Architecting autonomous control for global B2B telecom.
Designing the unprecedented self-service platform that let tier-1 multinational enterprises manage their global networks, adjust bandwidth on demand, and resolve issues without ever picking up the phone — for a major US telecommunications carrier serving the world's largest companies.
What was at stake.
For the world's largest companies, managing a global telecommunications network was a slow, friction-heavy process — built on manual support, siloed channels, and dashboards that surfaced data instead of decisions.
The carrier was the backbone of communications for some of the most demanding enterprise customers on earth — multinationals running mission-critical networks across dozens of countries, on whose continuity entire businesses depended. But the experience of being one of those customers had not kept pace with the sophistication of the network itself. Routine changes — adjusting bandwidth, reconfiguring a regional link, troubleshooting a dropped circuit — required a phone call, a ticket, an email thread, and a service representative to broker each transaction. Static, generalized dashboards offered no personalized intelligence. The most strategically valuable customers in the portfolio were operating their global networks through the slowest interface in the carrier's stack.
Three forces made the status quo untenable. Cost-to-serve — every enterprise customer was an expensive, high-touch relationship, and that economics did not scale. Competitive pressure — challengers were marketing self-service and on-demand control as table-stakes capabilities, and incumbents that could not match them were leaking accounts. And customer expectations — the network managers running these accounts were the same people who, in their personal lives, expected to provision a cloud instance in minutes; the contrast between consumer software and their enterprise telecom experience had become impossible to ignore.
The business stakes were direct. Tier-1 multinational accounts represent disproportionate revenue concentration in any carrier's book — losing one is a quarter; losing several is a strategy. Leadership had committed to a transformation that would shift the carrier's posture from reactive support to proactive, intelligent self-service, and the experience layer was the visible face of whether that strategy was real.
Bringing EXA to the problem.
The transformation required a strategic shift away from digitizing manual processes and toward architecting an intelligent, modular ecosystem. Every decision was decomposed through the three pillars of Enterprise Experience Architecture — scale, intelligence, and clarity — to deliver a platform that anticipated the needs of network managers rather than waiting to react to them.
A modular framework for a global B2B suite
A multinational enterprise customer base, operating across dozens of regulatory regimes and time zones, cannot be served by a monolithic portal. We architected a deeply tokenized Enterprise Design System and an interoperable suite of B2B applications — modular by design — so that capabilities could be assembled, localized, and rolled out independently while every surface inherited the same architectural DNA.
Anticipating AI before it was the trend
Long before AI-first was a category, we designed the platform around proactive, automation-led decision support — surfacing insightful, "next best actions" for network managers, automating routine diagnostics through smart self-service tools, and routing customers to the right resolution path before they had to ask. Intelligence was the operating model, not an add-on.
Designing for humans, not a corporation
Network managers spend their day inside dense operational data — configurations, telemetry, alerts, billing. We engineered the experience to translate that complexity into elegant, actionable interfaces and to dismantle every artificial silo between channels, so that a customer who started a question in a chatbot could carry the conversation into a phone call without losing context or repeating themselves.
What was designed and deployed.
A unified B2B self-service ecosystem — anchored by an Enterprise Design System and three integrated capability layers that together gave tier-1 enterprise customers unprecedented visibility, control, and autonomy over their global networks.
The Enterprise Design System was the architectural foundation. We built a centrally governed, deeply tokenized component library aligned one-to-one with the engineering team's modular front-end architecture — so every screen across the B2B suite, from billing to bandwidth to ticketing, inherited consistent interaction patterns, accessibility behavior, and visual language. The design system was the reason the portfolio could feel like one product even as it shipped on independent release cycles, and the reason new capabilities could be designed in days rather than months.
Omni-channel trouble management dismantled the communication silos that had defined the legacy support experience. A network manager could initiate a query through an automated chatbot, hand off to a live agent, or escalate to a phone call — without ever restarting the conversation, re-verifying their identity, or repeating the technical context. Every touchpoint shared the same case state, the same diagnostic history, and the same view of the customer's network. The boundary between channels disappeared from the user's point of view, even though the routing intelligence underneath was substantial.
