Replenit
Hire Maestro

2027 Retail Leaders’ Series · The business case

Why Replenit, why now: revenue from day one, outcomes owned.

This report makes the commercial case for hiring Maestro, Replenit’s AI CRM Manager. It shows what a retailer gets on day one, how revenue is generated from a standing start, the results retailers already see in production, and why Replenit carries the outcome instead of selling a licence.

  • 18.9x

    average ROI across customers, reached in under 12 months

    Replenit customer results

  • Month 1

    ROI positive: Maestro pays for itself in the first month

    Replenit customer results

  • 54 days

    for Maestro to run iBOOD’s #1 automation flow

    iBOOD case study

  • 90%

    lower AI pipeline cost, validated by Google Cloud

    Google Cloud, 2026

Hire MaestroRead the case

About this report

The commercial case forhiring Maestro

  • What this document is for

    The commercial case for Replenit: the metrics, the proof and why the timing matters.

  • Who it is for

    Executives, CRM and ecommerce leaders who need a number before they commit budget.

  • How to read it

    01 to 02 cover timing and ownership. 03 shows how revenue starts. 04 to 05 are the proof. 06 to 07 cover the horizon and sizing.

01

Why now

Every quarter in the campaign modelleaves revenue on the table

The market has already moved. Confidence in martech AI is falling, and the operating model is shifting from campaigns to decisions. Waiting does not keep options open; it keeps the gap open.

The share of retail leaders who say their AI investment is working fell from 54% in 2025 to 38% in 2026. 78% say martech does not deliver the ROI it promised. BCG expects 70 to 80% of mature customer touchpoints to run on next-best action instead of campaigns. The question is no longer whether the model changes, but how much revenue is lost before it does.

Exhibit 1

From the same $1M objective, about 4x more is realised from day one

Campaign era

Built by hand

$80K

8% realised

Legacy + martech AI

Same model, AI on top

$100–130K

10–13% realised

With Replenit

Day one, then with learning

$300K → $380K+

30% → 38% realised

Illustrative $1M retention objective; the dashed frame is the full objective. Replenit analysis, from “The New Operating Model for Customer Engagement” handbook. Market figures: MarTech.org, “The truth about martech in 2026”; eClerx “Mind the Gap” 2026 (n=366); BCG 2026.

02

Ownership

Replenit owns the outcome,not just the software

Martech is paid whether it works or not. Maestro is hired like a team member and judged on the commercial result.

Every skill runs against a holdout group, so the result is incremental revenue, not attributed revenue. Every decision is logged with its reason. Replenit is accountable for the number the retailer cares about: repeat revenue, basket, retention.

Exhibit 2

The ROI path: positive in month one, compounding after

  1. Month 1

    ROI positive

    Maestro pays for itself in the first month on pre-built skills.

  2. Month 6

    5x to 10x

    ROI as skills and Memory compound.

  3. Within 12 months

    18.9x

    Average ROI across customers.

Replenit customer case studies, replen.it/case-study. Results as published by Replenit.

Exhibit 2b · Customer example

Escentual: revenue per decision rose 9.1x as the loop closed

Revenue per decision

Indexed · launch = 1.0

At launch

1.0x

Today

9.1x

Launch and today are measured; the path between them is drawn for illustration. Source: Escentual case study, replen.it.

  1. Decide

    One customer, one product, one moment

  2. Deliver

    Through the existing Klaviyo stack

  3. Observe

    The outcome is recorded against the reasoning

  4. Enrich

    Memory, relationships and skills are re-scored

Back to 1

This closed loop is why the return compounds. Every outcome confirms or rejects a product relationship, deepens Customer Memory and sharpens the skill that made the call. Escentual’s three workflows (replenishment, cross-sell, substitute) share one Memory, so each one learns from the others.

