Conversations with Betty — featuring Ann Gergen, Executive Director of AGRiP
There’s a moment in every AI project where the idea stops being theoretical and starts being practical. For the Association of Governmental Risk Pools (AGRiP), that moment came quickly — and it changed how members access, share, and apply industry knowledge.
In this edition of AI in Action, we sat down with Ann Gergen, Executive Director of AGRiP, to talk about why they launched their Betty AI assistant, Poppi, how members are actually using her, and what surprised them most along the way
AGRiP is an association of organizations, not individuals. About 250 member organizations are part of AGRiP, representing roughly 3,500–4,000 individual professionals. Those professionals span different roles, levels of experience, and governance structures — from staff to board members.
What they share is a need for context.
Members regularly ask questions like:
Who else is working on this issue?
How are other risk pools handling this challenge?
What does best practice look like across the industry right now?
Historically, answering those questions required staff to rely on memory, internal notes, old files, and years of conversations. The answers existed — but accessing them took time.
Poppi didn’t introduce new information. She introduced a better way to access it.
Poppi’s domain is the public-facing websites of AGRiP member organizations. That means she can surface patterns, practices, and approaches across the industry in real time.
Instead of waiting for staff to research and respond, members can now ask Poppi directly:
How are other pools delivering training on this topic?
What approaches are being used across the industry?
How do organizations like mine handle this situation?
As Ann put it, Poppi is essentially a stand-in for the entire AGRiP staff team — but faster, more thorough, and not limited by human memory.
“We aren’t doing something new. We’re doing something better.”
One of the most interesting takeaways from this conversation was how quickly members moved beyond information retrieval and into content creation.
Members often:
Ask Poppi for industry context
Review how peers are approaching a challenge
Then ask Poppi to help create a first draft of what they need next
That might be:
A training outline
A member-facing article
A board presentation
A planning document
In one live session, Ann even discovered (in real time) that Poppi could generate charts — something she hadn’t tested before.
Poppi isn’t just answering questions. She’s helping members move their work forward.
AGRiP originally planned weeks of internal testing before releasing Poppi to members.
Instead, they made her live within days.
Why?
Because waiting didn’t make sense.
AGRiP chose transparency over perfection. They told members:
This is new
This is an experiment
You’re part of shaping it
That decision paid off.
Members didn’t abandon Poppi when answers weren’t perfect. They adapted. They refined their questions. They learned how to work with the AI — and in the process, they helped train it.
“They didn’t give up. They asked the question a different way.”
One word defined AGRiP’s approach: experimentation.
Rather than positioning Poppi as a finished product, AGRiP framed her as an ongoing experiment — and invited members into that mindset.
That framing:
Lowered pressure
Set realistic expectations
Encouraged curiosity instead of criticism
It also helped members build confidence not just in Poppi, but in their own ability to use AI effectively.
Looking ahead, members are asking for ways to personalize outputs even further — such as uploading templates or formats and having Poppi adapt her responses accordingly.
But there’s no rush.
AGRiP is focused on continuing to explore what Poppi can already do, learning alongside their members, and letting adoption evolve naturally.
Poppi shows what happens when AI is:
Anchored in trusted content
Designed around real member needs
Rolled out with transparency and intention
This isn’t about replacing people. It’s about scaling institutional knowledge — and giving members faster, easier access to the insight that already exists within their community.
That’s Betty in action.