Dynamic on-demand bandwidth controls turned the network from a fixed contract into a living instrument. Customers could boost bandwidth in real time when their business needed it — to accommodate a product launch, a high-traffic event, a quarterly close — and scale back when they didn't. The capability did more than reduce friction; it actively drove revenue. Customers who experienced the flexibility of dynamic provisioning consistently expanded their consumption, because the product had begun to adapt to their business in ways the legacy contract model never could.
Intelligent insights and next-best actions replaced the static, one-size-fits-all dashboards that had defined the prior generation. The new interface transformed raw network telemetry into personalized alerts, proactive recommendations, and contextual diagnostics — telling the network manager not just what was happening, but what to do about it. When a circuit was degrading, the platform offered a remediation path. When usage patterns suggested an opportunity to renegotiate, the platform surfaced it. The dashboard stopped being a window onto data and started being a partner in operating the network.
What changed.
Impact was measured across four dimensions — mean time to repair, first-contact resolution, customer satisfaction, and the commercial uplift driven by self-service adoption — the metrics that determine whether a B2B telecom transformation is real or rhetorical.
200% improvement in Mean Time to Repair. Smart self-service diagnostics let customers test and correct common issues autonomously — and when an issue genuinely required carrier intervention, the omni-channel routing and shared case context cut the handoff overhead that had historically dominated the resolution clock. Issues that had taken hours to triage were closing in minutes.
130% improvement in First Contact Resolution. Intelligent routing matched customers to the right resolution path on the first interaction — the chatbot when it was sufficient, the right agent when it wasn't. Crucially, when escalation did happen, the receiving agent already had the full conversational and technical context, eliminating the "let me transfer you" cycle that defined the legacy support experience.
+0.5 NPS point increase. A half-point on the Net Promoter Score sounds modest until you remember the size of the customer base — these are tier-1 multinational accounts, and the score is famously hard to move. The lift was a tangible, measured signal that the experience had crossed a threshold from tolerable to genuinely valued.
The deeper story behind the metrics. Initially, one signal looked like a problem: the raw number of trouble tickets actually increased after launch. Surface-level reading would have called that a failure. Deeper experience analytics told the real story — the new platform had simply given customers a frictionless way to log issues that had previously been silently absorbed or never reported, while resolving them dramatically faster than before. What looked like more problems was in fact more visibility into the same problems, paired with better resolution. The architecture was working; the measurement model needed to catch up.
Why this matters.
Telecom is one of the hardest places to make self-service work. The data is dense, the customers are sophisticated, the regulatory and contractual constraints are heavy, and the cost of getting it wrong is measured in churned tier-1 accounts. When intelligent self-service works in this environment — and it works well enough to move MTTR by 200%, FCR by 130%, and NPS measurably upward — it is not a coincidence of effort. It is the product of architecture.
This engagement validated EXA as a discipline that scales beyond any single industry. The same three pillars that unify a global energy operator or stand up a healthcare SaaS platform — scale, intelligence, clarity — translate directly into a B2B telecom self-service ecosystem. The premise holds: at the scale of complex digital ecosystems, the product is the system. You cannot patch a self-service portal onto a fragmented support model and expect the customer to feel confidence; you have to architect the underlying experience to be coherent across every channel before any one screen can earn the customer's trust.
It also validated something subtler. The platform anticipated the AI-first wave by years — not because the team predicted the trend, but because the EXA discipline starts from the question "what should the system decide on the user's behalf?" rather than "what controls should we expose?" That question, asked rigorously enough, leads to proactive intelligence as a natural conclusion. The technology to deliver it has only gotten easier since. The architectural framing — designing for autonomous control, intelligent insights, and humans rather than corporations — is the durable part. Do the hard work to make it simple. That principle is what makes the rest of it work.