How is Mars using AI to drive innovation across marketing, product development, supply chains, and enterprise operations?

Recorded live at AI4 2026 in Las Vegas, Sanjay Puri speaks with Shashank, who leads Enterprise Data Science and AI at Mars Snacking, about how Mars is building an AI-first culture through experimentation, innovation, and disciplined execution.

Shashank shares how Mars established its AI Lab as a physical and virtual sandbox where employees, data scientists, product managers, and business teams can experiment with emerging technologies. The lab has already helped turn ideas into patents and production applications, while creating a culture where experimentation and failure become organizational learning.

The conversation explores AI agents, generative AI, enterprise AI strategy, AI governance, AI ROI, innovation labs, digital twins, AI-powered marketing, supply chain AI, and the importance of knowing when to stop an AI pilot.

Shashank also explains why enterprises need to be deliberate about deploying AI agents, particularly when it comes to deciding how much cognition should be handed over to AI versus humans.

In this episode:

  • How Mars built an enterprise AI innovation lab
  • Turning AI experiments into production solutions
  • How Mars uses generative AI in marketing and product innovation
  • AI-powered campaigns and customer experiences
  • Using generative AI to develop new product flavors
  • Digital twins and AI in supply chain operations
  • How to evaluate AI experiments and ROI
  • Managing the cost of AI tokens and infrastructure
  • Why AI pilots need clear failure parameters
  • How to build a culture of experimentation
  • AI agents at Mars and the future of enterprise automation
  • How much cognition should be delegated to AI agents
  • Why low-cognitive AI agents may be easier to deploy
  • Open-source vs. closed-source AI
  • Why Agentic AI may be overhyped today
  • Build vs. buy: choosing the right AI strategy
  • Why operational experience matters in enterprise AI
  • The one AI decision Shashank would never trust AI to make

Whether you’re a Chief AI Officer, CIO, CTO, data leader, innovation executive, or enterprise AI practitioner, this conversation offers practical lessons on moving AI from experimentation to measurable business impact.

Recorded live at AI4 2026 in Las Vegas by Knowledge Networks.

Chapters

00:00 The AI decisions you should never delegate

01:04 Meet Shashank from Mars

01:40 Building Mars’ AI Innovation Lab

03:17 How the AI Lab has evolved

04:30 Who can experiment at the AI Lab?

05:36 Turning AI ideas into patents and products

07:27 Why every enterprise should consider an innovation lab

09:00 Leading Enterprise Data Science and AI at Mars

10:18 How Mars is using AI across the business

10:42 AI-powered marketing and product innovation

11:31 AI and digital twins in supply chain

12:17 Managing risk in public-facing generative AI

13:28 The changing economics of generative AI

14:44 How Mars measures AI ROI

16:16 Why AI experiments need measurable impact

16:53 The role of executive leadership in AI innovation

18:55 Mars’ approach to AI agents

19:39 How much cognition should AI agents have?

20:18 The cost and infrastructure behind AI agents

20:58 Why Mars is cautious about high-cognitive AI agents

21:38 From supply chain to enterprise AI

23:10 The operator mindset behind AI innovation

24:39 Why AI pilots need an exit strategy

25:43 Setting failure parameters before starting

26:46 Documenting failure and building organizational intelligence

28:15 Managing high-profile AI pilots

28:56 Lightning Round: Open vs. Closed AI

29:06 Is Agentic AI overhyped?

29:20 Build vs. Buy

29:25 Amazon vs. Mars

29:54 The AI decisions humans should never delegate

30:35 Key lessons from Mars’ AI journey 

Subscribe for more conversations with global leaders shaping the future of enterprise AI, innovation, governance, and digital transformation.

#ArtificialIntelligence #AI4 #AI42026 #EnterpriseAI #AIAgents #GenerativeAI #AIInnovation #AIGovernance #Mars #DataScience #SupplyChainAI #AIROI #ChiefAIOfficer #SanjayPuri