The era of basic conversational chatbots is rapidly drawing to a close, giving way to autonomous digital workforces capable of executing complex, end-to-end financial transactions.
For a mobile-first country like the Philippines — where millions rely daily on e-wallets, digital banking apps, and online remittances — the shift from simple text bots to action-oriented AI agents marks the next big frontier for local financial inclusion and customer service.
According to the 2026 Agentic Enterprise Index released by enterprise software giant Salesforce, global deployments of AI agents have more than doubled over the past year.
Analyzing aggregated usage data across its Agentforce platform from February 2025 to April 2026, the report highlights a broad cross-industry shift: across retail, manufacturing, public services, and finance, enterprises are moving away from passive text-generating models toward action-oriented AI capable of directly triggering back-end workflows, updating database fields, and routing complex inquiries.
Divergent deployment strategies: High volume vs. deep integration
IMAGE CREDIT: Salesforce
The index reveals a clear bifurcation in how global enterprises deploy autonomous digital agents, largely shaped by their operational DNA and regulatory environment.
In consumer-facing sectors such as retail, deployments prioritize speed and high volume to resolve immediate customer needs. While retail agents often start narrow—averaging one to two actions most of the year—their versatility expands dramatically during peak surges.
Global jewelry brand Pandora, for instance, used its AI concierge Gemma to resolve 60 percent of routine inquiries and boost its NPS by 10 percent during high-traffic shopping periods, with average retail agents expanding to manage up to nine skills as holiday demand spiked.
Conversely, heavily regulated and operationally complex fields lean into versatile, multi-step deployments focused on building sophisticated agent networks capable of performing cross-functional tasks.
Industrial giant Siemens demonstrated this by orchestrating a multi-agent workflow to handle lead qualification across 7 business units and 18,000 sellers. Across similar complex environments, such as the public sector and healthcare, agent task volume saw massive surges — growing 227-fold and 19-fold, respectively.
Financial services blends scale with high governance
IMAGE CREDIT: Salesforce
While retail platforms deploy high volumes of task-specific agents during holiday rushes, the financial services sector stands out by deploying highly complex, multi-action agents at massive scale—representing roughly 10 percent of total monthly agent task volume globally.
In banking and credit unions, where customer interactions require stringent identity verification, audit trails, and legal oversight, financial institutions are deploying agents behind secure login portals to execute real banking actions rather than simply serving static FAQ answers.
This trend carries heavy implications for Philippine banks and fintech providers navigating Bangko Sentral ng Pilipinas (BSP) compliance while scaling digital support.
A prime example highlighted in the report is Virginia-based PenFed Credit Union, which deployed an AI agent named Ace secured behind its online banking login. Rather than functioning as a standard chatbot, Ace evaluates account balances, checks loan application statuses, transfers funds, and delivers answers sourced directly from curated internal knowledge bases.
To extend these capabilities to voice channels, the institution introduced a secondary agent, Echo, designed to replace legacy interactive voice response (IVR) systems.
“Bringing that vision to life meant going beyond simple chatbots to build sophisticated agent experiences with trust baked in,” noted Shree Reddy, Chief Information Officer at PenFed. “By pairing robust governance with our unified platform, we’ve safely deployed multi-action agents like Ace and Echo that perform real, complex banking tasks.”
Measuring real-world execution: The Agentic Work Unit
IMAGE CREDIT: Salesforce
To evaluate whether AI deployments deliver tangible operational value, the report tracks the Agentic Work Unit (AWU) — a metric measuring discrete tasks successfully executed by an AI agent, such as rebooking appointments, processing refunds, or fetching live credit scores.
Across all sectors, AWU output is growing at a compound monthly rate of 15 percent, backed by an architecture where agent logic is decoupled from front-end interfaces to process tasks anywhere.
The index noted that the average agent today can act on six distinct skills — such as parsing raw data inputs or writing back to enterprise systems — up from just two skills at the start of 2025. As Salesforce President of Enterprise AI and Technology Joe Inzerillo noted, the ROI is showing up directly in execution efficiency as systems move from passive assistance to active execution.
Strengthening user trust and the impact on Philippine digital finance
As autonomous agents demonstrate higher precision, enterprise and consumer trust in digital workforces is deepening.
Internal adoption is surging — with Salesforce reporting that 83 percent of its own workforce regularly uses its internal Slackbot to save up to five hours a week — while consumer confidence is following suit.
Over the five quarters analyzed globally, AI agents handled 170 times more customer service interactions than in previous periods, successfully resolving 7 out of 10 inquiries without requiring human intervention.
For the Philippine fintech ecosystem — where customer support bottlenecks and scam prevention remain critical priorities — the transition toward governed, multi-action AI agents offers a roadmap.
As local banks and e-wallets look to scale their operations without compromising security, execution-driven agents could soon become a staple of everyday Philippine digital finance.
