[Un]Churned · 18 Mar 2026 · From the week of 16 March
How OpenAI Built a Champion Network to Drive AI Adoption ft. Christina Meng (OpenAI)
These are notes on the conversation, checked against its transcript. The episode itself has the full discussion.
In brief
Christina Meng, who runs the champion program at OpenAI and previously ran Slack's champion network, which grew to nearly 4,000 members, speaks with host Josh Schachter on Unchurned. The episode covers how OpenAI divides its champions into executive champions, champion leads and internal champions, how she runs executive dinners and virtual roundtables, how champions feed product work, and how the program measures activation. Her central argument is that AI adoption is a human change-management challenge rather than a tech rollout, which is why she sees it as a customer success and education initiative.
For founders
- Christina Meng said organizations that treat an AI rollout purely as a tech rollout are not successfully driving adoption, and that the organizations keeping humans at the center of the change management effort are succeeding.
- She described three champion types at OpenAI (executive champions, champion leads and internal champions), and said the program is currently designed around supporting executive champions and champion leads.
- She said early on power users and champions were used interchangeably, but she sees mature customers realising that champions need to want to teach their peers, not just use AI themselves.
- She said customers are increasingly building their own internal champion networks, because initial excitement around AI tends to fade or become uneven without a way to sustain it.
- She said OpenAI product managers now run early-stage ideas, directional messaging and light UX concepts past champions before building them.
For revenue leaders
- Christina Meng said OpenAI's champion program defines activation as weekly active usage over enabled seats, and that its main bet is on sustained activation rather than attendance or engagement.
- She said attributing champion program activities to activation is hard, though her hunch is that it is easier when the program focuses on champion leads and executive champions who hold decision-making influence.
- She said the executive champion dinners are intimate conversations about how to be an effective executive champion and sponsor, with OpenAI leadership and account teams invited to hear the discussion.
- She said her virtual roundtables work best with around 25 to 30 attendees, and that she keeps presenter content to about half the conversation because champions want to learn from each other.
- She said the Slack champion network ran in a separate designated workspace with external customers and NDAs in place, but she did not think that structure was scalable.
What was said 21, most useful first
Power users are not automatically champions; champions need to want to influence and teach their peers. Listen
Christina Meng said early in customers' AI journeys, power user and champion were often used interchangeably. She said mature customers reach a point where they recognise that a step exists between using AI for yourself and teaching others to customize workflows for their team, making the role more consultant-like.
“champions need to want to have some level of influence and have some level of influence in terms of teaching their peers.”
Listen to the episode Hiring & team building Link to this Report a problem
Executive champions are setting aside time and making it acceptable for employees to experiment with AI, through hackathons, workshops and dedicated time. Listen
Christina Meng said the biggest shift she sees is executive champions becoming more hands-on in change management and leading by example. She gave examples of hackathons, workshops and dedicating 10% of time to experimenting with AI as ways executives are unblocking experimentation.
“they're really stepping up to provide that space and unblock what needs to be unblocked to foster a lot of safe experimentation with AI”
Listen to the episode Leadership & culture Link to this Report a problem
Virtual champion roundtables work best with around 25 to 30 attendees to get deep insights. Listen
Christina Meng said she ran many event types in the first year to learn how champions want to engage. She said virtual roundtables are easier to run, and the sweet spot is about 25 to 30 attendees for deep conversations on adoption challenges. She takes insights from these sessions to design more scaled content.
“the sweet spot is around 25 to 30 attendees to get really deep insights.”
Listen to the episode Retention & customer success Link to this Report a problem
Roundtable content should take up about half the session, with the rest left for champions to talk to each other. Listen
Christina Meng said she kicks off roundtables with a couple of questions and picks topics from insights she hears from champions. She brings in OpenAI colleagues, such as a CSM on measuring value, but keeps their content to roughly half the conversation because the peer discussion is where the value comes from.
“I try to keep the content to Max like half of the, the conversation because there's just so much discussion that happens.”
Listen to the episode Retention & customer success Link to this Report a problem
OpenAI defines activation as weekly active usage over enabled seats. Listen
Christina Meng said the champion program moved from measuring attendance and engagement, which was what marketing focused on, toward operational rigor on activation and sustained activation. She said the program is aligned to driving activation and sustained activation as its main bet.
“Weekly active usage over enabled.”
Listen to the episode Metrics & finance Link to this Report a problem
Organizations that treat an AI rollout purely as a tech rollout are not successfully driving AI adoption. Listen
Christina Meng said asking ChatGPT a question is easy and the real value lies beyond that. She said the organizations keeping humans at the center of this large change management initiative are succeeding, and she sees this as a customer success and education initiative.
“The organizations who are treating the roll out of AI as purely a tech roll out, they're not successfully driving AI adoption”
OpenAI's champion program identifies three champion personas: executive champions, champion leads and internal champions. Listen
Christina Meng said OpenAI identified three champion personas, and that the program currently focuses on supporting executive champions and champion leads. Internal champions are the next segment the program plans to work with.
“we sort of identified 3 different champion personas. So there's not just like one type of champion.”
Listen to the episode Retention & customer success Link to this Report a problem
Customers building internal champion networks is a way to sustain AI adoption after initial excitement. Listen
Christina Meng said initial excitement around AI is expected, but adoption then tends to tail off or become very uneven. She said many OpenAI customers are building their own internal champion networks as a way to sustain adoption.
“I think that is a crucial way to sustain adoption because what we're seeing is that there's initial excitement, of course there is”
Listen to the episode Retention & customer success Link to this Report a problem
OpenAI runs off-the-record dinners for executive champions on how to be an effective executive champion and sponsor. Listen
Christina Meng said the dinners bring executive champions together to discuss common challenges. OpenAI leadership and account teams also attend to hear the discussion. She contrasted this with the customer marketing team's executive dinners and events, which cover a broader scope.
