The agents

Meet the agents that run your growth

Three specialists on one shared context — your customer data, your systems, your rules.

Acquire — find the growth

Find the customers you're not reaching

Acquire maps your base into lifecycle cohorts and scores each customer's propensity to transact. The gaps get names — untouched actives, reachable hibernators, regions your calendar misses. Then it builds the plays to close them, sized before you spend.

  • //01Maps cohorts your campaigns never touch
  • //02Scores transaction propensity, customer by customer
  • //03Proposes sized acquisition plays for approval

A Middle East value-fashion retailer narrowed a −9.6pp regional growth gap to −2.2pp in five weeks — and grew new customers 29.4% year on year.

Explore Acquire →
acquire · propensity model · fictional
CohortP(transact)
Hibernators, 60–90 days
0.71
Lapsed, high basket
0.44
One-visit new joiners
0.32
Acquire agent2 plays proposed · 190K reach

Convert — idea to live

Move offers from idea to live, without tickets

Your team describes the offer in plain language — audience, window, reward. Convert authors the configuration and pushes it to any offer engine you run, once a person approves. No tickets. No engineering queue.

  • //01Reads the offer as your team wrote it
  • //02Authors the configuration, field by field
  • //03Pushes to your offer engine after sign-off

A consumer superapp cut offer creation time 50% and took configuration errors to zero across 20+ monthly offer templates.

Explore Convert →
convert · plain-language offer · fictional
“Buy two, get one free on denim — app members only, this weekend.”
Audience, window and reward parsed
Configuration authored — 14 fields
Push to your offer engineWaiting for you
Convert agentWorks with any offer engine

Retain — keep the base

Keep the customers you fought for

Retain models your loyalty economics — earn rates, tiers, milestones, points liability. Every proposed change carries a revenue projection, so you see the math before the mechanics move. Discount discipline is a rule it works inside, not a hope.

  • //01Models points economics, tiers and liability
  • //02Simulates milestone rewards with revenue projections
  • //03Proposes the redesign — your caps enforced

A premium menswear retailer redesigned its loyalty programme end-to-end in 38 days, holding discount discipline at 10% or under.

Explore Retain →
retain · revenue projection · fictional
12-month projection · repeat-revenue index
Proposed — visit milestones added112
Current programme100
Your rule: discount cap 8% — held in every iteration
Retain agentProjection attached to every change

How they work together

Different jobs. One operating loop.

The loop is the same for all three. It never skips the ask.

Signal Plan Ask Execute
01 · Signal

Any agent can raise one — a slipping cohort, a stalled offer, an ageing programme.

02 · Plan

aiRA drafts the play from your data, with audience, channels and revenue math attached.

03 · Ask

Every consequential step stops here and waits for a named human to approve.

04 · Execute

Approved work is pushed to your systems of record, and every action is logged.

Three agents. One growth engine.

They plan from the same data, execute in the same systems, and ask first.