Unlocking high-impact AI implementations in any enterprise
Businesses need to scale AI to remain relevant, but the first steps of that journey can feel daunting. Yet some companies chose to be AI trailblazers, navigating beyond proof of concept to achieve measurable results, carving a path for other firms to follow.
We sat down with three such innovators from Team Liquid, Bosch and Bristol Myers Squibb to hear their story and learn their guiding principles for getting started with AI: identify a core business problem or opportunity, collaborate with stakeholders across your business to define a precise AI solution that addresses that problem or opportunity and look beyond traditional metrics when measuring success.
| Company | Industry | Company size | Headquarters |
| Bosch | High tech | 400,000+ | Gerlingen, Germany. |
| Bristol Myers Squibb | Life sciences and healthcare | 34,000+ | Princeton, New Jersey, USA |
| Team Liquid | Media, sports and entertainment | 300+ | Los Angeles, California, USA |
Reinvent
business processes with Al at the core
Partner
with IT, business teams and technology partners
Simplify
and standardise the enterprise data landscape
Finding the right function to maximise AI impact
It is notable that each of our experts began their AI journeys at the whiteboard. By examining their businesses to identify a core success lever, they focused the conversation on specific problems or opportunities that would accelerate business outcomes.
As Head of ERP Product Excellence at Bosch, Gamze Öztürk began with a review of existing processes to determine gaps and transformation opportunities. “We always start from the real business pain, from the business process challenges that we might be having,” she says. “With over 400,000 employees and over 2000 SAP systems, friction within any of Bosch’s processes has far-reaching effects. Carefully identifying where essential processes may be breaking down results in a more thoughtful application of AI and automation solutions. That’s been a key learning. We take all these business processes into focus before jumping to the technology because the important thing is not only the technology, but the context behind it.”
Kenneth Manalang, SAP Enterprise Architect at Bristol Myers Squibb, also investigates opportunities to transform processes to enable teams to better serve the company’s mission.
“Bristol Myers Squibb is a global pharmaceutical company that discovers, develops and delivers medicines to preserve and prolong the quality of our patients’ lives,” he says. “We cater to our patients, meaning we need to be defining and creating AI use cases that are the most valuable to them, that help accelerate and broaden our scope to provide more and more medicines to those who need it most.”
The guiding principle that critical business processes provide the greatest opportunity for AI transformation applies to every industry. E-sports powerhouse Team Liquid’s laser-like focus on helping their players remain competitive and win matches is among the reasons they are a global leader in their field. Co-CEO and Owner, Steve Arhancet says, “We’re a sports team, so we have a responsibility to compete and win. We have an enormous amount of data that is collected from all our games. We needed to figure out a use case where we could actually use this data and it was improving the competitiveness of our rosters and teams.”
Collaborating with business teams and technology partners
After identifying a problem or opportunity, zooming in on the ideal AI solution requires input from a cross-functional team. For example, for Team Liquid, the right tool to enhance their competitive edge was informed by the needs of their players. Team Liquid’s ban/pick draft tool utilises AI to analyse the abundance of first- and third-party data from previous games to suggest optimal character picks and bans for upcoming games.
“There’s a lot of strategy associated, so we thought, let’s build something AI-powered in order to give the optimal choice associated with this,” Arhancet says. “[Our coaches] want to be able to go into like a ChatGPT-like interface and ask questions [and] instantaneously get those answers. So, we started with data, organised it and then we layered a custom-built Joule agent on top of that so that it could interface with nontechnical users.”
Organisations are better positioned to articulate AI business cases that deliver value when people who will use the solution and technology providers are in an ongoing conversation. At Bosch, Öztürk’s team is “frequently doing ideation workshops together [with] SAP” and their business teams. “We are getting feedback directly from the business and then, altogether, we are moving to next steps to make it happen.”
The combination of users’ procedural knowledge with providers’ specialised expertise also helps avoid wasted resources and one-off pilots that don’t scale. Maintaining this type of partnership is also a key objective in Manalang’s work as an architect at Bristol Myers Squibb.
“Bristol Myers Squibb is not an IT software company,” he says. “We are a company who manufactures and develops medicine. This is where the partnership comes in. We engage with our business users and ask for help from our partners from SAP to identify the problem statement and how we can help them to accelerate their work. That is the foundation: to understand what it is we’re building so we’re not unnecessarily building and activating agents everywhere.”
Measuring the true value of your AI investments
Measurable impact comes from partnering across technical and business teams to pinpoint where AI will matter most, then executing against those scenarios. But the collaboration shouldn’t stop there. It is important to include diverse stakeholders when defining KPIs, since typical metrics may not capture the intangible value of the implementation.
“At Bosch, we are also sometimes calling it a return on intelligence,” Öztürk says, pointing out that classical ways of using ROI calculations don’t always make sense or capture the value it will bring in most cases. “It's really tricky to calculate the value behind AI. So, we ask, are business processes more accurate or are we making faster decisions and examine the indirect value behind the AI solution?”
There is evidence of this broad return of value across the AI solutions they’ve deployed. At Bosch, physical AI streamlines complex warehouse management workflows. A pick and place solution integrated into their warehouse environment helps alleviate the manual burdens on the plant floor. A comprehensive SAP system creates a connective loop involving warehouse robotics and ERP systems. SAP Joule triggers the task, the robots execute and send return signals back through SAP to close the loop.
Bristol Myers Squibb extended its mission-driven approach into administrative and fulfilment scenarios to support the processes underpinning their programmes without sacrificing the human involvement required for industry compliance. An SAP and Microsoft Teams integration drives invoice approvals, notifying personnel of open invoices. Facilitating on-time completion and in some cases early approvals, has financial implications in the form of potential discounts or late fees when working with external vendors. Meanwhile, the SAP Batch Release Hub for Life Sciences helps automate the batch release process and reduce the risk of quality noncompliance before releasing any drugs to the market.
The AI-powered data analysis of Team Liquid’s ban/pick draft tool led to improvements in players’ performance in matches and prompted broader implementations across the organisation. “We started primarily in athletics about two and a half years ago, but we have now started to implement AI across every different business vertical at different paces. We’ve had the advantage where we’re already using so many of the different SAP products, whether it’s Concur or SuccessFactors, we’re able to more easily get adoption throughout the organisation irrespective of business vertical.”
A blueprint for AI success
While Bosch, Bristol Myers Squibb and Team Liquid operate in different industries and seek to solve different problems, they have several things in common that help drive success: they start at the whiteboard to identify high-impact functions in which an AI solution will deliver outsized impact, work with a cross-functional team to define specific problems and opportunities for AI to solve and include non-traditional metrics to measure success.