A real decision · Escentual customer (anonymised)

Memory
A nightly retinol routine built over 14 months. The serum is about nine days from empty. Reacts to fragrance and avoids essential oils.
Decision
Replenishment email timed to her own consumption cycle, at the variant she actually uses
Skill applied
“The routine, not the last order”: reasons across the routine she is building, not her last basket
Next phase
The same Memory drives the storefront: a fragrance-free barrier repair moisturiser first, because retinol use rose and the routine has no barrier step

Read the full case study: replen.it/case-study/escentual-double-digit-roi-ai-transformation

Exhibit 3

Who carries the risk

Question

A martech licence

Hiring Maestro

What you pay for

Profiles, messages, events, seats, AI modules

A commercial outcome

Who does the work

Your team builds and runs campaigns

Maestro operates 24/7; your team directs

How results are proven

Platform attribution

Incremental revenue per skill against holdout

What happens if it underdelivers

Licence is paid anyway

Replenit is accountable for the number

What you keep

Journeys locked in the vendor’s tool

Your skills and enriched Memory

03

Revenue from a standing start

We enrich your data, bring the skills, andyou build your own

Most AI projects stall waiting for clean data, a model and a team to run it. Maestro starts with all three, so revenue starts before your team has built anything.

Order history and a product feed are enough to begin. Maestro enriches that data into Customer, Product and Brand Memory, applies more than 100 pre-built retail skills plus the presets for your industry, and runs them through the channels you already have. Your team then adds the skills that are specific to your business.

Exhibit 4

Three sources of revenue from day one

  • 1 · We enrich your data

    Memory

    Maestro reads orders and catalogue and derives what the data does not say: household, routine, fit, product relationships, replenishment cycles.

    164K

    data fields enriched at iBOOD, with zero engineering on their side

  • 2 · We bring the skills

    Retail and industry presets

    100+ retail skills plus industry presets for beauty, fashion, pet, electronics, grocery and more. Live from day one, tuned on your guardrails.

    100+

    pre-built retail skills ready at onboarding

  • 3 · You build your own

    Your skills, your IP

    Your team writes skills in plain language. Each is checked before it goes live, versioned, and applied to every customer.

    12

    skills in production at Escentual, 3 taught by their own team

Exhibit 5

What is ready on day one, and what your team adds

Skill

What Maestro learns and acts on

KPI it moves

Retail presets · every retailer

First to second purchase

What the first basket says about need and price sensitivity

Second purchase rate

Replenishment timing

Each customer’s own consumption cycle per product

Repeat revenue

Product-level churn preventor

An active customer quietly dropping one recurring product

Category revenue retained

Ready to subscribe

Regularity and stability of repeat purchases

Subscription uptake

Industry presets · examples

Fit understanding (fashion)

Size and fit, inferred from what a customer keeps and returns

Return rate; second purchase

Markdown guard (fashion)

Who buys at full price and who waits for the sale

Full-price revenue share

Routine builder (beauty)

Skin or hair type inferred from the products a customer buys

Items per customer

Seasonal care (pet)

Pet type and life stage from the basket; seasonal care needs

Category attach rate

Your own skills · example

Silent rich-buyer

A top-value customer who has gone quiet against their own rhythm

Revenue per top-tier client

Skills belong to the retailer. Their data, guardrails and logic are never used to train models for, or serve, anyone else.

04

Proof

Maestro inproduction

The numbers above are not a forecast. These retailers run Maestro in production today, on the stacks they already had.

Exhibit 6

iBOOD: lifecycle revenue in a business built for one-time purchases

iBOOD

Multi-category deals marketplace · Europe

16.6x

return on investment

Upstream MongoDB

Maestro

Downstream existing CEP

  • 54 days

    to the #1 automation flow

  • 6.3%

    of total revenue

  • +27%

    higher AOV

  • 164K

    data fields enriched

  • 5

    new countries in expansion

Before

A flash-sale model with no static product feed, no recurring inventory and no availability guarantees. Messages went out when deals reappeared, not when customers were ready.

What Maestro does

Reads customer behaviour instead of the catalogue and decides what to offer next, when, and through which channel across email, app push and web push. It found 100+ repeat-worthy items on its own, with zero engineering on iBOOD’s side.

Why it matters: revenue from a standing start. No product feed, no engineering, and Maestro still became the #1 automation flow in 54 days.

“Working with Replenit has been a game-changer for our retention strategy. The setup was incredibly smooth, no tech lift required, and their team truly understood our challenges from day one.”

Hannah van den Berg, Sr. CRM Marketer, iBOOD

Exhibit 7

Escentual: skills built with the customer’s own team, compounding over time

Escentual

Premium beauty and fragrance · United Kingdom

31.6x

ROI on the workflows Maestro owns

Upstream GCP + Shopify

Maestro

Downstream Klaviyo

  • 12

    skills in production

  • 362K

    individual decisions delivered

  • 9.1x

    revenue per decision vs launch

  • 120K+

    product relationships derived

  • 3

    workflows owned: replenishment, cross-sell, substitute

Before

Tens of thousands of SKUs that run out, get replaced and pair differently. A campaign calendar could not tell a 100ml eau de parfum from a nightly retinol serum.

