Interview
The telecommunications business has lengthy been characterised by cyclical waves of technological innovation that promise disruption however in the end ship restricted industrial change. For many years, operators have confronted a well-known set of challenges, together with declining margins, rising operational complexity, and restricted differentiation.
This stagnation, says Calix CEO Michael Weening, has been evidenced within the little influence from every new era of wi-fi know-how.
“The business appears to be talking about the identical issues for 30 years,” he stated in an interview with Whole Telecom. “Each time we go from 3G to 4G or 4G to 5G, the business says that is going to be that holy second that modifications all the things. And it modifications completely nothing. We’re quickly whinging again to the identical issues.”
With the appearance of agentic AI, nevertheless, this might all be about to vary, permitting for not solely larger community effectivity however a completely new strategy to delivering personalised companies to prospects.
“I believe we’re at a pivotal second […] All the things’s coming to a head and there’s an actual alternative to vary,” stated Weening. “AI utterly modifications the way you perceive the subscriber and the experiences that you may ship to them.”
“That complete concept of those brokers and the way you construct them into workflows can be a profound shift,” he added.
A hyper-personalised service mannequin
It’s well-known that operators are a few of the most buyer data-rich organisations on this planet, however making use of that information stays a major problem. Regardless of a wealth of knowledge, the complexity of telco operations has sometimes seen suppliers defaulting to simplistic choices based mostly on worth and velocity. Clients are understood in a generalised manner, with mass campaigns concentrating on demographics and geographies relatively than people.
Now, nevertheless, the event of specialized AI brokers, which might automate decision-making throughout a number of workflows, is making personalisation potential at a person degree and at scale. Moderately than broad, undifferentiated campaigns, operators can goal people based mostly on real-time context, behaviour, and inferred wants.
Weening calls this the journey in direction of an ‘the expertise of 1’.
“Most communications service suppliers can take some demographic information and a few exterior information about prospects that may present that they, for instance, personal a pool. They may use that information to run a focused advertising marketing campaign providing an out of doors Wi-Fi bundle,” Weening defined. “The issue is that doesn’t essentially imply something, proper? The client may be a grandparent who solely makes use of the pool when their grandchildren come over.”
With subtle AI brokers constructed into every a part of a suppliers’ workflow, buyer understanding may be much more granular.
“Let’s say you do business from home. The community agent can discover that your service repeatedly degrades at a sure time of day – maybe when your youngsters come dwelling – and it will possibly let you know change areas. It is aware of you lately bought a pergola, so it will possibly infer you repeatedly transfer into the yard to proceed working,” Weening stated. “It might then move this info to a advertising agent, which can construct a personalised plan together with improved outside connectivity. Then, an promoting agent that is aware of your social media preferences will arrange a focused advert that you will notice on Instagram – and when you don’t click on by means of, it will possibly hand off all this particular information to a name centre agent and arrange a name robotically.”
A shift in direction of these experience-based choices presents a serious alternative for income era, with Weening noting circumstances of progress “as excessive as 25%.”
“The most effective factor is that these income beneficial properties are wildly sticky,” he stated. “When you ship this improved expertise to the shopper, it turns into troublesome for different suppliers to compete. We’ve seen NPS (Internet Promoter Scores) of 94 – that’s insane!”
Constructing a many-layered platform
After all, orchestrating this AI agent ecosystem isn’t any straightforward activity—and, in accordance with Weening, it requires a platform mannequin to succeed.
The structure crucial for this transformation, Weening defined, is essentially constructed on 5 layers:
- The info layer, which kinds the inspiration.
- The information and context layer, which mixes Retrieval Augmented Technology (RAG), Giant Language Fashions (LLMs), and the Mannequin Context Protocol (MCP) to construct context from the information.
- The orchestration layer, which ensures the assorted AI brokers work collectively precisely and effectivity.
- The belief layer, which ensures each AI agent has applicable guardrails and may solely entry particular information.
- The safety layer, making certain that each one this exercise retains client information secure.
Calix has been engaged on the “AI-native third era” of its platform for the previous two-and-a-half years and went stay with it in February 2026. As of March 2026, all the corporate’s 1,200+ prospects globally have been migrated to this newest model of the Calix One platform.
“What we’ve constructed is one thing that abstracts worth from context,” he defined. “This degree of personalisation is one thing we’ve all the time aspired to. With AI, we will lastly do it at scale.”
AI wants tradition change
With AI’s technical capabilities not unsure, remaining limitations to adoption are largely cultural and organisational. Operators are sometimes seen as labyrinthine and bureaucratic, strangled by legacy know-how and an incapability to maneuver rapidly. For Weening, solely these organisations that put off these dated stereotypes and embrace change will see success.
“Persons are already taking a look at AI like it’s magic fairy mud – one thing you possibly can simply plug in and see outcomes. It’s not. It’s a cultural and enterprise transformation,” he argued. “Management should embrace it and use it to empower workers.”
Agentic AI might lastly present telecoms operators with a reputable path to differentiation, however realising this chance will rely much less on the know-how itself and extra on operators’ capability to execute the mandatory organisational and cultural transformation at scale.
Study extra in regards to the AI-native Calix One platform: https://www.calix.com/merchandise/platform.html
Michael Weening – President and CEO of Calix Inc.
Michael has served as president and CEO of Calix since September 2022, bringing greater than 20 years of expertise main progress, technique, and transformation. Beforehand he served as Calix President and Chief Working Officer. Michael joined Calix from Salesforce the place he served because the senior vp of International Buyer Success and senior vp, Japan and Asia-Pacific Buyer Success, Companies and Alliances.
Over the span of his profession, Michael has held govt positions in North America, Europe and Asia. Beforehand, Michael held management roles at Bell Mobility in Canada, the place he was vp of enterprise and client gross sales. Michael additionally held gross sales management roles at Microsoft, in Canada and the UK.
Michael holds a Bachelor of Arts in Enterprise Administration, from Brock College supported by ongoing govt training at Queens, Wharton, and USC.