AI Adoption & Integration

Turning uncertainty and distrust to clarity and alignment

NOTE: This case study describes one firm’s approach. Data protection obligations vary by profession, client contract, and jurisdiction, so review any AI guardrail with your legal counsel and your IT or security lead.

The Summary

Like many organizations, this professional services firm knew AI mattered but had no shared expectation for how, or whether, people should use it. A few early adopters were experimenting quietly, and many others were unsure what was allowed. Through the firm’s Innovation Committee, the COO led a sequenced approach: set the expectation that everyone uses AI within clear guardrails, protect client and personal information, create a shared channel for learning, and build responsible AI use into annual reviews.

Within three months, most staff were sharing how they used AI, and the firm had a foundation to build on.

“I was hesitant to use AI. I didn’t know if I was supposed to, if it counted as cheating, or if it was allowed on client work. The Innovation Committee's clear expectations and the AI Collective Channel changed that. I started using it openly, asked my team to do the same, and we began sharing how we used it in team meetings. AI became our thought partner, and we had some fun wins along the way.”

— Senior Manager

• The Symptom

  Some teams were experimenting with AI quietly, others hadn't started, and no one was sure what was allowed.

  Without leadership guidance, team members couldn’t tell whether AI use was encouraged, permitted, or risky. What individuals were learning wasn’t being shared, and the firm had little visibility into how AI was being used or what information was going into it.

• The Root Cause

  The firm lacked clear expectations and guardrails, not tools.

  The tools were available. What was missing was a clear message from leadership that AI was expected, simple rules for protecting client and personal information, and a place to learn from one another. Without those, uncertainty carried risk in both directions: cautious people held back, while informal use outside any guidelines could expose information the firm owed its clients.

• The Solve

  • Set the expectation first. The firm made it clear that everyone was expected to use AI in their work, within the firm’s guardrails. Starting there removed the hesitation about whether it was allowed.

  • Keep the guardrails simple, and back them with structure. Rather than a lengthy policy, the firm set one easy-to-remember rule:

    • No personally identifiable information or confidential client information goes into AI tools, except tools the firm has approved for that data.

    • The rule was supported by an approved-tools list, a review of each vendor's data-handling terms, and a named contact for questions; and every person stayed responsible for checking AI's work before it reached a client or colleague confirm this was stated policy.

  • Choose and configure tools deliberately. Where appropriate, AI features in the firm’s existing software were activated, so work could happen inside the systems they already in use. The official AI license was rolled out in different versions, such as licensing tiers to test what the firm needed after reviewing data-handling terms and file permissions.

  • Lead through the Innovation Committee. The committee was supported by the CEO and chaired by a shareholder, which built buy-in from the top. It championed adoption, kept momentum across the firm, and owned updates to the guardrail.

  • Make learning social. A firmwide AI channel gave teams a place to share how they were using AI, ask questions, and learn from colleagues’ examples, in line with the same guardrail.

  • Build it into how work gets done. Responsible AI use became an expectation in annual reviews, signaling that it was part of how work gets done and not an optional experiment.

• The Results

  • Within the first three months, staff were engaging with the AI Collective Channel, posting their own examples or sharing outcomes from drafting, research, summarizing, and client prep

  • Staff who used AI regularly reported about a 30% increase in efficiency, even after allowing time to review and verify AI's work

  • AI moved from a quiet experiment for a few to an open, shared practice across the firm

  • Responsible AI use was added as an expectation in annual reviews

The Takeaway

AI adoption is rarely only a technology problem. It’s a leadership, culture, and governance problem. People move quickly once they know what’s expected, what’s safe, and where to learn. The work doesn’t end there. Tools, vendor terms, and regulations keep changing, and no firm has AI fully figured out, so guardrails need regular review. But clear expectations, simple rules, and a shared place to learn can move an organization further in months than open-ended uncertainty can in years. HuntingtonLevy helps firms take those first steps and build from there, working closely with your legal counsel and your IT or security lead.