call_center_ai · energy · workflow
AT&T achieves 90% cost savings with fine-tuned small language models for call center operations
AT&T processes 15 million customer calls annually but needed a cost-effective way to extract actionable insights from transcribed interactions; large language models were too resource-intensive and expensive to deploy at scale.
How it works
Common implementation structure
How this type of workflow is generally built, generalized across documented cases — not tied to any one vendor's stack. Click any stage to read what happens there. Specific products that implement these stages appear in “Tools commonly seen” below.
Stage 1 · Call transcripts collected
AT&T handles 15 million customer calls annually, generating recorded, transcribed, and summarized interactions.
Tools used
H2O.aiH2O DanubeH2O LLM StudioGPT-4Llama
Outcome
Deploying fine-tuned small language models yielded 90% lower cost, 3x faster processing, 75% latency improvement, and 5x scalability, while the ensemble achieved 91% accuracy closely matching the previous solution.
Results
Volume3x
Cost replaced90%
Grounding & classification
Source type: vendor customer story
32 fields verified against source quotes.
data extractiondocument classificationcall recordingmetric backednamed customerproduction runtime claimedtools describedvendor confirmedworkflow describedtelecomaccuracy improvementcost reductioncycle time reductionresponse time reductionthroughput increasevendor customer storycall center aicustomer supportquality assuranceextract classify route