016 AI for Operators

Eren Celebi, AI Lead at WPP; Review of Happenstance.ai, 7 links

Hi there, 

Welcome back to AI for Operators. Here’s what we’ve got for you this week:

  • The Operator: Eren Celebi, AI Lead at WPP (listen to the episode here)

  • The Review: Happenstance.ai: social network search

  • The Links: 7 links, including custom tutors from Google, insights from CPG companies, and virtual agent economies.

The Operator

Eren Celebi, AI Lead at WPP

Eren Celebi, AI Lead at WPP, joined the podcast to talk about his hybrid role helping build AI solutions for WPP customers as well as leading the company’s internal AI community.

  • WPP - a global communications/marketing conglomerate with 130,000 employees - has embraced AI both internally and externally.

  • Being ‘forward deployed’ is key to create real change with such a transformative technology.

  • Eren’s core philosophy: put the technology expert beside those who know the product best, ship demos not decks, and let problems drive the tech - the biggest risk isn’t security, it’s believability.

  • WPP’s clients want faster and cheaper research: WPP uses synthetic personas/audiences to test creative, UX, CTR, and price sensitivity before real-world A/B tests.

  • Adoption playbook: start with individual productivity tools, watch for repetitive tasks, then automate and iterate.

  • WPP is investing heavily in AI and built an internal multi-model platform with creative tools and shareable custom agents (e.g., briefs, JD writer); scaling to tens of thousands is the costly part.

The Review

This is not a sponsored post.

What it Does

It’s an AI-powered people search for your professional and personal networks: tell it what you’re looking for (in natural language), and it will return a list of people that most closely match your query.

Why Ops Leaders Should Care

Recruiting, sales, partnerships, external expertise: putting your extended network to good use is incredibly important as an ops leader in numerous ways. An easier way to see who’s actually in your network will help you move faster and make better decisions.

Key Features of Gamma 3.0 (Pros & Cons)

Pros

  • Natural language search means that queries like “Head of Sales” also returns the VP Sales and CROs in your network.

  • Unify your disparate networks/contact books from LinkedIn, email, Twitter, etc. into a single place.

  • Inherently social: results show who your mutual connections are; you can also add 'friends’ to your network to be able to search through their networks, too (opt-in only!).

  • Groups: want to embark on a recruiting power hour, or need to see who on your sales team might be connected to a key prospect? You can create a collective network database to search through, without fully giving each member’s privacy.

  • Integrations: tag Happenstance in Slack or forward a request to [email protected] to work with Happenstance where you spend the rest of your time.

Cons

  • Limited to people in your known contacts: you won’t have access to everyone else in the world, which means your results will be limited.

  • Privacy concerns with “friending”: lack of granular access control means that if you “friend” someone, they can see your full network, which you may not want.

  • Friction to onboard your full network: while it automatically parses your email, you’ll to download its Chrome extension or manually download your contacts from X and LinkedIn to add them to your network on Happenstance.

  • Small, early-stage team: with ~$550k raised and a team of only a few people, you may be concerned about providing your data to - or starting to rely too heavily - on a less-than-established company.

Other Options

Bottom Line

Is it “by far the best social network search” as YC’s Garry Tan claims? Maybe. While it’s an early-stage product, it’s worth trying if you lean heavily on your network, whether for sales, recruiting, fundraising, partnerships, or outside expertise.

Join Chief of Staff Network and Lindy

At a hand’s-on workshop on building agents on October 15 at 2pm ET

The Links

Practical

Perspectives

  • Virtual Agent Economies…: A new paper examining what is likely to happen when AI agents start coordinating and transacting at scale. Will they create a separate economy (a la “Her”), enmesh their economy in the human economy, or create some hybrid? And what will that mean for us? Some analysis here

  • …are here today: Google is using stablecoins to let AI agents pay each other

  • Does scaling AI have diminishing returns? Not according to this paper.

News

  • Reddit stands up for itself: Guess what OpenAI, Anthropic, and all the rest have trained their models on: your Reddit comments. And Reddit has benefitted, to the tune of a 400% increase in their stock price since they IPO’d in 2024. But now, they want more.

  • AI agents arrive at Citi: The bank is running a 4-6 week pilot with 5,000 users across a variety of functions, looking for a “massive boost of capacity.”

Thanks for reading,

Tom Guthrie

If you liked this edition of AI for Operators, share it with another operator you know!