What Maestro does

Escentual’s merchandising and CRM leads taught Maestro three business-specific skills. Inside their guardrails (brand-partner exclusions, premium positioning, no discount dependency), Maestro decides who, what, when and why, and executes through Klaviyo.

Why it matters: intelligence compounds. Revenue per decision is 9.1x its launch level because every outcome sharpens the next decision.

“Their capability to handle complex data sets and develop customized solutions to meet our specific requirements has been outstanding.”

Rakesh Aggarwal, CEO, Escentual

Exhibit 8

Results across verticals

  • +235%

    L’Occitane

    post-purchase revenue · Beauty

  • 42x

    Mumzworld

    ROI · Mom and baby marketplace

  • 61x

    Farmex

    ROI and 12x higher conversion · Pharma

  • 14x

    Kito Pet

    ROI, +27% AOV; 1-day Shopify integration · Pet

  • +340%

    Ovabalance

    repeat revenue · Wellness

  • 12.7%

    Faith In Nature

    contribution to total revenue · Beauty

  • +53%

    Glosel

    marketing automation revenue · Online grocery

  • 18.9x

    Average ROI

    across customers, in under 12 months

Replenit customer case studies, replen.it/case-study. Results as published by Replenit.

05

The engine

Validated by Google Cloud:1:1 decisions at a cost that scales

Reasoning about every customer every day is only viable if it is cheap to run. Google Cloud published how Replenit built that engine.

Spotlight

Google Cloud customer story

  • 360%

    business growth in one quarter

  • 95%

    faster pipeline processing

  • 90%

    lower pipeline cost: from over $200 to $20 per run

  • 4 hrs

    to ship a new production pipeline, down from 3 days

  • BigQuery gathers, normalises and enriches customer and product data into unified records.

  • Gemini Enterprise Agent Platform orchestrates hundreds of specialised agents running in parallel.

  • Fine-tuned Gemma on Cloud TPUs delivers high-volume reasoning at a fraction of the cost.

  • Enterprise governance: SOC 2 controls, data separation, tenant isolation and auditable decisions.

Engine facts

What each engine fact means for a retailer

Engine fact

What it means for a retailer

90% lower pipeline cost

1:1 reasoning for every customer, not only a top tier, at a viable cost

Pipelines shipped in hours

New skills and data sources go live fast

Hundreds of agents in parallel

Enrichment and decisions keep pace with a large customer base

SOC 2, tenant isolation

Your data stays separated and every decision can be audited

06

The horizon

Short-term outcomes insidea long-term strategy

Replenit is bought for results this quarter and kept because it changes how customer engagement runs for years.

Exhibit 9

What you get, and when

Horizon

Commercial outcome

What you build

What happens to the stack

Month 1

ROI positive on pre-built skills

Guardrails, shelf, first objectives

Unchanged; Maestro sends through current tools

Month 6

5x to 10x ROI as skills and Memory compound

First skills of your own

Logic leaves journeys; premium AI tiers reviewed

Within 12 months

Average 18.9x ROI across customers

A skill library across lifecycle, channels and brands

Channel tools move to core delivery tiers at renewal

Long run

Lower cost per outcome; skills compound as IP

Skills that encode your commercial know-how

Downstream centralised on API-driven delivery

07

Next step

Size the case foryour business

Three inputs turn the illustrative example into your own figures: revenue uplift, hours released and margin protected.

  1. Your annual retention revenue objective

  2. The hours your team spends each month building, launching and reporting campaigns

  3. Your current discount spend on retention

Before you decide

Four questions to ask your team

  1. What share of our retention objective actually reaches customers today?

    In the campaign model, often under 15%.

  2. Who is accountable if our engagement tools do not move the number?

    If the answer is “we are”, the risk sits with you.

  3. How long before our own AI project produces revenue?

    Maestro starts with enriched data and pre-built skills.

  4. Which commercial know-how would we keep if a key CRM person left tomorrow?

    Skills keep it inside the business.

Don’t buy new software. Hire Maestro.

The next step is to map your retention objectives to Maestro’s pre-built skills and agree the first skills your team would teach.

Hire MaestroCalculate your ROI