AI checkpoint: The wins vs. the big rocks — Data-Driven Telco Summit 2026, Panel Discussion 1 video

The Data-Driven Telco Summit 2026, Panel Discussion 1 
— AI checkpoint: The wins vs. the big rocks

• Moderator: Matt Leary (TelcoTitans).
• Panelists: Patrick Kelly (Appledore Research); Christoph Neuenkirchen (Kearney); Justin Reilly, (Wavelo).

Panel overview

Every stand at MWC26 says AI. The opening panel of the Wavelo-sponsored Data-Driven Telco Summit set out to test that claim against reality — and found a sector with genuine ambition undermined by a much older problem: data trapped in decades of legacy systems.

A sector in need of a “health boost”

Christoph Neuenkirchen, Partner at Kearney, introduced the firm’s newly published index scoring European telecom sector health across financial, technology, and consumer sentiment measures [00:07:00–00:08:00]. His verdict was direct: “the state of the telco sector in Europe we see is in need of a health boost” [00:07:56], driven by financial returns under pressure, capital-intensive technology investment, and a fragmented market — “there’s three players in the US, we have 90 players in Europe” [00:08:52] — competing against both US tech giants and Chinese operators.

Pressed on where AI ambition breaks down, Neuenkirchen was equally direct: it isn’t a lack of ideas. “It doesn’t lack inhibition… there are great ideas in telcos about AI” [00:09:04], but siloed organisations and, above all, data that “is not accessible… not real time accessible” [00:09:47] mean initiatives stall before they scale.

The private equity view: buying four telcos, getting 300 systems

Justin Reilly, CEO of Wavelo, framed the same problem through the lens of infrastructure investors. Private equity firms rolling up telecom assets expect the usual synergies — construction, revenue, back-office — and instead find, in Reilly’s words, “300 different systems and none of the portfolio companies you’re buying are on any of the same systems” [00:11:16]. Adding a data point from an operating partner at a major PE firm: two acquired companies with system estates worth $200–400 million each, and “not a single system that overlapped” [00:11:47–00:12:00].

Reilly’s conclusion wasn’t that legacy has to be ripped out before AI can start. His pitch — echoing his firm’s own positioning — was for an event-driven approach that lets operators “get value now” from a brownfield estate while modernising underneath it [00:20:30]. His sharpest line on the stakes: “your telco app sat next to your Uber app” [00:20:52] — telcos are no longer being judged against telco competitors, but against the best digital experience a customer had that morning.

The size of the prize — and a number worth checking

Patrick Kelly of Appledore Research put figures against the opportunity: a five-year, $6 billion forecast for agentic AI specifically within the telco segment, split across IT tooling, digital enablement, service management/operations, and network security [00:15:00–00:16:30]. He separately cited close to $1 trillion being spent this year on AI infrastructure build-out broadly [00:22:15], and a further $100 billion enterprise-market opportunity opening up over the next 3–4 years [00:30:12].

Where the “boulders” are

Asked what’s blocking the opportunity, all three converged on the same answer from different angles: data fragmentation. Kelly: “you need an underlying data structure. You need a knowledge graph… it’s that semantic reasoning that becomes the foundation for this forecast. If you don’t do that, these numbers just don’t play out” [00:17:11–00:17:37]. Neuenkirchen’s own pulse-check with telco CEOs at MWC26 found operators needing “10 to 12 months to launch a new product” and unable even to differentiate pricing because of legacy IT sprawl [00:18:53–00:19:24].

The panel’s shared note of hope: none of them think operators need to wait for a perfect data environment first. Neuenkirchen: “is there a way to not wait for the perfect data… to still get going and to start building… while modernising at the same time underneath?” [00:19:34–00:20:00].

From the floor

Audience questions pushed into sharper territory — including whether AI-driven multiple expansion in the US software market could leave less-prepared European operators dangerously exposed [00:33:44–00:34:54]. Reilly’s response drew a line between the interface layer (vulnerable) and the “system of record” layer deep in the stack (defensible, if operators are willing to open it up to interoperate with AI agents) [00:35:00–00:36:47]. Kelly added a counterpoint on incumbent moat: “you’ve got companies like Nokia and Ericsson, they have a lot of domain expertise… tribal knowledge” that’s harder to displace than pure-software incumbents [00:37:01–00:37:30].