Business adoption of AI agents tripled this year – as measurable ROI emerges

Comply with ZDNET: Add us as a preferred source on Google.

ZDNET’s key takeaways

• The variety of energetic AI brokers in organizations has tripled within the final 12 months. 
• AI brokers have improved their capabilities by 350% and may now deal with advanced duties. 
• Worker use of AI brokers has elevated 3X as belief deepens. 

The typical variety of AI brokers activated per group elevated almost threefold since 2025, in line with the 2026 Agentic Enterprise Index, new analysis from Salesforce that analyzes combination AI utilization knowledge from the corporate’s Agentforce platform.

The index analyzes the exercise of AI engagements in manufacturing for 5 consecutive quarters of 400 actual companies leverage AI brokers to drive ROI. As well as, the index report features a survey of almost 5,000 respondents throughout 9 key markets. 

Additionally: ‘Specialists aren’t required’ anymore: How to stay valuable in an AI agent workplace today

The most important takeaways from the Agentic Enterprise Index report are:

• The size of agentic AI adoption is accelerating, quick. The typical variety of brokers per group almost tripled (going from 5 to 13), whereas creation time dropped by 53% to a median of 1.9 days to create an agent. 
• Individuals are beginning to belief brokers — quite a bit. Since Feb. 2025, there’s been a threefold surge in workers’ weekly variety of initiated agent classes.
• Industries are discovering the methods that work greatest for his or her enterprise wants: Client-facing industries like retail and journey lead within the velocity of deployment and using brokers general, particularly throughout peak-demand seasons like the vacations. Regulated industries have been slower to undertake and implement brokers, and their % of  Agentic Work Items (AWUs) — a time period outlined by Salesforce that represents one discrete unit of labor accomplished by an AI agent — within the final 12 months is decrease, however the brokers themselves are sometimes way more subtle.

Scaling AI agent adoption 

Enterprise AI brokers have moved from experimental assessments to speedy, production-scale deployment. The typical variety of AI brokers in manufacturing grew from 5 in February 2025 to 13 brokers in April 2026. Along with increasing the use AI brokers, companies have additionally elevated effectivity when it comes to how lengthy it takes so as to add a brand new agent to manufacturing. The time to create a brand new agent has decreased by 53%, going from 4 days in early 2025 to 1.9 days at this time. 

AI brokers have gotten smarter and extra succesful. Companies noticed a 31% improve within the compound month-to-month development charge of brokers throughout 5 quarters. The typical share of agentic motion monthly grew from zero in early 2025 to fifteen% by April of 2026. The steady-state development of agentic actions reveals the continued enchancment in agentic abilities and of their capacity to execute and handle advanced duties. 

Additionally: The 3 types of people who will excel in the AI agent era, according to tech leaders

AI brokers have gotten smarter and quicker

The index report revealed that AI brokers are evolving past easy textual content technology and may now execute advanced, multi-step enterprise logic throughout numerous techniques and limits. In 2025, the common variety of distinctive actions per agent was 2; at this time, that quantity is 4. Within the retail business, the height common distinctive motion per agent is 9, which interprets to shopper expectations rising to agentic scale intelligence, velocity, and scale throughout all touchpoints.

The expanded ability set of AI brokers additionally means they’ll now tackle secondary roles. Brokers are dealing with a higher scope than earlier than and turning into extra able to dealing with cross-functional exercise over time. The share of secondary features grew for 1% to six% of features previously 12 months. 

Additionally: Token-maxing is an AI cost sink – how to use agents without busting your budget

The index report discovered that AI brokers are additionally turning into extra proactive. The compound month-to-month development charge (CMGR) of the action-to-output-token ratio grew by 15% month-over-month. Brokers are doing extra of the work and the work is turning into extra advanced and cross-functional.

So are the brokers turning into extra environment friendly as they tackle extra work? One solution to reply this query is to check AWUs. (When an agent causes by means of, comes to a decision, and takes motion, then one AWU is consumed.) Salesforce noticed 734 million AWUs consumed in April 2026, representing a 15% month-over-month improve within the action-calls-to-output-token ratio. 

AI brokers are delivering measurable ROI

The index report discovered that autonomous scaling is delivering measurable ROI whereas enhancing worker adoption and driving buyer satisfaction. Workers are participating with AI brokers greater than ever earlier than. Weekly consumer classes with AI brokers have tripled year-over-year. The typical weekly classes between AI brokers and workers is approaching 8 classes per week.

Additionally: Moving from AI pilots to business-wide value requires a superhighway – how to ramp up

Despite the fact that AI agent interactions are tripling in quantity YoY, the variety of buyer escalations stays regular. In early 2024, zero customer support classes had been dealt with by autonomous brokers. Immediately, 7 in 10 customer support classes are dealt with autonomously by AI brokers. As of August 2026, over 5 million conversations at help.salesforce.com had been dealt with by AI brokers, in comparison with 2.4 million by human brokers. The expanded use of autonomous AI brokers transcend customer support. The index discovered four-fold improve in gross sales when shopper brokers are current. As well as, 77% of buyers report higher confidence after interacting with brokers. Workers and prospects are deepening their belief in AI brokers. 

Business adoption of AI brokers is accelerating 

Enterprise adoption of autonomous AI brokers splits into two distinct operational methods based mostly on regularity and workflow necessities. The index discovered that retail and journey industries are scaling for end-of-year demand. The 2 industries had 60% surge in agent output from November 2025 to January 2026. The retail consumption of AWUs grew by 18X (22% of whole month-to-month output) in early 2026, whereas the journey business noticed a 7X improve (10% of whole month-to-month output). The general public sector business noticed a 227X development whereas monetary service noticed a 13X improve. 

The index report revealed that regulated industries require higher AI agent sophistication. Degree 1-3 AI brokers are anticipated to learn, coordinate, and synthesize, whereas Degree 4-5 AI brokers are capable of write, analyze, and parse. 

Additionally: 70% of companies deploying customer service AI agents see ROI in 60 days

The Agentic Enterprise Index report from Salesforce highlighted three key learnings: 

1. Brokers can do greater than you assume. Companies can construct a extra versatile, interconnected digital workforce by trusting brokers to tackle adjoining obligations exterior their unique scope.
2. Leverage seasonal surges with brokers. Notably in shopper industries, agent assist grows throughout busy occasions like holidays or tax season to drive top-line enterprise income. 
3. Combine brokers into every day channels. The information exhibits workers are participating in additional classes with brokers than ever earlier than, significantly in communication channels they use all through the day.

To be taught extra concerning the Agentic Enterprise Index report, you may go to here.