GBS Director Interview – Johnson Matthey (March 2026) - Summary

A Summary Transcript of last week's Video Interview with: Mariano Andrade Gonzalez, Director Global Business Solutions at Johnson Matthey (Vilnius).

Thom Barnhardt (European Business Services Association): Today we continue our series of GBS Director interviews with Mariano Andrade Gonzalez, Global Services Director at Johnson Matthey, based in Vilnius. We’ve seen you a few times over the years at some of our events in Warsaw. Thanks for taking a few minutes with us today.
First of all, give us a bit of background on what Johnson Matthey is doing globally, and particularly in the Vilnius office.

Mariano Andrade Gonzalez: Yeah, sure Thom, thanks for having me. It’s a pleasure to be working with you, and hello everybody.
Johnson Matthey is a 200-year-old British company. To simplify what we do: next time you’re on the street anywhere in the world, count the cars—one, two, and the third car with a combustion engine will likely have a catalytic converter from Johnson Matthey. We invented it in 1974, and in gasoline we probably have around 30% market share. In diesel, it’s even higher.
Anywhere you go—heavy-duty vehicles, tractors—anything with a combustion engine uses catalytic converters. Moving forward, you’ll also see our technology in energy production, especially in data centers that need cooling and additional energy supply.
We also operate in healthcare, providing catalysts and equipment used in cancer treatments and brain surgery. A lot of our applications are based on platinum group metals, which is the core of our business.

Okay. And in Vilnius, tell us about your support for that global business.

The Vilnius office is relatively new—just over two years old. We don’t have manufacturing plants in Lithuania; those are located across Europe, including Poland, North Macedonia, Germany, Sweden, and the UK.
Lithuania was chosen as part of developing a GBS model. Historically, Johnson Matthey was highly fragmented, with four major business sectors operating independently. Functions like finance and HR were not unified.
The goal of the GBS model was to break down silos and unify the company under one umbrella—which, as we’ve seen many times, is easier said than done.
We started in Lithuania with core functions: HR, finance, and later procurement.
HR was more advanced because Johnson Matthey had already implemented Workday as a global ERP platform about four years ago, so processes were relatively well mapped.
Finance was a different story. We had 32 different ERPs—ranging from old JD Edwards systems with green screens to multiple SAP instances, none configured the same way. There was no standardization, making it very difficult to get accurate financial data in a timely manner.
Today, from Lithuania, we provide record-to-report, FP&A, controllership, accounts payable (partly from Kolkata), and accounts receivable.
Previously, month-end closing took 12–14 days. Today, we’ve reduced it to around five days—significant improvement, though there’s still room to improve.
In accounts payable, we initially faced issues where payments were consistently late. We had to identify root causes, but without proper tools, it was nearly impossible. We had teams analyzing data manually, but they were only identifying symptoms, not root causes.

That’s typical—layers upon layers.

Exactly. And this was a major issue for a company spending over £1 billion annually in procurement. Some services, like gas and water for metal processing plants, are critical—any disruption stops operations entirely.
We needed better visibility. While I don’t believe technology solves everything—bad processes will just create bigger problems—in this case, we needed technology to identify issues.
We implemented a process mining tool, Celonis. It allowed us to connect all ERPs and quickly identify where payment failures occurred.
We found recurring issues:

• Poorly structured contracts
• Purchase orders dating back decades (1990s still in use)
• Inconsistent payment terms
• Excessive access to master data (over 2,000 employees with access in a 12,000-person company)

We’ve improved significantly—now critical payments are on time, and about 90% of payments are executed correctly. However, we are still standardizing vendor terms and improving governance.
We were losing money both ways:

• Paying too early (working capital loss)
• Paying too late (damaging supplier relationships)

Now we’ve largely stabilized that.

Importantly, GBS didn’t create these problems—it surfaced them. That’s a key point. The issues already existed; we just made them visible and actionable.

We’ve also started using basic AI capabilities, mainly through tools like Celonis:

• Generating SOPs
• Mapping processes
• Linking risks

But this is still very early-stage. AI is not replacing jobs—it’s helping us move faster and operate more efficiently.

Let me ask—are you getting a lot of AI vendors pitching solutions?

Yes. We are working with Genpact in a hybrid model for some services.
They are already deploying agentic AI solutions. I visited their operations in Kolkata to see this firsthand. My honest view: it’s not yet transformative.
For a company like ours—with fragmented systems and processes—AI won’t solve foundational problems. First, we need to standardize data and systems, which requires significant investment.
Also, deploying agentic AI would require giving full data access to a third party. That creates dependency risks I’m not comfortable with yet.
So we’re being selective. AI will improve, but its effectiveness depends on organizational maturity.
Accounts payable is one area where AI can work well—if inputs are standardized. But that’s not fully our case yet.
In record-to-report, automation is already strong. The question is: do you need AI everywhere, or is traditional automation sufficient? Right now, automation is often cheaper and more practical.
We’ve seen similar hype cycles before—like RPA. Many companies invested heavily, but adoption didn’t always follow.
So we’re cautious—testing selectively, not scaling prematurely.

Looking ahead, what’s next for your GBS model?

We’re still early. Only about 25% of roles suitable for centralization are currently in the GBS model.
The next phase is eliminating “shadow organizations”—duplicate roles across the business. That’s actually harder than building the GBS itself.
We need:

• Clean end-to-end process ownership
• Clear accountability
• Strong upstream and downstream integration
• Supporting technology

Over the next couple of years, the focus will be:

• Optimizing what we’ve built
• Strengthening processes
• Implementing the right supporting technologies

Only then will we determine where AI or further automation adds value.

Final question—how are geopolitics and supply chain disruptions affecting you?

Significantly. Energy costs, logistics disruptions, and supply chain instability are major challenges.
However, there’s not much we can control there. The focus is on maintaining operational stability.
GBS will continue to add value, but supply chains are an area to watch closely—they will be heavily disrupted in the coming years.