65%
Queries resolved without a person
→ All case studies
FMCG · EuropePost-merger integration · Agentic AI
The AI Hypercare That Held a Merger Together
A global FMCG conglomerate was consolidating three legacy platforms into one system mid-merger, with 5,000+ users to support and 10,000+ queries projected in the first 90 days. A human-only support model could not scale to that. We built one that could.
4.2h → 38min
Average ticket resolution
4
Languages served by the agent
12 months
2023 – 2024
How the system runs
The input5,000+ users mid-merger
→
Three legacy platforms consolidating into one, 10,000+ queries projected.
We builtRAG agents
→
Merged documentation, SOPs and process guides, answering in 4 languages.
We builtEscalation logic
→
Genuinely uncertain queries route to specialists, humans stay in the loop.
We builtRole architecture
→
200+ roles mapped against the merged operating model.
The outcome4.2h → 38 min
65% auto-resolved; SLA held at 3x projected volume.
The situation
Three Legacy Platforms, One Merger, No Unified Documentation
Integration planning had underestimated the knowledge-transfer problem: end users were trained on legacy systems, documentation for the merged platform did not exist in one place, and the support model was over-indexed on headcount. At the same time, 200+ roles had to be redesigned as the merged entity took shape.
What we didA RAG Agent for the Queries, an Operating Model for the Roles
Across twelve months, the team ran the hypercare build and the organisational redesign in parallel:
- Designed and deployed retrieval-augmented (RAG) AI agents that ingested the merged platform’s documentation, SOPs and process guides and answered user queries in natural language.
- Deployed in four languages: English, Dutch, German and French.
- Kept humans in the loop: escalation logic routed genuinely uncertain queries to specialists.
- Mapped 200+ roles against the new operating model, delivering the merged entity’s role architecture.
The design principle
The result
"Let the agent absorb the question volume, so the people can absorb the change."
Stabilised Within SLA, Despite Three Times the Expected Volume
- 65% of hypercare queries resolved with no human escalation.
- Average resolution time down from 4.2 hours to 38 minutes.
- The merged entity stabilised within its hypercare SLA despite 3x the projected query volume.
- The role architecture identified an 18% FTE reduction through consolidation and was adopted as the global standard for the function.
Support and organisational savings of $8–12M a year, at six to eight times the engagement fee in year one.
Our fee
$1.2M – $1.8M
Duration
12 months
Team
Lead + 23 FTE
Client anonymised under NDA. Figures are engagement-reported ranges, measured against the baseline set at the start.
Your number next
Want a Savings Estimate for Your Workflow?
A baseline findings report puts a real figure on the table, before you commit to anything bigger.