Step 01 — Objective

Define what this campaign should protect.

The objective sets the frame for everything that follows. It re-ranks the segments, shapes which offers are allowed, and narrows the audience — so the recommendations you review are the ones that fit this campaign.

Campaign objective

Churn prevention

Hold on to customers whose risk is rising before they lapse.

Win-back

Re-engage customers who have already stopped buying.

Loyalty activation

Reward standing and move quiet members back into the program.

First-to-second purchase

Turn a single order into a repeat buying habit.

Category reactivation

Bring customers back to a category they have drifted from.

Service recovery

Repair the relationship where a service problem drove the risk.

Campaign period

Category focus

Value tier

Channel

Budget limit

EGP

Discount rules

Human review and approval required before launch

This control is fixed in the prototype by design. No campaign leaves the assistant without sign-off.

Continue to customer base

Objective saved. It now shapes the recommendations and brief for this session.

How the objective shapes the output

The objective re-ranks the segments. A service-recovery objective moves customers with open complaints to the top of the queue; a win-back objective lifts dormant accounts ahead of healthy ones.

The discount rule constrains which offers the assistant is allowed to propose. Choose points and vouchers only, and no price discount will appear in a single recommendation.

The value threshold filters the audience. A high-value focus narrows the base to the accounts that carry the most revenue at risk, so effort lands where it matters most.

Category focus and channel narrow the audience further, without changing the underlying risk scores — those are read from the data, not from the objective.

Settings persist for the session on the server. Reload this page and your objective is prefilled.