Why this matters now
For telecoms operators, the contact center has long been the default channel for everything: billing queries, service activations, troubleshooting, complaints. The model works until it doesn't - and the point at which it stops working tends to arrive at the worst possible moment: during a surge, a service outage, or a period of rapid subscriber growth
The structural problem is that voice-based service does not scale gracefully: more calls mean more turnover. This is exactly where AI Agents change the equation: by absorbing the repetitive, high-volume requests, they free human agents to focus on the complex, high- value conversations where they make a real difference. For operators with hundreds of thousands or millions of subscribers, AI-assisted agents turn what was a business model constraint into a scalable operation.
Digital-first customer service changes the equation. When a well-built AI agent can handle three quarters of all inbound interactions autonomously, the contact center stops being a bottleneck and starts being a precision instrument, reserved for the cases that genuinely require human judgment, empathy, or access to systems the agent cannot reach. Thus, the human agents are doing more of what they are actually needed for.
The company's goals were to:
- Reduce voice call volume: Shift the majority of routine enquiries from phone to digital self-service without degrading customer experience.
- Scale self-service capability: Build an AI agent capable of handling activations, subscriptions, complaints, and knowledge base queries autonomously.
- Empower human agents: Give contact center staff an AI assistant that fills the gaps left by CRM limitations and third-party integrations.
- Optimise workforce: Redeploy agent capacity freed by automation into revenue-generating outbound activity rather than reducing headcount.
The Challenge
Before deploying the AI Agent, the telecom operator's contact center was receiving over one million voice calls per month. The digital channel - RingCentral for social and messaging - was handling over 100,000 interactions.
Managing this volume required around 800 part-time agents - a staffing strategy designed to maintain quality through rotation and fresh engagement. But the model carried significant overhead: annual turnover of around 70% meant onboarding 30 to 50 new agents every month, each needing a full month of training before handling calls independently.
At a certain point, it oversaturated:
- Inbound call volume had reached an operational ceiling, and the team could not absorb further growth.
- The digital channel was underutilised relative to its potential, accounting for only 12% of total interactions despite the infrastructure being in place.
- Third-party service integrations were blocked by costly and slow CRM change requests, limiting what agents could do for customers without manual workarounds.
The Solution
The telecom operator deployed a Druid AI-powered customer agent, to transform how subscribers interact with the company across digital channels.
The deployment was intentionally phased: the A AgentI launched in late 2024 with extended knowledge base functionality and was systematically expanded over twelve months to version Laila 2.0 by adding transactional capabilities, additional language support, and deeper system integrations before the team was confident enough to actively redirect voice customers to the digital channel.
That patience proved decisive.
The solution includes several core capabilities:
- An Omnichannel Customer AI Agent
Active across WhatsApp (80% of volume), mobile app (16%), Facebook Messenger (2.4%), and the company website, the AI Agent handles service information, bundle activations, subscription changes, ticket creation, complaint handling, and knowledge base queries, covering approximately 70-75% of what the call center needed to solve. - Conversational Flow Designed to Feel Human
Monitoring of real conversations confirmed that many customers completed full service transactions without realising they were talking to an AI - a benchmark the team set deliberately and measured consistently. - Workforce Redeployment Through Outbound Automation
Agents freed from routine inbound volume were not let go. Instead, an automated campaign system identifies idle agents and connects them to outbound customer calls in real time, generating thousands of outbound calls per day from the same operator.
Learnings
Readiness before redirection
Asiacell spent twelve months building and tuning the AI Agent before actively redirecting voice customers to digital. By waiting until containment was genuinely high, the team ensured that when redirection started, customers stayed.
Redeployment multiplies the return on automation
When the AI Agent absorbed routine inbound volume, the telecom company redirected freed agent capacity into outbound revenue campaigns. The contact center gained a revenue-generating capability from the same team, turning efficiency into growth.
Speed of iteration is a competitive moat
Taking an idea from conversation to production in 24 to 48 hours becomes a strategic advantage. In telecoms, the operator that responds to customer needs fastest wins.
Quality earns adoption
Digital share grew from 12% to 43% because conversations felt human enough that customers completed transactions without noticing they were talking to an AI. Adoption grew organically, driven by the strength of the experience itself.
