Product descriptions:
With close to 100 years of history as stewards of a powerful network — one that once carried the voice of a generation and connected millions across copper lines and legacy systems — Lumen Technologies today is a leading digital networking services company, reshaping the future of business for a multicloud, AI-first world. The brand has a broad portfolio encompassing one of the largest, most connected and deeply peered networks in the world, with cloud and edge computing, security, IT consulting and voice and collaboration offerings. As early adopters of GenAI, Lumen wanted to continually refine their use of the technology to provide real-time diagnoses, recommend various resolutions to issues and support ticket summarization for their dedicated team of technicians. However, the company was up against fragmented and inconsistent data, excessive data volume and velocity and data interpretation difficulties. Intent on remedying these issues, Lumen invested in Databricks to test and deploy GenAI workflows and saw a reduction in customer resolution times.
Reworking legacy logic causing innovation barriers
Lumen Technologies is committed to driving business growth by connecting people, data and applications — quickly, securely and effortlessly. To boost enterprise productivity, they initially adopted GenAI for workflows such as sales enablement and daily operations. Building on this foundation, Lumen planned to expand GenAI to support field technicians and enhance the enterprise customer experience. However, their proprietary diagnostic tool, Laser, relied on static, outdated logic to speed up issue resolution. This limited its effectiveness in a fast-evolving environment.
Undeterred, Lumen aimed to leverage GenAI to analyze diagnostic performance data and automatically summarize tickets. This would help enterprise technicians onboard faster by providing concise summaries of the last 90 days of ticket history, ensuring consistent and high-quality service for enterprise clients. But achieving this was challenging due to data fragmentation and complexity. As Rabih Nahas, Senior Director of Network Engineering at Lumen, explained, “We had tons of data, but it was siloed. Alarms were in one system, performance metrics in another and ticket notes in a third. Each system used different formats, so traditional logic couldn’t make sense of it all.”
Adding to the challenge, Lumen’s engineers had to sift through lengthy service tickets — sometimes 15 to 20 pages — to understand a customer’s history. Meanwhile, their systems still ran on logic developed over five years ago, with Laser processing over 3 million diagnostics annually. Facing these hurdles, Lumen needed a smarter, scalable solution to turn operational data into a seamless customer experience — and that’s where Databricks entered the picture.
Expanding GenAI usage with a focus on customer impact
After choosing the Databricks Data Intelligence Platform, Lumen enhanced Laser by integrating GenAI into their operational stack. Laser now processes diagnostic, performance and ticketing data at scale using Databricks. Central to this system is Model Serving, which enables Laser to send structured inputs via API to a Databricks-hosted Llama 3.3 model. Hard-coded prompts ensure consistent, clear outputs for each inference. Technicians simply click a UI button to trigger GenAI — no back-and-forth with the model is needed, delivering fast and reliable results. This separation of UI, back-end logic and model inference gives Lumen flexibility to evolve their architecture without disrupting workflows.
“Separating the model from the UI and back end means we can upgrade or swap models without breaking anything. As diagnostic volume grows, the system scales smoothly, crucial in our fast-paced customer-facing environment,” Rabih shared. Laser now handles over 3 million diagnostics annually, many initiated directly by customers through Lumen’s portal or APIs. The scale and real-time demands required a platform that could serve models quickly and reliably — Databricks delivered exactly that without adding operational complexity.
Lumen’s team developed and refined custom prompts in Mosaic AI Playground, avoiding lengthy IT provisioning. This prompt-first approach gave them control over how inputs were interpreted and responses formatted while minimizing early deployment risks. As GenAI use grew, they implemented Unity Catalog for governance, providing centralized, secure management of data across on-premises and cloud environments.
Turning data into responsive, informed technician support
Since launching their new AI-powered capabilities, Lumen has supported thousands of technician-handled service calls with model-driven insights. The engineering team expects significant reductions in issue isolation and resolution times alongside measurable productivity improvements.
Rabih noted, “By embedding model recommendations and ticket summaries into tools our technicians already use, we’ve made it easier to resolve issues on the first call. It reduces manual work and helps us respond faster with a consistent customer experience.”
Looking ahead, Lumen plans to expand GenAI’s impact by increasing ticket deflection from 30% to over 50%, further solidifying their position as a leading enterprise digital networking company. This will identify issues outside Lumen’s domain, reducing operational costs while maintaining quick responses. In addition, the Lumen team saw a dramatic productivity gain, saving an average of 15 minutes per task or an estimated 3,675 hours saved annually through the help of GenAI.
Future projects include automating ticket creation and closure, generating outage documentation and making technical data accessible to nontechnical users. As more applications migrate to the cloud, Lumen will leverage Databricks for business intelligence workflows — dashboards, reporting pipelines and real-time operational visibility. They’re also exploring Databricks Apps to bring AI capabilities directly to business users with modular interfaces.
To further enhance AI power, Lumen is preparing to implement retrieval augmented generation (RAG) by indexing internal knowledge bases for Vector Search. Security and observability remain a priority, with GenAI set to handle anomaly detection across vast datasets. Databricks’ integrated platform and AI tools will be essential to scaling these innovations securely and efficiently. By continuing to invest in these core services, Lumen is positioned for future success with a data architecture ready to support broader GenAI adoption and next-generation agentic AI systems, delivering fast, trusted access to well-managed data.
“Databricks helped us move from static diagnostics to smart, scalable support,” Rabih said. “By embedding GenAI directly into our technician workflows, we’re solving problems faster, cutting costs and delivering the kind of enterprise-grade service our customers expect in the AI era.”