Agent adoption is accelerating — and the data is beginning to show where value emerges

Aggregated Agentforce usage data shows more active agents, faster creation and two distinct patterns for operational scale.

OrbTrail10 min readSalesforce NewsUpdated September 8, 2026

Editorial network of agents and performance signals growing across business sectors.

Original OrbTrail analysis expanded with complementary research, practical context and verified references.

The new Agentic Enterprise Index tracks aggregated, anonymized Agentforce usage from February 2025 through April 2026. The cohort is not a universal benchmark and does not prove causation, but it offers a rare view of how agents move from pilots into operations.

By the end of the period, organizations had nearly tripled activated agents while average creation time fell 53%. During demand peaks, some agents expanded their skill sets by as much as 350%, suggesting that scale means greater specialization as well as more conversations.

The data points to two paths. Retail and consumer services favor high volumes of narrow actions; manufacturing and regulated industries lean toward more complex multistep workflows. Both require teams to track resolution, quality, cost and escalation—not merely the number of deployed agents.

COMPLEMENTARY RESEARCH

3 Perspectives on the topic

What the index actually measures

The index uses aggregated Agentforce telemetry from February 2025 through April 2026. It describes the observed cohort—activated agents, build time, employee use and skills—but it is neither a market-wide sample nor proof that agents alone caused the reported business outcomes. Readers should therefore treat its growth curves as directional evidence from customers already using the platform, then validate the pattern against their own process baselines, adoption mix and operating costs.

Two scaling models

Retail tends to gain from high volumes of short, repeatable interactions; manufacturing and regulated industries derive value from fewer but longer, context-heavy journeys. Architecture, testing and unit economics must reflect that difference instead of copying one playbook.

The metric missing from the dashboard

Counting deployed agents or handled conversations does not measure value. An executive dashboard should connect completion without reopening, escalation rate and reason, cost per outcome, satisfaction and variance against a human baseline. Only then can usage growth indicate maturity.

Agent adoption is accelerating — and the data is beginning to show where value emerges

3 Perspectives on the topic

What the index actually measures

Two scaling models

The metric missing from the dashboard

Continue reading

AI agent pricing is beginning to move from consumption to outcomes

AI surges in field service, but a talent crisis threatens returns

Measuring AI use requires more governance than a leaderboard