PLDT has deployed three UiPath AI systems across its internal workflows, cutting manual research time from days to seconds and reducing risk assessment processing by up to 99%. The rollout addresses growing pressure on the Philippine telco to deliver faster customer support as demand for 5G, broadband, and satellite connectivity services climbs.
The three tools-KAI, Ellie, and ERICA-target distinct areas of manual work within the company. UiPath reported that the deployments have driven substantial reductions in research time, employee effort, and risk processing durations.
KAI and Ellie: cutting research time and connecting workflows
KAI, which stands for Knowledge, Automation, Intelligence, functions as an internal AI assistant for proposal creation and customer service research. It draws exclusively from approved information sources and provides contextual responses to employee queries. Research that once took up to five days now completes in one to three seconds, a 99% improvement in response speed. UiPath estimates KAI eliminates between 25,000 and 30,000 hours of manual work annually, equivalent to roughly 12 full-time employees.
Ellie, the Enablement and Intelligence Engine, connects employees with information and workflows across the organization. It reduces manual effort by 25% to 35% and saves an estimated 18,000 to 25,000 hours each year. Customer response times improved by up to 80%. While KAI shortens the time needed to find approved information, Ellie links fragmented knowledge and workflows so employees can act on that information quickly.
For customer support teams, these tools directly address the bottleneck of repetitive information gathering. The shift moves agents away from manual lookups and toward consultative work with customers. Readers tracking AI for Customer Support will recognize this pattern: automation handling retrieval so humans handle judgment.
ERICA: accelerating enterprise risk reviews
ERICA, the Enterprise Risk Intelligence Companion Agent, launched in February 2026 to support PLDT's annual risk reviews. It aggregates information, scores risks, and recommends treatments. Assessments that previously required between two and 10 days now finish in five minutes to one day. UiPath reports this reduces manual effort by 97% to 99%, freeing employees to focus on deciding how to address identified risks rather than gathering the data needed to assess them.
Why service speed matters now
Karl Crowther, Area Vice President, Middle East & Africa and South Asia at UiPath, tied the deployment to rising expectations in the Philippines' digital economy. "Customer expectations continue to rise alongside the growth of 5G, broadband, and digital services," Crowther said. "This has created increased pressure on communications providers to deliver faster, more seamless support experiences."
The AI Agents & Automation integrated into PLDT's systems target service restoration bottlenecks and aim to give teams faster access to the information they need during customer interactions. Gilbert Gaw, PLDT and Smart First Vice President and Head of IT and Transformation Office, said the partnership is intended to improve operations while supporting the company's growing digital services.
Why this matters for customer support professionals
PLDT's numbers offer a concrete benchmark for what AI-assisted knowledge retrieval can deliver in a live support environment: five-day research tasks compressed to seconds, and customer response times cut by up to 80%. For support leaders building a business case, these figures translate directly to capacity planning. A tool that eliminates 25,000 to 30,000 hours of manual work annually changes staffing models and lets teams redirect effort toward complex customer conversations rather than information gathering. The distinction between KAI and Ellie also matters-one finds the information, the other connects it to action. Both pieces are necessary for the speed gains to reach the customer.
Your membership also unlocks:
