As financial advisors increasingly rely on digital tools, many are discovering that AI is being quietly embedded across their tech stacks, often switched on by default and difficult to disable.
CRM systems like Wealthbox and Orion have embedded AI meeting note-takers and portfolio comparison features. Zoom added an AI Companion for meeting summaries. Microsoft has integrated Copilot. And many financial planning providers now include AI-powered document analysis.
Many of these embedded AI features are set to “opt out,” meaning users must decline them in order to turn them off, as opposed to “opt in,” where they would have to request activation. While such tools may improve efficiency, they also raise questions around transparency, control and compliance.
Experts say advisors need to closely monitor their tech stacks to ensure these new embedded AI features are useful — and figure out ways to disable them if they are not.
“Advisors should not have to hunt through hidden settings to manage powerful features that impact client interactions or data,” said Ramona Ortega, founder of agentic AI financial services workflow platform WealthBuild.
READ MORE: Tech serving higher-value clients beats expanding capacity: Kitces
Some updates are useful, others not so much
In recent years, advisors often sought out standalone AI tools to perform specific tasks. But increasingly today, existing software gains AI capabilities through regular updates, said Fergal Glynn, chief marketing officer at AI security firm Mindgard. The results of this quiet shift have been mixed.
Gray Reinhard, chief technology officer and managing partner at renewable energy investment platform Energea, said he’s noticed AI features appearing in nearly every third-party tool his team uses, including HubSpot CRM, Clay lead sourcing, integrated development environment code assistants, Google Analytics summaries, Zoom meeting recaps, Adobe editors — even security camera theft detection. He has also rolled out custom AI integrations in his firm’s internal CMS.
“Some of the features are incredibly useful, while others add zero value,” he said. “In general, products well-positioned to leverage AI are becoming noticeably more powerful, but they bring added complexity, permission considerations and some onboarding friction. Still, overall they deliver more value than hassle, and so far they haven’t significantly increased our cost of ownership.”
One feature Reinhard said he’s had to disable is code editors.
“I need granular control because AI-powered extensions can create overlap, reduce performance, access sensitive data, rack up token usage or just become a distraction,” he said.
While he said he’s not overwhelmed by managing it all, he does worry about the side effects.
“For example, if a CRM auto-enriches leads via AI, it may tempt us into overbuilding workflows that actually distract from our core funnel — which, ironically, worked better before AI,” he said.
When it comes to opt-out versus opt-in, context matters, said Reinhard. So far, most AI additions he’s encountered feel more like opt-in experiences, in that they require user action and aren’t inherently tied to the product’s core functionality.
“That’s encouraging, because features should solve a problem, not create one,” he said. “I’ve seen organizations fall into the trap of asking, ‘How do I leverage this AI?’ instead of, ‘How can AI help solve problem X?'”
READ MORE: How advisors can get noticed in a no-click search world
Advisors need transparency and control
There’s significant pushback against opt-out defaults, said Glynn, from those who argue it is a “deceptive pattern” forcing unwanted AI functionality.
“The consensus favors opt-in defaults, especially for sensitive client data,” he said. “Advisors want better control, clearer disclosure about data processing and granular controls to selectively enable features rather than all-or-nothing approaches.”
A default opt-out may serve the vendor’s goals but undermines trust, said Ortega. Providers should prioritize transparency, with clearer notifications and consent workflows that respect the advisor-client relationship.
“Just as importantly, advisors need a clear understanding of how AI integrates into their workflows and contributes to return on investment,” she said. “Otherwise, it’s just another opaque layer in the tech stack instead of a tool that adds real value to the job.”
Forcing advisors to hunt in their settings to manually turn off new embedded AI features is a common headache, said Glynn.
“Some CRMs keep things simple, but others either hide the controls or don’t explain the process clearly,” he said.
Staying compliant with active oversight
Arnulf Hsu, founder and CEO of GReminders, an AI-powered end-to-end meeting and automation management platform specifically for financial advisors, said opt-in features tend to have lower adoption but give users more control. Opt-out defaults usually drive higher usage, but only if the experience is well-designed and transparent.
“The bigger risk with opt-out is that users may not fully understand what the AI is doing, which can lead to confusion or compliance issues,” he said. “As we shift from deterministic systems to generative AI, it’s up to vendors to manage that complexity, communicate clearly and build strong feedback loops with their users.”
Hsu said the way his firm approaches embedding AI depends on the specific features.
“For minor improvements, AI often runs in the background without users noticing,” he said. “But when generative AI becomes central, it requires much closer scrutiny.”
Hsu said GReminders focuses on high-impact, low-risk use cases with a human in the loop, such as drafting a client email and then having the advisor review before it’s sent.
“If privacy and security frameworks are in place, small AI features are usually well received,” he said. “But when outputs are AI-generated, it’s important to involve compliance and IT early.”
The size and scope of the feature rollout has a major influence on the overall feedback, said Hsu.
“Major rollouts usually require direct coordination with firms to avoid confusion and ensure adoption,” he said. “On the other hand, smaller features can be introduced more quietly, but thoughtful user interface and experience still make a big difference.”
