Every latency spike pulled engineers in by hand. Moving Gemini to their own infrastructure — not Google's — cut both the risk and the cost.
One of Korea's leading telecommunications carriers, generating over KRW 15 trillion in annual revenue and serving more than 30.71 million wireless subscribers, with LLM-powered search and Voice AI services live in production.
The company's AI search service ran almost entirely on external LLM providers, leaving token costs, response speed, output quality, and uptime in the...
“We want every AI experience we deliver to be accurate and trustworthy. Databricks helped us get there while improving both our cost efficiency and our stability.”
Mobile network operator and telecommunications provider in South Korea.
Databricks is a Big Data company that offers a unified analytics platform for data science, engineering, and analytics teams.
LG U+'s Customer search is part of this use case:
Related implementations across industries and use cases
Reps lost time hunting across complex systems. Now, AI agents surface verified answers and automate notes, freeing humans to connect.
The math was solid; only data scientists could read it. An AI agent now translates site predictions into plain language answers instantly.
Pulling answers from a sprawling customer dataset once meant weeks of SQL; now anyone asks in plain language and explores it in seconds.
Acquisitions fragmented data into silos; regulated datasets took months to unlock for AI. Unified governance made them available instantly.
Rate limits capped geocoding at 2,000 requests/min. A fine-tuned LLM now clears 8,000/min—cutting costs by 80%.
Pulling answers from a sprawling customer dataset once meant weeks of SQL; now anyone asks in plain language and explores it in seconds.
Call centers couldn't keep pace with demand peaks and rising costs. Now, agents handle 24/7 messaging, with AI summarizing and translating.
Querying Wikidata required specialized syntax, locking out most AI systems. Vector search now lets LLMs navigate 100M+ entities in plain language.
On-premise systems, dispersed and brittle, bottlenecked every release. AI agents now run routine dev steps — hours cut to minutes.
Every latency spike pulled engineers in by hand. Moving Gemini to their own infrastructure — not Google's — cut both the risk and the cost.
One of Korea's leading telecommunications carriers, generating over KRW 15 trillion in annual revenue and serving more than 30.71 million wireless subscribers, with LLM-powered search and Voice AI services live in production.
The company's AI search service ran almost entirely on external LLM providers, leaving token costs, response speed, output quality, and uptime in the...
“We want every AI experience we deliver to be accurate and trustworthy. Databricks helped us get there while improving both our cost efficiency and our stability.”
Mobile network operator and telecommunications provider in South Korea.
Databricks is a Big Data company that offers a unified analytics platform for data science, engineering, and analytics teams.
LG U+'s Customer search is part of this use case:
Related implementations across industries and use cases
Reps lost time hunting across complex systems. Now, AI agents surface verified answers and automate notes, freeing humans to connect.
The math was solid; only data scientists could read it. An AI agent now translates site predictions into plain language answers instantly.
Pulling answers from a sprawling customer dataset once meant weeks of SQL; now anyone asks in plain language and explores it in seconds.
Acquisitions fragmented data into silos; regulated datasets took months to unlock for AI. Unified governance made them available instantly.
Rate limits capped geocoding at 2,000 requests/min. A fine-tuned LLM now clears 8,000/min—cutting costs by 80%.
Pulling answers from a sprawling customer dataset once meant weeks of SQL; now anyone asks in plain language and explores it in seconds.
Call centers couldn't keep pace with demand peaks and rising costs. Now, agents handle 24/7 messaging, with AI summarizing and translating.
Querying Wikidata required specialized syntax, locking out most AI systems. Vector search now lets LLMs navigate 100M+ entities in plain language.
On-premise systems, dispersed and brittle, bottlenecked every release. AI agents now run routine dev steps — hours cut to minutes.