BFCM 2026 · Peak season readiness
Is your AI actually ready for peak?
Twelve checks across the three things that decide it: what your AI is allowed to do, what happens after it gives up, and whether anyone reads what it said. Score yourself in two minutes, then run your numbers.
The CX team inside
and more!
Step 1 · The checklist
Twelve checks, three moves.
Tick only what is true today, not what is on the roadmap. Anything you have to think about for more than five seconds is a no.
Scope
What your AI is allowed to do. Not which topics it answers, which consequences it can cause.
Instrument
What happens in the minutes after your AI gives up. The bot's dashboard will not show you this.
QA
Reading what your AI actually said. An AI agent is a new hire that never asks for feedback.
Tick at least one box, or be honest and scroll down anyway.
Your AI readiness score
—
Scope
Instrument
QA
Your gaps, in the order we would fix them
Scope first, because it is the only one that creates liability. Then instrumentation, because you cannot fix what you cannot see. QA last, because it is the habit that keeps the other two true.
Step 2 · The numbers
Now run your volume.
Pull these from last year's peak window (Black Friday through mid-December). Estimates are fine. The output is a planning range, not an audit.
Shopify › Analytics › Reports, BFCM through mid-December.
All channels: email, chat, SMS, social. Exclude spam.
Use the revenue plan, not last year's actuals.
Full-time equivalents on the queue, including any seasonal help.
Resolution rate, not deflection rate. If your dashboard only shows "handled" or "deflected", pick "I don't know" — that is a finding, not a failure.
Enter orders, tickets, forecast and agent count to run the numbers.
Your peak-day coverage gap
—
Contact rate
Projected tickets
across the three weeks that actually carry peak
Peak day volume
roughly 2x the daily average of the window
Agent-days needed
on your peak day, at 60 tickets per agent-day
What AI readiness is worth, in your headcount
Same volume, different resolution rates. This is the entire business case for getting the AI layer right, and it only holds if the number you report is resolution, not deflection.
| AI resolution rate | Agents needed on peak day | Versus what you have |
|---|
What we would flag first
How this is calculated: contact rate is tickets divided by orders, per 1,000. Projected volume applies that rate to your forecast. The window is 21 days, because WISMO peaks five to ten days after Black Friday and returns run into January; peak day is modelled at twice the window's daily average. Agent capacity is 60 tickets per agent-day, the published benchmark for ecommerce helpdesks; a trained pod runs well above it, so this is the conservative read. Market contact rate band is 200 to 500 tickets per 1,000 orders. Your own numbers will differ by vertical, channel mix and how much of your volume is WISMO.