49% of tickets resolved with no human touch, against 4% in the previous scenario.
Project context
The size of the chain, how many stores and distribution centers it runs and how the service desk covered that footprint before the project.
AI agents inside the process, not in a channel in front of it.
A multichannel food retail chain with 21,000 employees, three banners, 420 stores and 4 distribution centres ran a 78-person service desk for 15,000 IT users. That meant 31,500 tickets a month, 68% SLA compliance and 11h20 average resolution time.
Working with Central IT, the operation deployed CITSmart ITSM with the Hyper Agents and Autonomous AI engines. Twelve months later, 49% of tickets were being resolved with no human touch and SLA compliance reached 96%.
About the company
Company not identified under a confidentiality agreement. The figures describe the size and profile of the operation.
Sector
Multichannel food retail, three banners
Size
21,000 employees
Operation
420 stores and 4 distribution centres
IT
15,000 users and 78 people on the service desk
Solution focus
- Triage, resolution and interaction executed by AI agents
- Wave-based rollout with a confidence threshold per category
- Three ITSM tools consolidated into one
Results in numbers
0%
No human touch
of tickets resolved with no analyst, against 4% before.
0%
SLA compliance
against 68% in the previous scenario.
2h38
Resolution time
on average, against 11h20 before.
6.4 months
Payback
on the investment, with 210% ROI over three years.
Strategic points
Challenges
31,500 tickets a month, 68% SLA compliance, 11h20 average resolution time. Only 4% of the volume was resolved with no human touch, and every POS outage left the store 41 minutes without selling.
Four causes sustained those numbers: entirely human handling, with every ticket waiting for a person to read, classify and route it; no automatic triage, with wrong classification in a third of cases, producing wrong queues and rework; disconnected systems, forcing analysts to navigate between directory, ERP and monitoring to solve a single case; and three ITSM tools running in parallel, one per banner, leaving the holding company with no single view of service.
Solution
Central IT deployed CITSmart ITSM with the Hyper Agents and Autonomous AI engines, placing five agents inside the ticket lifecycle. The Interaction agent receives requests via WhatsApp and portal, transcribes audio and identifies user and location. Triage classifies category, urgency and impact on arrival. Resolution executes the action on the target system and validates the result. Knowledge maintains the article base, and Problem correlates recurrence.
The rollout ran in waves of five categories, with a confidence threshold and exit criteria defined before each wave. Every category starts under supervised autonomy, with human review by sampling, and only expands coverage after hitting its measured accuracy target.
Measurable delivery
Results and impact on the operation
What changed across twelve months with autonomous agents working inside the ticket lifecycle.
Results
96% SLA compliance, against 68% before the project.
Average resolution time of 2h38, against 11h20.
First-contact resolution of 84%, against 38%.
In-store POS downtime cut from 41 to 11 minutes per incident.
Cost per ticket 30% lower, already including the recurring cost of AI.
R$ 2.98 million gained in the first year, with 210% ROI over three years.
No layoffs: fourteen outsourced first-level positions not renewed and 12 in-house analysts redeployed to automation and problem management.
Impact
The R$ 2.98 million gained in the first year comes from four lines: R$ 1.34 million in outsourced first-level positions not renewed, R$ 640,000 in sales recovered through POS availability at peak hours, R$ 600,000 from consolidating three ITSM tools into one, and R$ 400,000 in overtime and on-call during seasonal peaks. Even in the conservative scenario, with benefits 15% below and costs 15% above the model, the investment pays for itself in 8.6 months.
Because this is our business: empowering people. Impacting organizations.
Talk to a Central IT specialist
See how a service desk of this size reaches half its tickets with no human touch. Our team helps size the return using the real numbers of your operation, from diagnosis to the first wave.