For an AI assistant to be genuinely useful, it first needs to understand a great deal about you. Ollie, a personal assistant designed for everyday life, is betting that achieving this doesn’t require you to surrender all your personal data or compromise your privacy. While many enterprise-focused AI assistants already include data protection measures, Ollie stands out as one of the first mainstream, family-oriented AI tools to earn SOC 2 compliance—a certification that, through an independent audit, verifies a company has formal safeguards in place to protect customer information and run its systems securely. For Ollie’s team, reaching this milestone was a significant achievement, as it reassures potential users that the assistant isn’t collecting their data for other purposes, such as AI training. “We’re not sharing your data with anyone,” the founder emphasized.

The market is crowded with personal AI assistants positioning themselves as consumer-friendly options that integrate into daily routines via text messaging, which makes Ollie’s privacy focus a potentially powerful differentiator. Today, Ollie faces competition from other text-based assistants, including Poke (now part of Cognition), Fambot, Ohai, Folk, Saner.ai, and Tomo. There are also numerous assistants geared toward work, calendar management, and email, such as Town, Lindy, and Reclaim.ai. And then there’s Instinct, the AI assistant that secured $350 million in funding at a $2.5 billion valuation before even launching. To say the space is intensifying would be an understatement. In comparison, San Diego-based Ollie has raised a more modest $7.5 million seed round, primarily from Khosla Ventures, with support from AI House.

Despite the convenience these assistants offer, consumers are grappling with how much privacy and personal data they’re willing to give up. Instinct, for instance, faced criticism for its overly broad Terms of Service and privacy policy, which granted it a “perpetual and irrevocable” license to “access, use, host, cache, store, reproduce, transmit, display, publish, distribute, and modify” any user materials, including for training its AI models—a concerning prospect. Lennon is taking a different route with Ollie. Currently, this family assistant connects to calendars and email to help organize household schedules and keep users informed about their day-to-day activities. It also provides tools for meal planning, grocery shopping, to-do tracking, appointment booking, and bill payments through group chats. In the future, it may expand to help manage household budgets. However, all this access hinges on trust, which Lennon argues is Ollie’s core advantage over its rivals. “We’re not sharing your data with anyone. This is super sensitive, and that is necessary to win the trust of the users,” he said.

Beyond the protections offered by SOC 2 compliance, Ollie avoids requesting usernames or passwords to carry out its tasks. When Ollie needs to log into a website on the user’s behalf, it does so through its own cloud-based browser, then sends the user a link to a remote session. The same approach applies when the assistant needs to make a payment or purchase for the user. While this method provides security, the trade-off is that users must always log in to complete certain actions. Still, Lennon believes this is something technology will refine over time. “This is… new territory. I think in the future, we will do some form of hard tokenization, in a secure way, so you’re not going to have to re-enter [your information] every time,” he explained. “That’s frontier stuff… we want to find the right user experience that balances convenience and trust.”

Building that trust will be essential before users feel comfortable letting Ollie access more sensitive information, such as bank accounts, even through a connector like Plaid. Lennon, who holds a Ph.D. in AI, also has a background in fintech, having previously sold his startup Groundwork, a neobank for nonprofits, in 2021. That expertise could prove valuable as Ollie ventures further into money management.

Still, it remains uncertain whether people truly want this kind of assistance. Lennon is optimistic, noting that Ollie’s retention rates align with leading AI subscriptions in terms of paid subscribers. (He declined to share specific user or paying customer numbers, citing the highly competitive nature of the space.) These systems can also be inconsistent at times. Despite the buzz around Instinct, one of my initial tests involved pricing hotels and booking one—it quoted me incorrect rates across the board. Ollie, on the other hand, experienced an infrastructure outage with its text provider during my testing and stopped responding. When asked how Ollie handles such issues, given that most consumers give a new app or technology only one chance before moving on, Lennon acknowledged this is a broader challenge with AI. “This is the challenge with LLMs, in general—because they’re stochastic [i.e., involve probability], they’re inherently unreliable,” he said. “We have to essentially build the harness—the agent harness—in a defensive way to catch and prevent those things… It’s almost like there’s just 1,000 cuts that you’ve got to solve first.”