The vendor platform's track ending at go-live while the owned track, built through SAIF, carries onSHADOW MODETHE VENDOR PLATFORMDECOMMISSIONEDOWNED · CLAUDE-FIRSTSAIFGO-LIVE
Case study · LIBERATE · retail · MENA

A retail conglomerate. Core platform replaced. 60 days.

A large retail conglomerate replaced a core operations SaaS platform with a Claude-first system built on SAIF. They own every line.

LIBERATE60 days · From brief to production go-live$0 · Ongoing licence fees post-cutover1 · Vendor platform decommissioned
Sector
Retail · Consumer Goods
Region
MENA
Scale
Multi-brand retail conglomerate
Timeline
60 days · brief → production
Delivery path
LIBERATE

Replace the SaaS you rent

Stop paying to rent your business back. We replace vendor software with a system you own outright.

Explore the LIBERATE path
At a glance

Numbers behind the engagement.

60 days From brief to production go-liveFrom brief to production go-live

Client consent on file

$0 Ongoing licence fees post-cutoverOngoing licence fees post-cutover

Client consent on file

1 Vendor platform decommissionedVendor platform decommissioned

Client consent on file

Stage 01 · Challenge
01
Challenge

One vendor SaaS. Every brand. Compounding cost.

A single core operations platform sat between the group's HQ and every retail brand running under it. The vendor set the roadmap. The group paid to rent an operating model that no longer served the business.

Licence costs compounded with every new brand acquired. Feature requests queued behind global roadmap priorities that had nothing to do with MENA retail. The data warehouse was structured around what the vendor could export, not what the ops team needed to see.

The exit case was clear. What was missing was a delivery vehicle that could replace the platform without a multi-year rebuild and without leaving the group hostage to a second vendor.

One core SaaS platform

Every brand, across multiple Middle East markets.

Licence spend

Compounding with each new acquisition.

Vendor roadmap

Gating feature velocity.

The data warehouse

Shaped by vendor export limits, not by what the ops team needed to see.

A second vendor lock-in

Zero appetite for another multi-year tie.

Most firms in this category bundle a platform of their own with their services. You leave one rented system and arrive in theirs. saasinator builds the other way round: systems you own, with no platform licence, no lock-in, and the code and runbooks handed over.

Stage 02 · Approach
02
Approach

LIBERATE, delivered on SAIF.

We mapped the load-bearing workflows, rebuilt them Claude-first, and cut over to production in 60 days. The vendor platform was decommissioned on go-live day.

Brief phase locked scope, timeline and the eval harness in the first week. Build phase ran test-first — every capability shipped with a versioned eval suite, and no capability advanced without passing gates. Deploy phase ran a glass-walled cutover: a controlled shadow-mode window, then live.

Claude sits inside every workflow that the vendor platform used to own — inventory reconciliation, brand-level operations, HQ reporting, cross-brand analytics. Every commit lands in a client-owned repo. Every prompt, model call and eval is auditable in the client's own observability layer.

IN — what the vendor platform owned
  • Inventory reconciliation
  • Brand-level operations
  • HQ reporting
  • Cross-brand analytics
SAIF

the saasinator AI Factory — Brief → Build → Deploy

  • Brief
  • Build
  • Evals
  • Deploy
  • Transfer
OUT — what the client owns
  • The platform, outright
  • Source, model weights and eval harness
  • Runbooks and observability

Every commit lands in a client-owned repo; every prompt, model call and eval is auditable in the client's own observability layer.

Fixed-scope engagement

No change-request queue.

Source in the client's repo

From day one of Build — not on cutover, not on invoice-received.

Test-first delivery

Evals at every gate. No capability advanced without passing them.

Shadow-mode cutover

A controlled window, then live. Zero big-bang risk.

Sovereign hosting

Inside the client's own environment.

Stage 03 · Result
03
Result

The client owns the platform outright.

Post-cutover, ongoing licence fees are zero. Feature velocity is decoupled from any vendor roadmap. The group's ops team runs the system with runbooks, observability and eval harnesses handed over on go-live day.

The immediate outcome is the seven-figure licence line removed from the P&L. The compounding outcome is the freedom — every new brand acquired ships onto a platform the group controls, not one they rent.

Managed AI Services stayed on the table as an option, but the client opted to operate the platform in-house. Their engineering team runs it. We ship improvements when they commission them.

The engagement is live today: saasinator forward-deployed and AI engineers are extending the platform with new capabilities for the client.

Zero ongoing licence spend

Post-cutover.

Operated in-house

The ops team runs the platform unaided.

The feature backlog

Set by the group, not by a vendor.

Full handover

Source, model weights and eval harness.

Managed AI Services

Optional, not compulsory.

How we ran this

SAIF, applied.

Fixed scope

Priced before code.

Value modelled at Brief. No change-request queue.

Source day-one

In their repo, week 1.

Not on cutover. Not on invoice-received. Day one.

Evals gate the ship

Behaviour, not vibes.

Every capability shipped with a versioned harness.

They own it

We left. It runs.

Runbooks, observability, training — handed over.

Plan a liberation

Which of your platforms is next?

30 minutes. We'll tell you which of your rented SaaS lines is a LIBERATE candidate — and roughly what a 60-day replacement looks like on your stack.