AI feels inescapable. It’s not just in the news (this company just pivoted to AI!) or in advertisements (this AI tool will fix your life!). It’s at the gym: people are turning to chatbots for full fitness plans. It’s in the kitchen: share a recipe screenshot and let UberEats build your grocery cart. It’s at work: countless stories about AI supplanting the human workforce. It’s at your friend’s birthday party: prompt a playlist to queue up tunes. It’s at the movies: Val Kilmer is back on the silver screen, posthumously. It’s on vacation: let AI pick your next destination.

Whether overtly or simmering under the surface, there is a relentless undercurrent of urgency; a pervading sense that if you haven’t yet upended your entire way of being, living, and working, you’re already hopelessly behind.

But is that actually true?

AI in the Workplace

The truth behind the current state of AI in the workplace is elusive, if that “truth” exists at all. It’s difficult for individuals to get a true sense of what other companies are doing internally—and, by extension, where you fall on the adoption curve.

Though being Pliancy’s associate director of communications is a decidedly non-technical role, working here has huge benefits to my awareness and understanding of technology, from fundamentals to best practices to emerging tools. (Don’t worry, my English degree and I won’t be responding to tickets anytime soon.) It’s through that lens that I’ve been warily eyeing the AI discourse coming out of the 24-hour news cycle and the LinkedIn echo chamber. Instead of trusting that those were accurate representations of the current landscape, I set out to uncover what’s really happening out there.

Pliancy’s Inside Look

As an IT provider, Pliancy maintains over 200 client environments. I’ve spent the last few weeks interviewing colleagues about what they’re seeing in the real world, what trends are popping up, and how clients’ attitudes towards AI have changed and evolved in the recent past, all rooted in firsthand experiences with clients.

The following observations represent our own client base only, which skews heavily toward life sciences startups and VC/PE and other capital management firms. Most of these organizations have between 5 and 100 employees. Our clients are located across the United States, with concentrations in the Bay Area, Boston, New York, Austin, San Diego, and Los Angeles.

Most companies feel “behind,” but it doesn’t mean they actually are.

My conversations confirmed the hunch that feeling “behind” is more perception than reality. There’s certainly anecdotal evidence that AI has led to modest workflow improvements for certain types of automation-ready work like research, coding, data organization, and so on. But as for businesses fully and successfully rearchitecting themselves around AI… things aren’t quite there yet. One interviewee estimated that AI use currently results in 15% productivity gains for individuals and 5% gains across a company. Helpful, but not transformative.

“Everyone thinks they’re behind. I have yet to meet someone who doesn’t think they’re behind.”

– Noah Tagliaferri, VP of Growth

There is not a single, ubiquitous set of parameters when rolling out AI.

There’s high variability in how organizations choose to launch and provide access to AI. Some have chosen to sign up every employee to every AI tool they come across. Others have taken a more conservative approach with formal AI policies, required training before access is granted, and role-based availability. In some cases, organizations have gone further by creating AI subcommittees and dedicated roles to oversee experimentation and adoption. There’s not yet a clear frontrunner for what may one day become a default approach.

AI integration is being driven by individuals, rather than by universal adoption.

Despite what the zeitgeist and leading AI companies want you to believe, AI adoption is not yet universal in terms of access, interest, or meaningful engagement. Instead, integration and daily use are driven by a minority of passionate individuals. I heard no evidence of environments where every employee, top-to-bottom and across departments, has changed their entire daily workflow or work product as a result of AI.

Critically, access does not necessarily equate to adoption. Within client environments, our team members see hotspots of activity, such as developers writing code or data analysts and data scientists processing raw data. Those with technical backgrounds or affinities also gravitate towards experimenting with new tools, regardless of their current roles.

People are probably using AI on personal accounts—whether you have org-wide AI subscriptions or not.

Our previous blog post on AI strongly suggested implementing an organization-wide AI tool for all users, because without an approved option, users were likely to interact with AI models via free or personal accounts. Using non-enterprise accounts for business purposes could lead to the exposure of proprietary or privileged information.

In my conversations, multiple team members mentioned that this remains a challenge even in environments with approved options: some still utilize unsanctioned AI tools or non-enterprise accounts, whether due to model preference, task-specific tooling, or convenience.

