{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreihvsqcugormmwjf5tgzo7op2xmowo5itlt6gyrpnkvvw2iq7ujtra",
    "uri": "at://did:plc:qzjwstutqk2cy7df7jbzd2hx/app.bsky.feed.post/3mide2g76eh62"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreiglji4dlfnyjxlezv4tdnzzbnlu3jlwznkzpq5anvlp53v37org5u"
    },
    "mimeType": "image/jpeg",
    "size": 4689194
  },
  "path": "/article/4151954/how-lumen-is-dismantling-decades-of-network-complexity.html",
  "publishedAt": "2026-03-30T18:27:41.000Z",
  "site": "https://www.networkworld.com",
  "tags": [
    "Artificial Intelligence, Network Management Software, Networking",
    "Lumen Technologies",
    "NaaS",
    "Kina Corcoran",
    "digital twin",
    "single source of truth",
    "Alexandre Mercier-Dalphond",
    "Dare to Lead"
  ],
  "textContent": "Any network built through years of acquisitions will typically accumulate the structural debt of overlapping infrastructure, fragmented inventory, and no single view of what is running where.\n\nFor Lumen Technologies, that debt compounded across four decades and dozens of acquisitions. Lumen is an enterprise network infrastructure company serving large enterprise, government and wholesale customers, with a fiber network spanning roughly 500,000 route miles globally and $12.4 billion in 2025 revenue. The company has been repositioning from its legacy telephone carrier roots toward AI-era enterprise connectivity and network-as-a-service (NaaS).\n\nLumen now routes up to two-thirds of global internet traffic, operating the world’s most peered internet network, according to the company. As it has grown, each acquisition left behind its own systems, equipment and data silos. The result was 17-plus inventory systems, nearly 500 data sources with no common schema, and network hardware spanning 40 years of manufacturing generations.\n\nLumen’s response treats the problem as a data and automation challenge, not a conventional network refresh. The company has deployed AI agents across its inventory and ordering systems and built a unified operational workflow platform.\n\n“Something that took several resources, probably several weeks and several months to answer, those are questions we can answer now in a matter of minutes,” Kina Corcoran, Lumen’s chief data officer, told _Network World_.\n\n## Getting the data right first\n\nLooking at how Lumen was able to transform its network begins with data.\n\nEach acquisition Lumen absorbed came with its own data boundaries and its own definitions of network elements, customers and costs. None of those systems talked to each other. Engineers could not relate a piece of equipment to the customers riding it, or understand the cost and revenue implications of removing it, without pulling data manually from multiple disconnected sources. That made systematic simplification at scale effectively impossible.\n\nThe first step in transformation was building a unified data layer across all of those sources. Lumen ingested nearly 500 data sources into a common platform and built data objects that link network elements, customer services, cost data and revenue data across what were previously hard organizational and system boundaries.\n\n“This is the first time we’ve been able to relate those things to one another,” Corcoran said.\n\nThe outcome is what Corcoran describes as a digital twin that goes well beyond the network layer. “It’s a digital twin of our inventory, of our architecture, of our ecosystem,” she said.\n\nA representative use case is identifying all customers in a given metro that are running legacy voice services, determining the next best migration offer based on current network capacity and feature parity, and surfacing the path with the least customer disruption. That analysis previously required multiple teams working over weeks or months.\n\nThat unified data model is also what makes automation possible at the execution layer, where engineers are doing the actual decommission work.\n\n## Turning data into execution\n\nThe tool Lumen’s field engineers use to execute decommissions is called NetPal, a proprietary workflow tool built on top of its data platform.\n\nBefore it existed, taking a single piece of network equipment out of service meant manually querying and updating more than 50 separate systems. That included inventory records, service assurance, NOC coordination, traffic migration planning, and record updates across every affected node in the ring. Each step was sequential, and an error in any one system created downstream problems. The reason that process was so brittle is the same reason it took so long. There was no reliable single source of truth to query.\n\nNetPal changes that by putting a single interface on top of the unified data layer. An engineer identifies the target equipment, and the tool surfaces what traffic is riding it, which customers are affected, and what the energy consumption impact will be. It also maps how to consolidate that traffic and flags which downstream inventory records need updating. Dispatch coordination and field confirmation run through the same interface.\n\n“The output of my planning team now is more than 8x what it was,” Alexandre Mercier-Dalphond, senior vice president of infrastructure and operations modernization at Lumen, told _Network World_.\n\n## What that execution delivers\n\nThe throughput gains inside the planning team are only part of the picture. As inventory data becomes more accurate and decommissions execute more reliably, the operational benefits compound across the business.\n\nIn recent network outages, mean time to resolution dropped from several hours to 15 minutes. Field teams now dispatch against accurate inventory rather than spending the first portion of an outage reconciling conflicting records. First-time-right rates on truck rolls improved as well, cutting repeat dispatches.\n\nThe same clean data accelerated service delivery. Lumen’s Rapid Routes product delivers 400G wavelength services with AI-assisted pre-provisioning.\n\nLumen is also migrating off 15-plus legacy inventory systems, including some mainframe-based platforms, onto a single-pane-of-glass architecture using Blue Planet as its target platform. Mercier-Dalphond said Lumen expects to be the first tier-one operator to fully exit mainframe-based network inventory.\n\n## Lessons for network operators\n\nThere are several lessons learned by Lumen as part of its own network transformation that can be broadly applicable to other network operators facing issues of legacy complexity**.**\n\n  * **Clean data unlocks what you thought was impossible**. Clean inventory allows Lumen to identify and remove power-intensive legacy equipment systematically, keeping utility consumption flat even as new capacity is added. “Having clean data sets enables us to really make better decisions,” Mercier-Dalphond said. “We were surprised by finding, now that we’ve been able to put all this together, pockets of opportunities we never thought we had.”\n  * **Culture is part of the technical program**. Inherited conflict across systems and people can stall a transformation faster than any technical obstacle. Lumen ran a cultural program alongside the technical work, including a partnership with Brené Brown’s Dare to Lead framework. “A lot of that has really helped us change the minds and hearts of people internally, to think bigger, to think differently than we have in the past,” Corcoran said.\n  * **Poke the bear.** Most operators avoid pushing legacy customers to migrate, fearing revenue loss. Lumen changed that approach. “People were like, Alex, let’s not poke the bear on legacy revenue,” Mercier-Dalphond said. “We actually changed the narrative.” The decommission of Blue Voice, Lumen’s legacy voice network, is the proof case. Lumen approached a major enterprise customer, laid out the risk of staying on aging infrastructure, and executed the migration with the data platform handling dependencies.\n\n\n\nNetwork engineers have spent years knowing exactly what needed to be done with legacy infrastructure. The tools to do it at scale simply did not exist. Now with the power of AI, proper data, and the right processes, that has changed.\n\n“What we’ve been dreaming or hoping for is now possible,” Mercier-Dalphond said.",
  "title": "How Lumen is dismantling decades of network complexity"
}