Optimizely says AI agent use jumps 42% as adoption grows

Annual recurring revenue for Optimizely's AI agent orchestration platform grew 42% quarter on quarter, and nearly 1,700 customers now use its Opal product.

Those customers have built more than 4,000 AI agents and run them more than 172,000 times across experimentation, campaign execution, content production and reporting. Customer-built agents accounted for 97% of that activity, while 32% involved agents handling multi-step tasks.

The figures point to broader use of AI tools inside marketing and digital teams, as companies try to move beyond idea generation and automate more of the work needed to complete campaigns and tests. Opal's weekly user base has doubled over the past year, with stronger uptake among large organisations.

Shafqat Islam, President at Optimizely, said customers were now focused on wider deployment rather than initial adoption.

"The marketers we talk to aren't asking whether AI belongs in their workflow anymore. They're asking how to scale it," said Shafqat Islam, President, Optimizely.

"The data we're sharing today reflects that shift. Customers are building agents that run repeatedly across campaigns, content and experimentation, and seeing measurable gains across the department, not just individually," Islam said.

Workflow shift

The data suggests marketing teams are using AI less as a one-off assistant and more as a tool for managing repeatable work across systems, data and internal processes. That marks a shift from earlier use cases, which often centred on drafting content, brainstorming ideas or shortening single tasks.

In experimentation, the share of tests carried through to completion rose over the past year, with concluded experiments up 38.0%. Experiment win rates reached 26.4%, while concluded personalisation campaigns increased 42.4%.

There were also gains in campaign and content operations. When customers used Opal alongside Optimizely's Content Marketing Platform, campaign production rose 85%, according to the company. In digital asset management, Opal contributed to a further 57% increase in asset reuse by helping teams find, adapt and reuse existing material.

Mark Fagiano, Associate Vice President, eCommerce at Road Scholar, described the effect on his organisation's day-to-day work.

"Adoption across our teams happened faster than we expected because the value was clear right away," said Mark Fagiano, Associate Vice President, eCommerce, Road Scholar.

"Using Optimizely Opal, we've moved beyond generating ideas to actually getting work done through agents built into our workflows. That's where the impact shows up. We're saving hours, and in some cases weeks, across the team," Fagiano said.

Customer examples

Optimizely cited several customer deployments to show how the software is being used in practice. A real estate business is running experimentation across more than 200 websites with one marketer, cutting variation development times from days to 30 minutes.

A global hospitality brand generated more than EUR €4 million in revenue through experimentation supported by Opal, according to the company. Another distributed content team reduced brief creation from hours or days to near-instant turnaround, allowing staff to respond to demand across several business units.

Product expansion

As usage grows, Optimizely is adding more AI functions to the platform. Opal now includes more than 40 pre-built agents, with recent additions including tools for search optimisation, experiment variation development and reporting on Google Analytics 4 traffic.

The product also connects with software widely used by marketers, including Salesforce, Google Analytics, Figma and Atlassian. Those integrations allow agents to work across existing systems and data sources rather than sit inside a single application.

Alongside the software rollout, the company has introduced a training and certification programme called Opal University. More than 1,800 companies have joined, 445 certifications have been completed, and participants have built more than 1,499 agents while saving more than 4,000 hours of work.

The latest figures suggest AI adoption in marketing is shifting from experimental use to operational deployment, particularly in areas where teams are under pressure to increase output without adding staff. A third of activity on Opal already involves agents completing multi-step tasks, indicating that customers are using the software for finished work rather than just first drafts.