AI Agent Platform for a National Wholesale Fiber Network Operator

Client Profile

A national wholesale fiber infrastructure provider supplying connectivity to dozens of ISP partners across the United Kingdom. The organisation owns and operates a growing fiber network spanning multiple regions, providing the physical infrastructure over which its ISP partners deliver broadband services to end consumers and businesses.

The company occupies a critical position in the UK broadband value chain: it builds and maintains the network, while its ISP partners handle the retail customer relationship. This wholesale model means the operator’s success is directly tied to the success, satisfaction, and retention of its ISP partner ecosystem.

Industry: Telecommunications (Wholesale Infrastructure) · Region: United Kingdom · Products Used: VoiceFlow AgentIQ · White-Label AI Platform · Field Engineer Assistant

The Challenge

The operator faced three distinct but interconnected challenges, each affecting a different part of their business:

Challenge 1: ISP Partner Support Was Bottlenecked

ISP partners contact the operator for technical support on network outages, service provisioning delays, installation issues, and infrastructure capacity queries. These requests require immediate triage and routing to the appropriate engineering team — delays directly impact the ISP’s ability to serve their own customers.

The existing manual triage process was creating friction:

• Response times were inconsistent, varying from minutes to hours depending on which support agent was available and how accurately they classified the issue.
• Escalation protocols were informal, with different agents routing similar issues to different teams.
• SLA compliance was difficult to track and report, straining partner relationships.
• After-hours coverage was limited, meaning network issues occurring overnight — when maintenance windows typically run — went unaddressed until morning.

Challenge 2: No Scalable Revenue Beyond Connectivity

The operator’s commercial model was entirely based on selling raw connectivity — wholesale bandwidth at a per-connection price. As the fiber market matured and competition intensified, margins on pure connectivity were compressing. The operator needed to move up the value chain and offer services that their ISP partners would pay for on top of basic access.

Many of their ISP partners — particularly smaller operators — struggled with the same problem: customer support costs that didn’t scale. They were running lean teams, hiring agents who turned over quickly, and delivering inconsistent support experiences that drove customer churn. If the wholesale operator could solve this problem for them, it would create a new recurring revenue stream while simultaneously reducing churn among its ISP partners.

Challenge 3: Field Engineering Relied on Phone Calls to HQ

The operator employed field engineers and contractor teams who frequently encountered technical challenges during fiber installation and maintenance. Their current workflow for resolving on-site issues was to call the central technical support desk — the same desk handling ISP partner queries — and wait for guidance.

This created a double bottleneck: field engineers waiting on hold, and ISP partner queries being delayed because support agents were on calls with field teams. The technical knowledge needed to resolve most field issues existed in documentation, but it wasn’t accessible in a format engineers could use while up a pole or in a cabinet.

Our Approach

We deployed three purpose-built AI agents, each addressing one of the three challenges. Critically, all three agents share a common knowledge base and integration layer, meaning intelligence gathered in one context informs the others. A network fault identified by the field engineer agent, for example, automatically updates the information available to the ISP partner support agent.

What We Built

Agent 1: NetworkAssist — Wholesale Partner Support Triage

A 24/7 AI voice agent serving as the first point of contact for ISP partner technical queries. NetworkAssist handles:

• Intelligent incident classification — categorising inbound queries by type (outage, provisioning, capacity, installation), severity, and affected infrastructure. The agent achieves 90%+ classification accuracy, ensuring issues reach the right engineering team immediately.
• Automated provisioning status checks — pulling real-time data from the operator’s provisioning systems to answer “where is my order?” queries instantly, without requiring human intervention.
• Guided troubleshooting — walking ISP partner engineers through diagnostic workflows for common issues, using the operator’s technical runbooks translated into conversational guidance.
• SLA-aware escalation — tracking response time commitments per partner contract tier and prioritising accordingly. If a Tier-1 partner’s issue is approaching SLA breach, escalation is automatic and immediate.
• Real-time network status — accessing the operator’s network monitoring systems to provide live outage updates, planned maintenance information, and estimated restoration times.

Every interaction is logged with full context, creating a complete audit trail for SLA reporting and partner relationship management.

Agent 2: SupportAgent — White-Label Customer Support Platform

A modular, brandable AI customer support platform that ISP partners can deploy for their own end customers. The operator sells this as a managed service — a new revenue line that didn’t exist before.

Each ISP partner gets:

• Branded experience — the agent identifies itself with the ISP’s brand name, uses their tone of voice, and follows their specific support workflows. End customers interact with what feels like the ISP’s own support team.
• Fault management — handling the most common end-customer queries: “my internet is down,” “my speeds are slow,” “I can’t connect a new device.” The agent performs automated diagnostics, walks customers through troubleshooting, and creates tickets for issues requiring field intervention.
• Billing and payment support — processing payment queries, explaining charges, and managing account updates through integration with the ISP’s billing system.
• Multi-channel deployment — available via voice, web chat, and WhatsApp, meeting customers on their preferred channel.
• Shared learning — patterns identified across all ISP deployments (e.g., a common router firmware issue causing widespread connectivity problems) are surfaced to the operator, enabling proactive network-level remediation.

The platform supports multiple concurrent ISP deployments from a single instance, with each partner’s data strictly isolated.

Agent 3: FieldAssist — Field Engineer Support

A voice-activated AI assistant designed for engineers working in the field. FieldAssist provides:

• Real-time troubleshooting guidance — engineers describe the issue they’re facing (verbally, hands-free), and the agent provides step-by-step resolution guidance drawn from technical documentation, previous incident records, and equipment manuals.
• Equipment identification — engineers can describe or identify the hardware they’re working with, and the agent provides relevant specifications, configuration procedures, and known issues.
• Remote diagnostic assistance — the agent can query network management systems on the engineer’s behalf, checking signal levels, port status, and configuration parameters without the engineer needing to call the central desk.
• Safety protocol reminders — for high-risk activities (working at height, in confined spaces, near live electrical equipment), the agent provides relevant safety checklist items as part of the troubleshooting workflow.
• Incident logging — the agent captures a structured record of the work performed, parts used, and issues encountered, writing this back to the ITSM system automatically.

Projected Impact

Metric	Target
ISP partner query response time	70% reduction through AI-first triage
Issue classification accuracy	90%+ on first contact
24/7 support availability	Full coverage (from business hours only)
White-label platform deployments	3 ISP partners in pilot phase
End-customer query resolution rate	85% without human handover
ISP partner support costs	60% reduction for participating partners
Field engineer dependency on central desk	Significant reduction through FieldAssist self-service
New revenue stream	Recurring SaaS revenue from white-label platform

Why This Matters

This engagement transforms a wholesale infrastructure provider from a pure connectivity supplier into a platform business. The white-label AI support product creates a recurring revenue stream that improves with scale — each new ISP partner deployment adds data that makes the system smarter for all partners.

More strategically, it deepens the operator’s relationship with their ISP partners. A partner using the operator’s network, the operator’s support platform, and the operator’s analytics is far less likely to migrate to a competing wholesale provider than one who is only buying bandwidth.

The three-agent architecture also creates an information feedback loop: field issues inform partner support, partner support patterns inform network planning, and network intelligence informs field operations. The operator doesn’t just have three chatbots — they have an integrated intelligence layer across their entire operation.