This poses an obvious risk from a cybersecurity standpoint. But the threat of shadow IT (unmonitored use of tools by an employee “going rogue”) is nothing new. The impulse stems from human nature, and while it’s never ideal, it reinforces why clear policies and workforce training matter so much. Some clients have rolled out AI training modules and hired third-party trainers, while others are still working on formalizing guidance around approved AI use.

Interest in AI is largely initiated by FOMO, rather than by clearly defined goals.

Multiple interviewees cited fear of missing out as an initial driver for clients, spurred on by misleading or overly boastful social media posts and Big Tech’s supposedly AI-driven layoffs. This FOMO-inspired haste gets in the way of slowing down to consider specific aims or end results.

While FOMO is usually the initiating spark, clients eventually pivot to an interest in AI’s productivity potential as a justification for their interest. However, because they tend not to have a particular goal in mind, clients frequently struggle to articulate what they’re trying to achieve. The sentiment tends to be general: “We want to use AI to help us work,” with scant additional details, as opposed to “We want to leverage this specific strength of AI in order to change this particular process or workflow in X/Y/Z ways.”

“People have this perception that they’re being left behind, so they want to move quickly—but they don’t know why they want to move quickly other than other people’s opinions.”

– Donald Gonzalez, IT Director

Interest has spiked since February 2026.

You don’t need me to tell you that widespread interest in AI has grown in recent years. But what’s significant is a recent surge in interest, pinpointed down to the month.

When asked about shifts in client attitudes over the past six months, several interviewees agreed that February 2026 was a turning point for client interest in AI. It’s possible that Claude Cowork’s release (in January 2026) could have been a contributing factor in this spike.

DDQs and the SEC have not yet caught up to the current state of AI and its impact on compliance.

AI is something of a blank space where compliance is concerned—at least for now. Though the SEC did charge two advisers with making misleading statements about their use of AI in 2024, practical guidance surrounding AI use has otherwise been foggy. One interviewee raised a critical question about whether, for the SEC’s purposes, LLM conversations constitute electronic communications and should be archived.

Another interviewee mentioned joining a Pliancy client’s due diligence meeting with a major investment bank to provide IT and cybersecurity input. Though he expected to be grilled on AI use, bankers didn’t ask a single question focused on AI. This lag in oversight seems conspicuous, and it’s unlikely to last forever.

Reflecting on Your Own AI Use

Based on my research and interviews, here are three considerations that can help you refine your organization’s approach to AI:

🎯 Be specific and intentional in your goals

Start with AI’s strengths, then think about how to apply them to your business operations or workflows. This increases your odds of extracting lasting gains from AI. Letting AI’s strengths guide your use cases also means that not every team will use AI tools in the same way or at the same volume. This reflects natural differences in the type of work being done, and you shouldn’t expect uniform adoption across the board.

✅ Make it easy for employees to use approved tools.

Assume that employees are experimenting with AI, and make access as frictionless as possible. No employee and no training program is perfect, but the more activity you can contain within sanctioned, enterprise-grade environments, the safer your data will be.

👀 Governance still matters, even if it’s not codified yet.

Just because DDQs and the SEC are still playing catch-up to AI doesn’t mean that guardrails and policy documentation aren’t necessary. This Wild West period won’t last forever, so get a head start on access rules, approval workflows, archiving plans, and contingency plans for breaches, leaks, and other incidents.

Waiting for the Dust to Settle

Beyond reinforcing the urgency that I was sensing, these interviews shed light on the vast variability that exists in AI interest and integration. Things can’t be boiled down to a simple spectrum with “no AI” at one end and “all AI” on the other. Instead, it’s more of a scatter plot, shaped by zeal, aptitude, risk, and resources. The technology is still too new for the journey to be linear; too untested for the path forward to be set in stone.

This state of flux shows no signs of slowing. If I were to redo these interviews in a year, or even in six months, I imagine the responses would paint a different picture. Not one of complete certainty quite yet, but as the dust begins to settle, my hope is that we start to worry less about falling behind and focus more on the promise ahead.