How an AI Council Actually Drives Adoption
Manage episode 493737670 series 2833920
What does it take to build and scale AI responsibly inside a major media organization?
In this episode, Amir sits down with Arvind Thinagarajan, formerly Head of Enterprise Data Science & Analytics at Gannett (publisher of USA Today), to unpack how they built a volunteer-driven AI Council that governs, guides, and accelerates AI initiatives across the company.
From prioritization frameworks to cross-functional subcommittees, Arvind shares the inner workings of a model that supports nearly 90 AI use cases — and might just inspire how your org tackles AI at scale.
Key Takeaways
– Gannett’s AI Council is fully volunteer-based and cross-functional, giving every department a voice in how AI gets used.
– New AI ideas bubble up from across the org — the council exists to prioritize and support them, not to own or build them.
– Every pilot starts with a scoped business case, clear success metrics, and a timeline.
– The enterprise rollout phase is intentional — ensuring tech used in pilots aligns with the broader IT stack.
– A separate IT Council works alongside the AI Council to avoid duplicate tools and ensure strategic alignment.
Timestamped Highlights
00:57 – What the AI Council is and how it’s structured
03:54 – How the council started and why it matters in media
06:19 – Subcommittees that cover everything from tooling to compliance
09:40 – Where AI use cases come from (hint: it’s not top-down)
11:39 – Who actually builds the solutions, and how governance plays out
16:55 – 87+ tracked use cases and what happens after a pilot succeeds
Quote of the Episode
“The AI Council exists to make sure there are no silos. We’re here to bring the right skill sets, tool sets, and mindsets together — to solve the right problems, the right way.”
Call to Action
If you’re working on AI in a complex org — or trying to sell into one — this episode will give you the playbook from inside a company doing it right.
Follow the show on your favorite podcast platform, share it with a teammate, and if you liked the episode, leave a quick review. It really helps.
488 episodes