AI Bots Are Now Your Colleagues, Thanks To Optimizely’s Latest Launch

Optimizely, an AI marketing platform, launched Virtual Teammates, AI personas with job titles and roles that work alongside human colleagues within its Opal agentic platform. President Shafqat Islam said the company may not be around in its current shape in a few years, but saw an opportunity to restructure by leaning into AI. The virtual teammates, which can be customized by customers, run on schedules and have distinct identities for governance, according to SVP of Product Kevin Li.

The phrase “AI is everywhere” is becoming truer every day. It’s in NYC subway ads https://www.techbuzz.ai/articles/friend-ai-startup-bets-1m-on-nyc-subway-ads-despite-backlash . It’s in daily workflows and embedded into Slack channels. And now, it’s in the employee org chart. On Monday, AI marketing platform Optimizely announced the launch of Virtual Teammates, which are… exactly what they sound like: AI personas, each with a job title and role in the org chart, that work alongside human colleagues. In a few years, Optimizely “may not be around, at least not in its current shape,” said President Shafqat Islam. But within the “existential threat” of rapidly evolving technology, Optimizely saw an opportunity to restructure its offerings by leaning into AI, he said. Last year, the company launched Opal, an agentic platform that guides marketers through the marketing journey, from planning to publication to analysis. The new virtual teammates live within Opal, and can have one of six “roles”: Chief of Staff to help with meeting prep and delivering briefings , SEO & AI Search Analyst to monitor and find opportunities to improve search and AI visibility , Marketing Analyst to convert data into insights , Personalization Strategist to design campaigns and monitor performance and CRO Manager to analyze performance and recommend next steps . BYOB Build Your Own Bot Optimizely’s customers can design their own teammate, assigning them a name and one of the six roles. A user’s personal teammate will have access or limited access, depending on its role to their specific data and back-end systems. Many AI agents are reactive, rather than proactive, meaning that they need to be prompted every time a human needs a certain task performed. In contrast, the virtual teammates are designed to run on schedule, the same way that “humans actually work,” said Kevin Li, Optimizely’s SVP of product. For instance, he added, Li’s own Chief of Staff agent checks his inbox every day at 5am and categorizes and prioritizes any emails that came in overnight. The virtual teammates are constantly improving by analyzing interactions with their human managers. They may pick up on details like how the human tends to structure their reports and update their own report structure accordingly. But the teammates only need to be programmed once, at the time of their creation, rather than individually triggered each time they need to complete a task. Optimizely also wanted to give each of the teammates a distinct identity with different permissions and logins catered to their role. The digital teammates are configured so that they only have access to the systems they need, and sometimes only partial access – like view-only – the same way a human employee might. The identities are key for governance, said Li, so any action can be traced back to a particular teammate. Another issue that Optimizely saw within many current AI systems was the one-on-one dynamic between humans and their AI assistants. Sharing something from one’s own Claude instance to someone else is “like a 1990s experience,” said Li, that requires either copy-pasting the content of the chat or generating an internet-accessible link. In a platform like Claude, he added, two employees and their agents can’t all communicate in a single thread. “We have this amazing machine that has all of human knowledge,” said Li, “and yet for me to share with another human, I have to ‘control C, control V.’” The horror. Through Optimizely’s platform, human employees and their virtual teammates can collaborate and brainstorm in a shared chat, discussing topics like how to revise messaging to better connect with a specific audience, or what changes need to be made to a campaign brief. The big question There’s one obvious concern that arises when the topic of digital employees comes up: What happens to human employees? According to Li, nothing bad. Although it’s a “hot topic of discussion,” he said, marketing teams see the digital teammates as “additive” rather than a substitute for human roles. For marketers facing tight budgets or hiring freezes, the best solution in the past often would have been swapping in someone from a different role, thus shortchanging a different team. One of Optimizely’s clients had a six-person team, each of whom was stretched thin – but providing each one with their own digital teammates gave them additional “perspective” on optimization and personalization strategies, Li said, plus somewhere to offload more manual tasks like keeping an eye on Google Analytics and SEO performance. What about when those hiring freezes end, though? Do CMOs have any incentive to add humans to the team, if the digital teammates have proved their worth? Li insists that human roles shouldn’t be affected. A virtual teammate can’t figure out messaging, or a company’s value proposition, or what makes it unique from its competitors, he said. Interesting take, coming from a company with AI tools that both generate AI creative and monitor competitive intelligence. He compared the shift to outdated roles like calligraphers and court stenographers. Humans will focus on “net new use cases,” like running agentic programs and making final decisions, he said. “But someone still needs to look at daily anomaly detection,” he added, “which, by the way, no one does today.” The question, then, is whether that “someone” should be some one or some thing.