Copy to clipboard 
Like many medtech companies, Johnson & Johnson encourages employees to experiment with generative AI in their daily work. The company’s initial strategy was to support employees as they learned, tested and built AI applications that, once validated, could be applied to broader teams and functions throughout the organization.
J&J has now shifted from an investigative approach to one that is more focused, Chief Information Officer Jim Swanson recently said in a Wall Street Journal article. Swanson noted that at one point, employees were pursuing nearly 900 individual use cases that involved generative AI, data science and intelligent automation. The ideas were shared with a centralized governance board that found many of the use cases were redundant and ineffective — 10% to 15% of the ideas drove about 80% of the value that was captured.
“We’re prioritizing, we’re scaling, we’re looking at the things that make the most sense,” Swanson said. “That was part of the maturation process we went through.”
One of the high-value use cases J&J recognized and is now pursuing relates to identifying and mitigating supply chain risks, including the impact of raw material shortages.
The adoption of AI is expected to revolutionize the management of supply chains, including the ways procurement and sourcing teams identify suppliers, maintain visibility into inventory, monitor for risks and nurture partner relationships. The potential benefits of AI have created intrigue and optimism, but like J&J, companies need to take a strategic approach to their use of the technology to ensure that tools correlate to productive activities and not distractions.
“Don’t implement AI because it’s trending, do it because it’s transformative,” said Vikas Yadav, Vice President of Consulting Services at GEP, a global provider of AI-powered procurement and supply chain services and software.
Yadav said procurement teams that are considering implementing AI should heed three pieces of advice: evaluate the technology for ROI, understand the effort it will take to complete and harmonize your data and confirm executive buy-in.
“Ten years ago, procurement was seen as a support function and didn’t have a seat at the table,” Yadav said. “As cost and margin pressures mount, AI can help procurement deliver value quicker. Technology amplifies the ability to work faster, smarter and more strategically in terms of how teams approach business requirements. Procurement should be looking at ways to adopt AI, because the technology is a game-changer.”
Where to Start
It’s exciting to consider leveraging AI to automate day-to-day tasks and identify problem-solving solutions. Yadav said that before companies evaluate and implement open tools and integrative platforms, they need to align their use of AI with business goals, information processes and the right people.
Determine Business Goals. Companies need to outline their business goals and translate them into functional steps for the procurement team. Goals might relate to pain points like improving stock availability or reducing COGs. Next, they need to evaluate how their data is structured across those pain points to achieve their goals. The exercise can help identify high-impact use cases for AI and allow companies to prioritize them based on which ones have the most potential for their return on investment.
“Companies should always start by mapping their pain points across procurement and supply chain to understand how overcoming those pain points meets business goals,” Yadav said.
He referenced a company that embarked on a cost-reduction journey to improve margins. The company’s first step was to leverage AI to classify its data. The accuracy of the data improved from 50% to 95% in a matter of weeks, providing a fast and high-reward use case for the company’s multi-step initiative.
Build from the Bottom. Data infrastructure is foundational to the success of an AI program. “AI runs on data, and if the quality isn’t good, then the output won’t be good,” Yadav said. “It’s a classic case of ‘garbage in, garbage out.’”
Most companies overestimate how well organized their data really is. Data is often incomplete or missing because it is scattered across multiple systems that don’t communicate with each other. Yadav said disorganization doesn’t mean that companies can’t embark on AI initiatives, but they first need to understand where data silos exist and put governance in place on which data to use and how to leverage it.
Companies should create an internal team and seek outside help to implement tools that pull data from multiple sources and rationalize it. An external perspective can provide fresh ideas, Yadav said. In his experience, companies often look at spreadsheets of data to answer a specific question or work on a particular project rather than taking a holistic view of what data reveals.
Hire and Train People. AI initiatives require teams with the right skillsets to run the program. Yadav places people into three categories: existing talent, specialized talent and external perspective.
If a company is already engaged in AI, it needs to focus heavily on upskilling its procurement teams by providing access to analytics, chatbots or co-pilots — basic AI tools. Companies also need to invest in training employees to understand the fundamentals of these technologies: how to input data effectively, what outputs to expect and how to interpret the results to drive informed decision making.
“Critical thinking and decision making in conjunction with AI-generated recommendations are important to emphasize,” Yadav said. “You can’t solely rely on AI recommendations. You need to evaluate your business requirements and combine them with the recommendations from the tool.”
Yadav said the objective should be to give the team confidence to trust but challenge the AI insights and make faster, more informed decisions. The goal is not to turn everyone into a data scientist, he added, but companies do need those skillsets, too.
