.png)
.png)
The challenge: A fleet split across a dozen systems
PepsiCo runs one of the most complex distribution networks in Asia-Pacific, covering food and bevarge from start to finish, manufacturing to last-mile delivery. To manage its logistics, the in-house PepsiCo team collaborates with third-party logistics providers running their own transportation management systems. With such huge variety in the systems these 3PLs use as well as the variety of geographic challenges unique to each locale, PepsiCo was looking to advance the technology used without completly disrupting their network.
To give an idea of the level of mismatched of systems, some of its bottler network across Southeast Asia relies on spreadsheets and WhatsApp for large parts of daily dispatch!
To roll out a single route optimization platform, that mix of systems and data types is normally a dealbreaker. Most tools need data from a clean source before they can generate a usable route.
"A logistics AI that can only work with clean, standardized, API-ready data from a single source is not a logistics AI that works in the real world," says Richard Savoie, CEO and Co-Founder of Adiona. "It's a laboratory instrument."
PepsiCo needed a partner who could optimize routes without asking drivers, dispatchers or bottlers to change the systems they already knew. This was explicitly built into the criteria for selecting the 2026 cohort, along with alignment with PepsiCo’s pep+ sustainability priorities, ability to deliver measurable impact, readiness for commercialisation, and feasibility of integration into the company’s supply chain.
The solution: A normalisation layer
Adiona first joined PepsiCo's 2023 Greenhouse Program and built a data normalization engine capable of pulling logistics data from virtually any ERP or TMS a bottler happened to be running, SAP included, alongside custom EDI feeds and plain CSV exports.
The platform kept every existing workflow intact, so drivers still saw the same runsheet format. Dispatch also still worked from the same interface. The AI optimization ran while remaining invisible to the people using it day to day.
That design choice mattered more than any algorithm. Savoie points to internal trust (particularly along team members at the front lines like drivers,) not data quality, as the real barrier to adoption. A route planner with fifteen years of territory knowledge will treat an unfamiliar AI-generated route as wrong by default, unless the logic behind it is made transparent.
Once PepsiCo's teams were onboarded and understood how Adiona's routes worked, the optimiszation engine started producing usable plans within weeks of receiving a bottler's data.
The results: From pilot to permanent infrastructure
Adiona's early work with PepsiCo delivered a 19% reduction in fleet distance travelled, with further headroom to cut Scope 3 emissions across the bottler network, a target that sits at the centre of PepsiCo's climate disclosures. The company's Scope 3 emissions cover everything from ingredient sourcing to third-party logistics and make up the vast majority of its total footprint.
In 2026, PepsiCo restructured its Greenhouse Program around this. The new IMPACT Edition selected the cohort from five alumni companies it already had data on, Adiona among them, to test whether a proven pilot could survive integration into a business PepsiCo's size. The seven-month program runs through a phase-gate framework covering Integration, Measurement, Planning, Acceleration, Commercialisation and Tracking, with senior leadership sign-off required before a company can advance to the next stage.
"Assign your best operators to the pilot, not your most available ones," Savoie says. "The quality of a pilot is almost entirely determined by the quality of the internal engagement behind it."
Why it works
Route optimization is one of the few things a company can do to increase sustainability without spending money on new vehicles, infrastructure, or energy sources.
PepsiCo's Chief Sustainability Officer for Asia Pacific and India, Ashley Brown, frames the shift similarly in an interview with Asia Food Journal:
"The transition from pilot to integration is often where the real work begins. A pilot can demonstrate that a solution works, but integration requires proving that it can operate within the realities of a large business and become part of how the organisation runs day-to-day."
That's the gap Adiona is closing, turning a working algorithm into infrastructure a global bottler network runs on every day.
