case study
Building an AI Roadmap to Fuel Growth Without Adding Headcount
How GSMS built a two-year AI plan on an owned asset competitors can't buy.
AI Solution(s)
Industry
Pharmaceuticals
How GSMS built a two-year AI plan on an owned asset competitors can't buy.
What worried David Arnold most about artificial intelligence had nothing to do with artificial intelligence. "I was most worried about hiring an external consultant who puts together a lot of PowerPoint slides that don't have any value for me."
Arnold is President and Chief Operating Officer at Golden State Medical Supply (GSMS), the government pharmaceutical supply chain partner to more than 150 manufacturers, founded by a US Air Force veteran with a vision of delivering improved dispensing efficiencies and cost reduction to the government healthcare system. Its customers, the VA and Department of Defense, buy only from trade-compliant suppliers. GSMS manages that compliance, and all of the related regulations involved with government procurement. Underneath that process sits something no competitor can buy. "We have the largest data warehouse of government pharmaceutical usage data," Arnold says, "and it's one of our advantages."
Owning the data and accessing it turned out to be two different things. Ask GSMS a question about its own data and the process was fixed. Write a requirements document. Send it to the analytics partners overseas. Wait. "We put in a requirements document, we sent it off," Arnold says. “The process to deliver the completed result could often take months.”
The capability to analyze the data existed in pockets. A VP of Business Analytics had been making the case for two years that this was an opportunity. "We've never had a corporate strategy around how to fully leverage this to our benefit going forward. That was really what started the discussions here. We made a commitment as a leadership team to say we're going to lean into this, and we are going to embed AI and advanced analytics in every aspect of our company."
GSMS is private equity backed, with a five-year goal of doubling EBITDA. Arnold asked the sponsor one question: inside your portfolio, who does credible work here, and who do you rate? Three names came back, from a sponsor that had not worked with any of them directly and gave GSMS full freedom to run its own process. Four firms went through a competitive RFQ. Two made the final.
"The technical depth within your team was clearly at a higher level than the other finalist," Arnold says. "So that really stuck out to us." The second point was that KUNGFU.AI was describing work it had done, rather than work it could do, "which is really useful to us," Arnold says, "because we could see the actual output."
Arnold is candid about the kind of client he is. "I'm part of the executive team, so I want everything done today. I'm impatient." GSMS wanted use cases, as soon as possible. KUNGFU.AI held a different line: AI governance first, then data and infrastructure, then the use cases those foundations make possible.
Not a comfortable message for an executive who wants to move. It is the part Arnold now singles out. "The learning from your team to us was: get the foundation right and then you can really take off. That was the biggest learning for me from the process."
GSMS is regulated by the FDA, the DEA and individual states, and leadership's position was simple. "When we came into this, we were like, hey, hands off, we're not going there." This boundary was assumed settled, and nobody argued the point. AI doesn't make compliance decisions. The team asked a narrower question: could AI support the people who do?
"Your team identified three pretty obvious areas that we could use these tools, still be fully compliant, but could likely bring real efficiencies to those processes," Arnold says. "That was a bit of an eye-opener for us." Three new opportunities opened up inside a part of the business that had been walled off from AI on day one.
GSMS left with a 24-month roadmap of ten AI use cases, each evaluated for feasibility and business impact, spanning both sides of the P&L. Alongside it came an AI governance framework that GSMS leadership helped shape rather than simply receive, and completed discovery and technical requirements for the highest-value pricing opportunity.
The roadmap was deliberately ambitious, built to give GSMS a full runway of vetted ideas rather than a handful of safe ones. GSMS is now converting it into its own two-year action plan, starting exactly where KUNGFU.AI had recommended.
"It starts with governance. We had no AI governance in our company. Your team was crystal clear: you better get that. So we're starting there." But what changed across all of GSMS mattered more than the plan. "Before this, we had pockets of needs within the company, but everyone was fighting for their own stuff." Now, he says, "we have alignment across our company. We know what the priorities are, who's in charge, what's the expected timeline, and why are we chasing what we're chasing."
Asked what will be different in two years, Arnold responds, "I like to look at it as a funnel. Today we start at the top of the funnel and work our way through. My belief is that we're going to be able to start further down the funnel and move through it much faster. Which ultimately just leads to more opportunities for our business and allows us to move faster than our competitors on multiple fronts."
Asked what it would have cost to work that foundation out alone, he is shorter.
"We wouldn't have got there."
He clarifies, because the distinction matters: the improvement projects came from GSMS. His people are the experts, not the consultants. "I don't believe we would have been able to define that foundation that we need in order to ultimately deliver on these initiatives. You've given us the formula needed for the foundation."
Asked what he would tell a peer weighing the same investment, Arnold gives an answer and a caveat together. "I would tell them absolutely it's worth it. It's a necessary investment to make sure that you are setting yourself up for multi-year success. You can't just chase the shiny object."
What convinced him in the first place had nothing to do with artificial intelligence either. He came in afraid of consultants with slide-ware and nothing behind it. What motivated GSMS was KUNGFU.AI’s long history of delivering results. "What we saw in the proposal work was exactly what we got as the end result. There was no hiding of the ball."
"We wouldn't have got there. You've given us the formula needed for the foundation."