AI product owners who understand and maintain the tools are an emerging role in companies. A new role specific to procurement is that of digital category manager. “An evolution of traditional category managers, this role will require fluency in digital tools, data analytics and AI applications to make faster, data-driven procurement decisions and manage supplier relationships more strategically,” Yadav said. “The role typically doesn’t exist in procurement today, but it will in the future.”
External software vendors are important to consider as an extension of a company’s team because they have a deep understanding of the quickly evolving landscape. “Consultants can help companies drive efficiencies, unlock value creation opportunities and stay informed of the emerging trends in the AI space,” Yadav said.
What Work to Tackle
AI tools offer various levels of sophistication. Companies have long deployed chatbots to answer customer service questions, whereas generative AI platforms used by the public have gained adoption recently. More sophisticated technology can unearth problems that companies don’t even realize they face.
Procurement teams should take an incremental step to implementing AI tools, but be aware of what’s possible with advanced technology. Yadav offered examples of simple, moderate and sophisticated use cases of AI.
Low-Risk Tools to Try. An accessible starting point for procurement teams involves conducting an AI-powered spend data classification. AI taxonomy models can help cleanse, classify and categorize the information across systems to give companies up-to-date visibility into accurate spending patterns and help uncover opportunities.
Another easy win for teams is to deploy AI for contract and invoice data extraction. Tools can accurately scan documents and pull out key fields like payment terms, pricing and renewal dates, and update the contracting system. Yadav said that when he works with companies, he typically finds that contract expiration dates aren’t easily accessible and this type of extraction often unveils numerous expired contracts.
A third use of AI is the implementation of chatbots that focus on information retrieval. Companies can feed process manuals and policies into an AI chatbot for teams to quickly access information.
Use Cases that Drive Impact. Cost modeling is often a static process. Procurement teams develop a model, implement it and revisit it a year later to make updates. The cycle repeats.
“Companies can integrate cost models with specific indices into AI systems,” Yadav said. “If your raw material prices are fluctuating, tariffs are coming into play or labor costs are changing, the model will update at a set frequency and provide you with predictive insights. For example, if the market is trending unfavorably, dynamic models might suggest that you renegotiate certain contracts now rather than wait several months. The tool provides forward-looking insights instead of a current viewpoint.”
Procurement teams are also gaining interest in autonomous sourcing. AI bots automate supplier identification, proposal evaluation and contract negotiation. Yadav said this use case reduces manual effort, speeds up the sourcing process and even provides strategy tactics.
The legal aspects of contract negotiations can also be streamlined with AI technology. “Legal teams spend a significant amount of time reviewing contracts,” Yadav said. “AI tools can examine the contract, flag risky terms and provide recommendations. Billion-dollar companies have contract lead times between two and three months. AI can be leveraged to cut the amount of time needed for the review process.”
Sophisticated Algorithms. Companies that are more experienced with AI and have the tools and talent in place to effectively manage the technology can engage in more advanced use cases that drive value for the organization.
Most companies strive to enhance efficiency across their operations. However, when data is fragmented across ERP systems, invoicing platforms, payment tools and process trackers, significant value can be lost, Yadav noted.
He cited the example of rebate clauses and year-over-year efficiency improvement clauses in supplier contracts that are often overlooked once an agreement is signed. If finance or procurement teams are not actively monitoring these agreements, the organization may miss out on capturing this value with the suppliers.
Procurement is heavily focused on metrics. The numbers might show that a team saved the business $100 million over the course of a year, but how much of that savings flowed to the bottom line is not always captured. AI tools can help assess value leakage areas and the system gaps that need to be resolved to capitalize on process improvement opportunities.
“AI can help understand and uncover value levers and claw back opportunities wherever possible,” Yadav said. “Preventing value leakages is one of the more complex use cases that we are solving.”
Why to Shift Mindsets
J&J said it is tracking progress with AI in three areas: the ability to successfully deploy and implement use cases, how widely the use cases are adopted and the extent to which they deliver on defined business outcomes. That structure allows teams to consider which opportunities to pursue and how to measure success.
The use of AI in supply chain management is often linked to benefits around improved efficiency, reduced costs, better decision-making and overall resiliency. Yadav said companies that fully embrace AI are also becoming more transparent, collaborative and performance-driven, both with internal and external teams. To reach that zone, though, companies need to have open conversations with employees and partners about what they seek to achieve with AI tools.
“Historically, procurement and sourcing functions have been based on relationships,” he said. “A team might think, if there’s no noise, the supplier is doing fine. AI brings real-time visibility and a 360-degree view of the supplier’s performance into focus. It’s going to shift the conversation from what went wrong to how we can anticipate and solve this problem together.”
