Commerce operations · AI colleagues
AI colleagues are agentic teammates that work alongside your ecommerce team: they review your catalog, order, campaign, customer, and search signals on a schedule, then deliver finished work — slides, PDF, Word, or Excel — by email. CHATTERgo ships ready-made colleague roles (Penny, Sterling, Bell, Iris, Nova) for merchandising, operations, marketing, insights, and search.
Deliverables by email · Microsoft Teams (staff, on request) · Agent Studio for Agents and Runs
The challenge
Into recurring work: the weekly merchandising review that someone assembles by hand, the catalog checks nobody has time to finish, the campaign report that means exporting three CSVs and praying the numbers line up.
Every one of these jobs is predictable, repeatable, and needed again next week — which is exactly why they’re expensive. Either a skilled teammate spends hours on them, or they slip when the week gets busy.
An AI colleague changes who does them: a teammate with its own role, instructions, memory, and desk that runs the checks on schedule and sends a finished file — while your team keeps the judgment calls.
Start from a role template — Penny, Sterling, Bell, Iris, and Nova each ship with a persona, a charter, a desk, and starter tasks. Eight templates exist in the product, and you can also build a colleague from scratch.
Catalog health, unusual orders, promo conflicts, support themes, search visibility — recurring jobs run without anyone remembering to run them, or kick off on demand when you need a one-off.
Each run writes a usable deliverable — slides, PDF, Word, or Excel — not a wall of chat. Open it, work it, forward it.
You decide what each colleague can see and do. Findings arrive with source context, so your team can verify before acting.
Every colleague has its own instructions, memory, desk, and files. Agent Studio lets you manage Agents and Runs — pick a shipped template or start from scratch.
Chat privately or @mention them in a Microsoft Teams channel, with the report and its attachment posted alongside. For staff, not shoppers — available on request.
Setup
Step 1
Eight role templates ship in the product. Pick the one closest to the job, or open Agent Studio and build your own from a blank desk.
Step 2
Point the colleague at the store, catalog, order, campaign, or search signals it needs. Set the trigger — schedule, by hand, or an incoming email — and what it may see and do.
Step 3
Review the file — slides, PDF, Word, or Excel — refine the colleague’s instructions, and let the schedule take over. Adjust any time.
Excel · on schedule
Broken variants, missing fields, and price conflicts flagged in a spreadsheet your team can work through — with source context on every finding.
PDF · by hand or from email
Overlapping promotions and pricing conflicts surfaced before they reach customers, in a file you can drop straight into the weekly review.
FAQ
AI colleagues are agentic teammates for your team: software coworkers with their own role, instructions, memory, and workspace. Unlike a chat assistant, a colleague owns recurring jobs — runs them on a schedule, by hand, or from an email — and delivers finished work as files your team can open and act on.
A copilot assists in the moment, inside a conversation. An agent answers or acts when asked. An AI colleague owns a recurring job end to end: it knows when to run, what to check, and what “done” looks like — a usable file with source context. Think teammate with a job description, not a tool.
Eight role templates ship in the product. The five commerce templates highlighted on this page cover merchandising, operations, marketing, insights, and search — Penny, Sterling, Bell, Iris, and Nova, each with a persona, charter, desk, and starter tasks. You can also build a colleague from scratch.
Reports land by email as slides, PDF, Word, or Excel, and are kept in the colleague’s workspace. Microsoft Teams delivery and chat are available on request, for staff.
You decide, per colleague: which sources it reads — store, catalog, order, campaign, customer, or search signals — and what it may do. Every finding arrives with source context so your team can verify before acting.
No. They take the recurring checks and report assembly that eat your team’s week, and hand back files with a clear next step. The judgment — what to change, what to escalate, what to stop — stays with your people.
Start from a commerce role template, or build your own — the first report can land before your next team meeting.
Explore the shopper experience, a commerce workflow, or both. Tell us where you want to start.
Checking form availability…