Uploaded July 2025 | Updated September 2026, 2 weeks ago
In this powerhouse session, leaders from PepsiCo and IBM Consulting reveal how PepsiCo is scaling AI across a $92B global enterprise, from hyper-personalized AI-generated Gatorade bottles to AI-optimized retail shelves and predictive forecasting.
Speakers:
Nachiket Karajagi is Global Senior Director of Data & AI at PepsiCo
Shobhit Varshney is Head of Data & AI at IBM Consulting
⏱ Timestamps
00:02 - Welcome & introductions: PepsiCo, IBM Consulting, scale of AI initiatives
1:06 - PepsiCo’s value chain & surprising AI use cases: precision agriculture, soil testing
1:52 - The staggering data scale: 60 petabytes doubling yearly + IoT data
2:11 - Hyper-personalized Gatorade bottles: driving loyalty with AI-generated designs
3:43 - The impact on sales & ensuring brand-safe outputs
4:34 - Why observability is critical when scaling AI to billions of impressions
5:07 - How PepsiCo’s AI stack mirrors human intelligence: memory, attention, skills
5:15 - Inside Pep GenX: central platform enabling 100+ GenAI use cases
6:12 - Centralizing observability & evaluations with PepVigil
7:32 - Benefits of an abstracted, tech-agnostic platform for flexibility & speed
8:12 - Federating innovation across global teams while maintaining control
9:25 - Pep Vigil observing not just GenAI, but conversational, vision & IoT AI
10:28 - Example: planogram compliance via mobile edge AI — from manual surveys to real-time image processing
11:47 - Tackling scale: processing 10M+ images globally, managing regional nuances
12:26 - Using embeddings & clustering via Arize to detect drift & issues
13:32 - Computer vision for potato quality on conveyor belts
14:13 - Observability for forecasting models across markets
14:35 - Conversational AI: “chatting with data” with Pepsi’s FP&A bot
15:26 - Multi-layer filters for personalized Gatorade images (tying back to responsible AI)
15:43 - Enforcing responsible AI: upfront risk assessments, closed-loop policy execution
16:54 - The future: orchestrating multi-vendor agents, connecting ROI to business value
17:48 - The power of a “single pane of glass” for all AI across PepsiCo
18:48 - Closing thoughts & lessons learned: enabling fast time-to-market & trusted AI
Find other AI engineering events from Arize AI: arize.com/community
In this powerhouse session, leaders from PepsiCo and IBM Consulting reveal how PepsiCo is scaling AI across a $92B global enterprise, from hyper-personalized AI-generated Gatorade bottles to AI-optimized retail shelves and predictive forecasting.
Speakers:
Nachiket Karajagi is Global Senior Director of Data & AI at PepsiCo
Shobhit Varshney is Head of Data & AI at IBM Consulting
⏱ Timestamps
00:02 - Welcome & introductions: PepsiCo, IBM Consulting, scale of AI initiatives
1:06 - PepsiCo’s value chain & surprising AI use cases: precision agriculture, soil testing
1:52 - The staggering data scale: 60 petabytes doubling yearly + IoT data
2:11 - Hyper-personalized Gatorade bottles: driving loyalty with AI-generated designs
3:43 - The impact on sales & ensuring brand-safe outputs
4:34 - Why observability is critical when scaling AI to billions of impressions
5:07 - How PepsiCo’s AI stack mirrors human intelligence: memory, attention, skills
5:15 - Inside Pep GenX: central platform enabling 100+ GenAI use cases
6:12 - Centralizing observability & evaluations with PepVigil
7:32 - Benefits of an abstracted, tech-agnostic platform for flexibility & speed
8:12 - Federating innovation across global teams while maintaining control
9:25 - Pep Vigil observing not just GenAI, but conversational, vision & IoT AI
10:28 - Example: planogram compliance via mobile edge AI — from manual surveys to real-time image processing
11:47 - Tackling scale: processing 10M+ images globally, managing regional nuances
12:26 - Using embeddings & clustering via Arize to detect drift & issues
13:32 - Computer vision for potato quality on conveyor belts
14:13 - Observability for forecasting models across markets
14:35 - Conversational AI: “chatting with data” with Pepsi’s FP&A bot
15:26 - Multi-layer filters for personalized Gatorade images (tying back to responsible AI)
15:43 - Enforcing responsible AI: upfront risk assessments, closed-loop policy execution
16:54 - The future: orchestrating multi-vendor agents, connecting ROI to business value
17:48 - The power of a “single pane of glass” for all AI across PepsiCo
18:48 - Closing thoughts & lessons learned: enabling fast time-to-market & trusted AI
Find other AI engineering events from Arize AI: arize.com/community
