Here's my thinking on who should carry this AI work once we wrap in October. We talked this through together on July 29 and landed on a direction, which I've written up first. The four original options stay below for the record, with what I like and what I'd watch for on each.
You're a small team by choice, around fourteen now with maybe six more joining next year, and I want to make sure the AI work we build together doesn't fade once I step away. Someone needs to keep the policy current, help colleagues turn the things they do every month into simple workflows, and bring new hires into how Echidna uses AI. A full-time role would be far too much at your size. But if no one owns it, the policy goes stale and the momentum you've built slips within a couple of months. So the real question is how you hold a light, ongoing responsibility that doesn't slot neatly into anyone's current job.
It's a handful of things, and it's roughly a quarter to a third of someone's time during this phase, lighter after that:
I'd keep this separate from IT. Your outsourced IT partner can set up the accounts, but the rest of it is really about helping your team work better and trust their own judgment, and that's not something an IT function is set up to do.
You were already planning a database-focused hire. Bring it forward, and write ~25–30% AI ownership into the role alongside the EGGs and data work.
Keep an outside resource (SA Advisory or similar) on a light monthly retainer with set weekly availability. Staff hold their questions to that window.
Spread the responsibility across a few naturally enthusiastic staff, with AI treated as a creative objective, measured on effort and experimentation rather than a hard outcome number.
Bridge with a light external retainer from October, while deciding whether to fold the AI mandate into the data hire when it lands. If one or two staff show real appetite, let them grow into informal Guides with the retainer as backup.
The room chose the distributed model with external backup: 3 to 4 AI Guides drawn from across the team, plus a light retainer with me and Sara for the harder questions and for keeping the group accountable. That's Option C with Option B underneath it, and it moved away from my original hybrid recommendation for a good reason: the feeling in the room was that a hire now would aim before we know where the target is.
The pieces we agreed: no hire for roughly a year, until we can see where the real efficiency gains concentrate. Guides are opt-in first, tapped second, chosen for representation across regions, levels, and expertise, and the staff survey may surface natural candidates (one recent hire is already flagged as a strong one). Accountability runs through a regular debrief cadence rather than annual goals, with the retainer holding the rhythm. At the one-year mark we check where the traction is and decide whether a dedicated role earns its place, possibly as a phase 3 of this work.
The name matters and the room chose Guides over champions deliberately: a Guide shows you where the route is good and where it's a bad idea, including calling out poor use. Sherpas remains the standing alternative name. The full role description and expectations are drafted on the Guides page for your review.
Draft prepared by SA Advisory, July 2026, for Kim and Emily. Not yet shared with the wider team. Comments welcome → aggarwal.shabnam@gmail.com