Meridian Telecom
Multilingual support operation rebuilt from the queue up
Stood up a 22-agent pod across four languages in six weeks, replaced the routing logic, and added an AI assist layer that drafts responses and summarises calls. Service level moved from missed to contractually exceeded within one quarter.
Results
- Calls answered in 20s, SLA met
- 82/20Calls answered in 20s, SLA met
- CSAT out of 5
- 4.7CSAT out of 5
- First-contact resolution
- 72%First-contact resolution
- Total contact volume in two quarters
- -21%Total contact volume in two quarters
The challenge
What we walked into.
Meridian missed its service level every month for a year. At peak, roughly a third of calls were abandoned before an agent picked up.
The root cause was staffing, not attitude. The floor was rostered to a flat monthly average, so Mondays and the days after a campaign collapsed while Thursday afternoons sat idle. Their previous vendor ran a shared agent pool with no product training, and the same customer rarely spoke to anyone who knew the account.
Four languages were needed - German, Dutch, English and Italian - and nobody in the business could say which product problems were generating the calls.
The approach
What we actually did.
- 01
Forecast before headcount
We built an Erlang C forecast from eighteen months of their historical contact data, then rostered against it with intraday re-forecasting rather than to a monthly average.
- 02
A dedicated pod, trained before go-live
Twenty-two agents hired for this account only, put through product training and live shadowing before taking a single call, then ramped on a supervised traffic split.
- 03
Routing rebuilt
Skills-based, language-aware routing in Genesys, with scheduled callbacks replacing the queue at peak, so the abandoned-call number stopped being a function of hold music.
- 04
AI where it measurably helps
Response drafting grounded in their knowledge base, automatic call summarisation and disposition tagging. Humans keep anything with consequence; the model removes the typing.
- 05
Closing the loop with product
Every contact is tagged by root cause, and the top five drivers go to the product backlog each month. This is the part most outsourcers skip, and it is where the volume reduction came from.
Stack
- Genesys
- Zendesk
- Twilio
- Claude
- Metabase
They took over a support operation that was missing every service level and had it exceeding target within a quarter. What surprised me was the monthly report telling us which product bugs were generating the calls.
Anja P.
COO, Meridian Telecom
Practices used
Get started
Tell us what you need built — or who you need building it.
Send us the situation as it actually is. You get back a shaped engagement, a named team, a fixed timeline and a number you can put in front of a board.
A partner replies within one business day — with questions, not a sales sequence.