“we do dinners and we're talking about things like how to be a good and effective executive champion and sponsor.”
Listen to the episode Retention & customer success Link to this Report a problem
Successful roundtable formats are packaged as events-in-a-box playbooks and shared with other teams such as sales. Listen
Christina Meng said she is starting to scale formats that have worked with champions through playbooks. She said she is sharing these events-in-a-box with other parts of the organization, including sales teams, so they can run similar sessions.
“events in a box that I'm sharing with other parts of the organization like our sales teams”
Listen to the episode Retention & customer success Link to this Report a problem
The most common champion question is how other people are doing it, not how to use specific product features. Listen
Christina Meng said champions in the network rarely ask product how-to questions. Instead, they want to hear from peers and help each other, and she said everyone is learning together because AI is new and few people in organizations lead these efforts.
“The number one question is I just want to hear from other people how they're doing it and they're helping each other”
Listen to the episode Retention & customer success Link to this Report a problem
Attribution from champion program activities to activation is difficult, though focusing on leads and executive champions may make it easier. Listen
Christina Meng said there is a lot of correlation but it is hard to attribute impact, because champions lead the programs that drive adoption. Her hunch is that focusing on champion leads and executive champions, who have decision-making influence, makes attribution a little easier.
“my hunch is that because we're focusing a lot of our programs right now on our champion leads and our executive champions”
Listen to the episode Metrics & finance Link to this Report a problem
OpenAI moved away from using GPT usage or creation as the signal for finding champions. Listen
Christina Meng said the team looked at metrics like GPT usage or creation and how many people use a GPT when identifying champions. She said they then stepped away from that approach and took a more creative route to find the individuals leading adoption efforts.
“we looked at things like, you know, GPT usage or creation and how many people are using that GPT. But really it is we, we actually took a step away from that”
Listen to the episode Metrics & finance Link to this Report a problem
OpenAI product managers test early-stage ideas, messaging and light UX design with champions. Listen
Christina Meng said product managers tap champions to run early-stage ideas by them, sometimes directionally without even naming the feature. She said champions help pressure-test value messaging and light UX design, and that once she had the first couple of internal product champions at OpenAI, others wanted to tap into the group.
“bringing our product managers in or tapping into our champions when we want to sort of run early stage ideas by our champions”
Slack ran its champion community in a designated workspace with customers, using hidden emails and NDAs. Listen
Christina Meng said at Slack the entire champion network ran in a separate designated workspace, with all champions being customers. Champions were told on joining that external customers would be in the same workspace, and the team used NDAs, a privacy agreement and hidden email addresses.
“we made sure that Champions knew upon joining that there were going to be external folks in the same Slack workspace”
Listen to the episode Retention & customer success Link to this Report a problem
Slack's separate champion workspace was not scalable, and it was folded under Salesforce Trailblazers. Listen
Christina Meng said Slack's champion network reached just under 4,000 champions and was run entirely through Slack. She said it is no longer around and was folded under Trailblazers, which has tremendous reach, and that she did not think the separate workspace was scalable, though Slack is a good way to start for high real-time engagement.
“I think it was just a separate slack works, but I don't think it was scalable.”
Listen to the episode Retention & customer success Link to this Report a problem
Advocates speak on stage or in references; champions also drive adoption internally and act as the node within the organization. Listen
Christina Meng said OpenAI's champion program sits under post-sale success and works closely with advocacy and community teams; she works with marketing to find advocates among champions. She agreed with the host that champions can do advocate activities but their defining role is leading internal programs that drive adoption.
“they're actually the individuals who are leading these programs that are driving adoption internally.”
Listen to the episode Retention & customer success Link to this Report a problem
A clear shared brand and identity for the champion role is important, with Salesforce Trailblazers cited as an example. Listen
Christina Meng said she has had to champion the term champion and make clear what an OpenAI champion does, so champions can speak to it. She said she took inspiration from Salesforce Trailblazers, citing its education element and its strong brand, identity and shared purpose.
“I think having that brand and identity and like shared purpose is so important.”
Listen to the episode Positioning & marketing Link to this Report a problem
Christina Meng used Codex to build a web app that pulls and links champion program event information in one place. Listen
Christina Meng said she is not a technical person and cannot code. She said she used Codex to build a web app on top of her hard-to-read spreadsheet of internal and external champion events, pulling the right information for internal teams and linking to event recaps in Slack, notes and landing pages.
“I was able to use Codex to help me build a web app that essentially pulls the right level of information for our internal teams”
Champion leads have bi-directional influence, taking direction from executives or the board and turning it into programs that activate internal champions. Listen
Christina Meng described champion leads as the middle level, typically head of AI or head of AI enablement roles who design the systems and programs for AI education and adoption. She said they lead both strategic and tactical AI transformation at their organization.
“They're taking the guidance from their executive or their board and then turning it into programs that activate the internal champions to to lead the trainings”
Listen to the episode Retention & customer success Link to this Report a problem
Internal champions sit across departments and execute the enablement strategy that champion leads create. Listen
Christina Meng said internal champions often sit in different departments and execute the enablement or AI strategy that champion leads create. She said customers are increasingly creating internal champion networks so these champions can connect and share learnings, and that the role differs by organization. Host Josh Schachter suggested they host internal hackathons, act as go-to peers and sit closer to the individual contributor level; Christina said it is a mix.
“these are the people that are that are hosting kind of the internal pizza party hackathons and, and they're like the people, the peers that everybody goes to for help.”
Listen to the episode Retention & customer success Link to this Report a problem