guidance from large language models?

Fanutti: Today, we really learned about how to be specific on the different dimensions of your prompt, how to provide context, and really, how to make the prompt cohesive. One of the things we did with our prompt is, in some areas, we were very specific. In other areas, we were very generic. And then the prompt didn’t quite deliver what we expected it to. In retrospect, I would have probably expanded more one of the areas and maybe narrowed some other areas a little bit more and make the prompt a bit better rounded. 

Navarro: Let’s go back to the second use case around workforce planning. As a call center, you have a ton of turnover, sometimes 5,000 people in a single year, if I understand correctly. What challenges did that present for BGO, and how has Adaptive Planning addressed those challenges?

Fanutti: The key thing within our business is making sure you’ve got the right people, in the right seats, at the right time. So we manage multiple programs and multiple geos. And not having the right people in the right seats at the right time really impacts margin. There’s significant margin impacts because what ends up happening is you’re either missing out on revenue because you don’t have somebody in a seat, or you’re taking on additional cost because that person’s not billable. Looking at client service delivery forecasts from operations, looking at your existing staffing and related attrition, looking at your training program, and then looking at what existing recruiting activities you have in place, all those pieces form up what your headcount is forecasted to be.

What ends up happening is there’s usually a gap, right? That gap generates the additional recruitment effort you need to go out and do to staff the program. What the model does is brings all that together and say, “OK, what am I committed to my clients? What does my staffing look like? What are my current recruitment efforts? What does my training plan look like? What's my gap?” It brings it all together so that everybody, so all the stakeholders, which includes operations, HR, training, recruitment, they all understand exactly what we need to do as an organization. So you've got stakeholder alignment there, and then you’ve got a way of generating the recruitment effort that's required to meet your client service requirements. As a model, someone can go in and say, “Hey, where are we short on staffing? Where are we overstaffed? Where do we have gaps? Where do we need to put additional recruitment effort?” That overall model brings it all together and brings visibility to everyone in the organization.

This activity happens within Bill Gosling, but it happens ad hoc. It’s not systemic. It happens on spreadsheets and emails and so forth, so we’ve brought it all together within Adaptive. And then we’ve agreed on the framework on how we do the calculation, so everybody agrees on the calculation in terms of how it’s done. So you’ve got partnering happening throughout the organization that way. And then ultimately, like I said, what it delivers is, “Hey, what does the recruitment team need to do so that we can meet client obligations?”

On the workforce planning side, what we did was we built out a model, right, to do the calculation. Ultimately, the calculation is staffing gaps. How many people do you have to go and get? So think of it just like revenue, right? How much sales do you have to go and get? It’s no different in our business. What we did, though, is we used those gaps to generate job requisitions in Workday, so it actually pushes back to Workday. So what happens is you’ve got operations coming in saying, “Hey, this is what my client needs in terms of staffing.” You've got HR coming in and saying, “Hey, this is what your staffing is today. Oh, and here’s your forecasted attrition, so this is what your staffing’s going to be next week.” Then you’ve got training coming in and saying, “Hey, by the way, out of those staff, these are in training. You can’t bill.” Then you”ve got recruitment coming in and saying, “Hey, these are my job reqs out there looking for your staff.” And in the end, you’ve got a number, and that number represents, “Hey, what am I short in terms of people?” As our business grows—right, we continue to grow—you’re going to always be short people. Being short people represents a gap in revenue, so then that drives the recruitment effort that you need to have to fill the gaps. That’s how it all comes together.

And then those gaps drive job requisitions in Workday, right? And then you do that process every week, so it’s a recurring weekly process. And it works really, really well because all this stuff was happening: emails, spreadsheets, offline discussions, regional, cost center, client. It wasn’t happening cohesively. So now, we’re doing it all together as a group, same rules, same way, same principles, same view, which you didn't have before. So that’s the power of putting it in Adaptive.

From an efficiency perspective, you don’t have sidebar conversations happening. You don’t have offline meetings happening. You don’t have panicked recruitment calls happening, so you don’t have any of that. You’ve agreed-upon principles. You’ve agreed upon the model. You have a systemic way of doing it. You can scale it, right? You couldn't scale your behavior before. And you’ve got ways of working. So you’ve got defined ways of working on how the groups all interact, what data they’re responsible for, what they’re accountable to, what the RACI chart looks like, so that’s all built out. That’s what we’ve put together.

Navarro: Well, Joe, thank you so much for joining us today. I enjoyed our conversation.

Fanutti: Bruno, it was a pleasure being here today. Really enjoyed it.

Navarro: We’ve been speaking with Joe Fanutti of Bill Gosling Outsourcing. Remember to follow us wherever you listen to your favorite podcasts. And remember, you can find our entire catalog at workday.com/podcasts. I’m your host, Bruno Navarro, and I hope you have a great workday.

More Reading