← Cases
PhilippinesTechnology and telecom2026 – ongoing

PLDT

KAI, Ellie and ERICA agentic AI assistants on UiPath

PLDT said three UiPath-based AI assistants cut research and risk-review work from days to seconds or hours

WorkingVendor case studyVendor ↗

Problem

PLDT staff spent up to five days researching answers for proposals and customer support, and annual risk reviews took two to ten days per evaluation. Demand for 5G, broadband and satellite services was rising.

Approach

PLDT built three assistants on the UiPath platform. KAI answers staff questions from approved sources only, for proposal work and customer support. Ellie is a digital assistant for PLDT Enterprise teams. ERICA, an agent deployed in February 2026, pulls together data for risk reviews, scores risks and recommends treatments.

Results

25,000 to 30,000 hours a yearestimated manual effort removed by KAI, about 12 full-time equivalents, per UiPathVendor case study[1]
18,000 to 25,000 hours a yearestimated time saved by Ellie, per UiPathVendor case study[1]
5 minutes to 1 daytime for ERICA risk evaluations, down from 2 to 10 days, per UiPathVendor case study[1]
1 to 3 secondsKAI response time, versus research that took up to five days, per UiPathVendor case study[1]

Takeaway

The savings were published by the software supplier as estimates, with no measurement period, user numbers or note on whether the hours saved by KAI and Ellie overlap.

Categories

Innovation type
Process innovation
Evidence
Vendor case study

Sources

Source summaries

[1]UiPath11 Sep 2026

UiPath's release described KAI, Ellie and ERICA and gave estimated savings: 25,000 to 30,000 hours a year for KAI, 18,000 to 25,000 hours for Ellie, and risk evaluations cut from 2 to 10 days to 5 minutes to 1 day for ERICA. It also said KAI answers in one to three seconds instead of up to five days of research.

Full details are in the original sources.

Added 2 Nov 2026 · Updated 2 Nov 2026
Open case page ↗