CUSTOMERS
Helping a global beverage manufacturer deploy an AI workforce to protect margins
Industry
Consumer packaged goods
Cases
Price variance / price arbitrage detection, Supplier and SKU consolidation, Maverick (off-contract) spend
Technology stack
SAP ECC, SAP Ariba, Microsoft Dynamics 365
Consumer packaged goods companies are being squeezed from both sides. Raw material, energy, and logistics costs keep rising, while consumers are becoming more price-sensitive, making it harder to pass those increases on.
For procurement teams, that means finding new ways to hunt down value, while balancing risk.
When a global beverage manufacturer approached Magentic, this task was still highly manual and time-consuming. Category managers had to work hard to field opportunities across previously neglected indirect spend categories, navigating fragmented systems, with no scalable way to turn data into evidence-backed cases without heavy manual analysis.
With teams already stretched, they didn’t need another dashboard, but a way to automate the work while remaining in full control. Magentic's AI digital workers stepped in to handle the work end to end.
Extending roll-out across spend categories
By working with Magentic, the team can move fast to capture value while reducing costs and operational risk.
After starting with office equipment, administration, and IT hardware and peripherals, the company is now working with Magentic on supporting more complex areas of the business.
Accelerating time to value
Together, we launched a two-month pilot to see how Magentic’s digital workers, Mages, could help the team stay on top of spend. The Mages work like the teams already do; analysing data directly from systems, such as SAP and Ariba, while cleaning up data as they go - no prior integration needed.
Spend variances are checked automatically, while maverick spend is prevented without each invoice being checked by hand. Category managers are only looped in for validation through a lightweight review process - with the evidence for each case attached. Each case comes with a set of recommended actions, replacing a previously slow and manual audit and execution process.



