Share with your CIO
PLDT, the Philippine telecom and digital services group, is betting that agentic AI, meaning AI systems that can plan and execute multi-step tasks without constant human direction, will close the gap between operational volume and staff capacity. Working with UiPath, PLDT has deployed three named systems: KAI for instant knowledge retrieval, Ellie for customer-service automation cutting manual effort by 25 to 35 percent and saving an estimated 18,000 to 25,000 staff hours annually, and ERICA for automated risk scoring and treatment recommendations. Response times are reported at up to 80 percent faster.
What this means for your business
The interesting signal here isn’t the UiPath partnership itself, it’s that PLDT is deploying agentic AI across three distinct domains simultaneously: knowledge management, customer operations, and enterprise risk. Most large-enterprise AI rollouts still pick one beachhead and consolidate before expanding. If your own roadmap is sequenced that way, PLDT’s multi-domain push is worth watching as either a cautionary tale about overextension or early evidence that a unified orchestration platform makes parallel deployment viable.
The 18,000 to 25,000 hours saved figure is where scrutiny belongs. That range is wide enough to reflect genuine uncertainty, which is actually a mark in its favor compared to the suspiciously round numbers that usually populate vendor announcements. But hours recaptured only translate to cost reduction if headcount adjusts or those hours flow into higher-value work. PLDT hasn’t characterized what happens to the freed capacity, and that gap matters when you’re defending a similar automation investment to your CFO. The UiPath framing, published through a trade outlet that sells into exactly this audience, has an incentive to headline the hours number without the reallocation caveat.
The enterprise risk use case, ERICA synthesizing data and scoring risks automatically, is the least-discussed element here and probably the highest-stakes one for peer organizations. Automating risk scoring touches audit trails, regulatory accountability, and model explainability in ways that customer-service bots simply don’t. If that capability works cleanly, it changes the renewal calculus on legacy GRC platforms for any CIO whose CISO is already asking whether their risk tooling is keeping pace with operational complexity.
Concept deep-dive: Agentic AI
Agentic AI refers to systems that don’t just respond to a single prompt but pursue a goal across multiple steps, calling tools, checking results, and adjusting course along the way. Think of the difference between a calculator and an assistant who books your travel, confirms the hotel, and flags a scheduling conflict without being asked each time. In enterprise automation, the business relevance is that it removes the human-in-the-loop bottleneck from workflows that are too complex for simple rules but too repetitive to justify analyst time.
Based on reporting from PLDT advances AI transformation with UiPath, originally published 2026-09-11 02:00:00.
