AI Adoption in Law Firms: What Actually Works with Norah Olson Bluvshtein

Nora Olson-Bluestein, Fredrikson’s Chief AI and Legal Operations Officer, said she uses AI for practical rather than flashy tasks. With two young children, she recently used Claude to brainstorm more interactive meetings and retreats, including worksheets and conversation formats, then toned down suggestions that felt too wild for lawyers. For a new training program, she used AI to map stakeholders, determine who needed input and when, and clarify who was a decision maker. She also said she argues with AI frequently; Claude characterized her voice as “direct, pragmatic, and a little opinionated,” apparently learning that through repeated use rather than an explicit style instruction.

Nora described Fredrikson as a roughly 400-lawyer firm based primarily in Minneapolis and the Twin Cities, with regional offices in Iowa, Wisconsin, North Dakota, and elsewhere. It is ranked 145 in the AmLaw 200 and is particularly strong in life sciences, private equity, and energy. Midsize firms face expensive tools, the need for additional personnel, and difficulty building custom systems without Big Law’s large developer teams. Fredrikson’s advantages are close relationships with lawyers and clients, little bureaucracy, and a broad legal-operations portfolio covering innovation, attorney development, paralegals, legal assistants, and project management. Her core team has about 30–35 people, with innovation handled by a smaller group.

Nora said Fredrikson’s change-management effort has moved through waves: first establishing basic awareness of generative AI, then supporting early adopters, reaching the early and late majority, and now working toward later adopters. She avoids calling anyone “laggard” because the term sounds judgmental. The timing reflects both external signals—such as industry focus on agents—and internal realities, including team bandwidth and the need to finish earlier initiatives. The firm’s culture is not top-down, so pilots, encouragement, department-specific approaches, and meeting people where they are work better than “thou shalt” mandates. Adoption also naturally progresses from experimentation to proficiency to business-strategy integration.

Nora rejected the idea that brand-new lawyers will automatically solve AI adoption for the profession. Law schools approach AI inconsistently, and junior lawyers may over-rely on it before developing judgment, legal intuition, and pattern recognition. She favors preserving an apprenticeship model in which experienced lawyers let juniors observe, participate, and receive instruction, perhaps supplemented by “flight simulator” environments. Nora worries about a possible AI-era “lost generation,” comparable to pandemic learning loss, if firms let technology outrun skills development. She acknowledged that she has no finished solution and said firms may need to compress and redesign traditional apprenticeship.

For firms with limited resources, Nora recommended buying a workable set of tools but reserving money for one or two people who can sit with lawyers, practice prompts, understand the underlying legal task, and iteratively build something useful. Otherwise, adoption may remain superficial—proofreading or summarizing rather than applying AI to substantive work. General demonstrations are insufficient because lawyers struggle to connect them to litigation, M&A, or other specific practices; hands-on, practice-specific workshops are more effective. Nora also observed that AI fluency depends more on temperament than age: strong delegators often become strong AI users because they give context, clear instructions, and continuous feedback, while lone wolves may struggle.

Nora attributed Fredrikson’s progress to a broad, coordinated portfolio and her team’s history of implementing change before AI arrived. She practiced at the firm, has been there about 20 years, and said that familiarity gives her credibility with lawyers. Her team combines former practitioners, paralegals, librarians, technology specialists, and people from outside legal. Because AI enablement appears across strategic planning, she can keep multiple departments focused on it. She also relies on relationships, opportunistic small projects, and client-driven needs to reach busy lawyers. Looking ahead, she expects more sophisticated agents, slowed by ethics rules, change aversion, and conflicting client requirements.