Pilot success plan
A pilot with Alfred is not a technical evaluation — it is a working month. By day 30, [CUSTOMER] should be running its mornings on Alfred's briefs and be able to point to specific decisions the briefs produced. This plan defines the weeks, the roles, and the pass/fail criteria up front.
| Week | Focus | What happens | Done when |
|---|---|---|---|
| Week 1 | Connect | Secure read-only integrations across ad platforms, analytics and CRM — typically under an hour each, zero engineering. Up to a year of history syncs in the background. | Priority sources live (target: 6+); history synced |
| Week 1–2 | Calibrate | One working session: KPIs, alert thresholds, channels and business objectives tuned to how the team operates. The champion nominates the brief's recipients. | Calibration session held; thresholds signed off |
| Week 2–3 | First briefs | The Daily Brief starts landing — what changed, why, and what needs a decision, with sources and confidence on every recommendation. The team reads, reacts, and flags anything miscalibrated. | First brief by day 10; brief read 4+ days/week |
| Week 4 | Decision review | A 45-minute review with the exec sponsor: which recommendations were actioned, what they were worth, what Alfred missed, and whether the success criteria on page 2 were met. | Go / extend / stop decision made, in writing |
Champion
Owns the pilot day to day: connects the stack, attends calibration, makes sure the brief gets read, and collects the decision log for week 4.
~3 hours in week 1, then ~30 min/weekExec sponsor
Sets the success criteria at kickoff, receives the weekly pulse, and chairs the day-30 decision review. Where the go / no-go call lives.
Kickoff + 45-minute reviewAlfred + our team
Alfred delivers the briefs, alerts and recommendations. Our team runs calibration, monitors sync health, and prepares the week-4 review pack.
Daily briefs; support within one business dayWhat this costs the team
Under four hours of total customer effort across the month. No engineering tickets, no migration, no rip-and-replace — connections are read-only and revocable at any time, and every read is captured in the audit trail.
Success criteria
A pilot passes on evidence, not vibes. These are the defaults — edit the numbers with the exec sponsor at kickoff, then hold both sides to them.
What "3 decisions" looks like — demo scenario
In the fictional Northwind Labs dataset, week one alone surfaces a kill (a Meta campaign burning $4.8K with zero conversions), a reallocation ($18K from non-brand search to Performance Max) and a creative rotation (TikTok CTR down 12%) — $55.1K/mo of recoverable waste sized and sourced. Demo data, not a customer result.
| Outcome | When it applies | What happens next |
|---|---|---|
| Convert | Decision criteria met — 3+ actioned decisions and the team wants to keep its mornings | Move to an annual plan; calibration and history carry over, nothing is re-implemented |
| Extend once | Adoption criteria met but decisions lagged — for example, a slow connect or a quiet month | One two-week extension with a named blocker to remove; no second extension |
| Stop | Criteria unmet, or the briefs did not earn the team's attention | Integrations revoked, workspace exported on request, all data deleted within 30 days |
Ready to put a date on week one?
hello@seekalfred.ai · seekalfred.ai/contact