What are the biggest AI risks facing professional services firms?

Tom Ricketts

Insight

5 minutes

03 August 2026

5 categories where risk can occur, and some practical controls that can reduce a firm’s exposure.

Professional services firms are quickly learning that while Artificial Intelligence (AI) implementations can enhance efficiency, productivity and insight, it can also introduce new ways for professional liability exposures to arise. 

Growing AI Adoption

According to a recent survey, Thomson Reuters’ Future of Professionals Report, 80% of respondents believe AI will have a high or transformational impact on their work within the next five years. That’s an increase of 3 percentage points from 2024 to 2025.

As more firms adopt AI to stay competitive and meet growing client demand, these tools can magnify traditional risk areas such as errors, omissions, inaccurate advice, and data privacy / confidentiality compromises if appropriate guardrails are not put in place.

AI is transforming the way professional services firms deliver advice

The use of AI provides faster data processing, reduction of manual errors, and automation of routine tasks, allowing professionals to focus on more challenging issues and deliver more accurate and cost-effective recommendations to their clients.

Legal professionals use AI-powered tools for tasks such as contract review, legal research and document analysis. In accounting, automated data analysis is being deployed to streamline audits and to identify discrepancies more efficiently. Consulting firms are leveraging predictive analytics to offer clients tailored strategic advice based on real-time trends and data.

AI isn't something professional services firms are “trying out” anymore. It's part of the advice itself.

But that's where the risk shifts. The person who signs off the work is still the one responsible for it. If the AI model gets something wrong, AI doesn't take the blame, the firm does. Take legal research. An AI model can point to a court case that sounds convincing but never actually happened (AI-invented case law). Put that in front of a client, and it's the lawyer's name on the work. And the lawyer's duty of care.

AI saves time, no question. It also brings a new kind of risk, which is why professional services firms need a plan for the moments it gets things wrong.
 

AI-driven professional liability risk scenarios

The following AI-driven professional liability risk scenarios highlight where professional services firms can face issues, and some practical controls that can reduce a firm’s exposure. AI-driven professional liability risk generally falls into one or more of the following five categories:  (1) Overreliance on AI-generated output, including hallucinations and gaps; (2) Confidentiality, privacy, and privilege compromise when client data is shared with AI vendors; (3) Supervision and quality control breakdowns; (4) Biased or noncompliant recommendations; and (5) Weak documentation of the work performed and the basis for advice.

Claim scenario: AI output is used in advice or deliverables including hallucinated, incorrect and/or misleading information, leading to incorrect or incomplete guidance.

  • Set clear permitted vs. prohibited use cases (drafting and summarization vs. final conclusions).
  • Require qualified human review and sign-off before anything is shared with a client.
  • Use a validation checklist (primary sources, jurisdiction / version, and independent calculations).
  • Prefer tools that provide citations, retain the cited sources in the file, and require that all citations be checked by a qualified human.
     

Claim scenario: Client data is entered into a third-party AI tool and is used by the vendor to train the model that is deployed to other clients, triggering allegations of confidentiality breach, privacy violations, loss of privilege or breach of intellectual property.

  • Classify data and restrict what may be entered into AI tools; require redaction / anonymization where needed.
  • Use firm-approved tools with clear vendor terms (security, retention, access controls, and training prohibitions where required).
  • Provide prompt guidance and escalation paths for sensitive or privileged matters.
  • Address AI use in engagement letters and client communications when appropriate.
     

Claim scenario: AI compresses timelines and weakens review / supervision, resulting in advice being issued without accountable sign-off or outside the practitioner’s competence.

  • Train teams on approved tools, their limitations, and mandatory validation protocols.
  • Match review depth to risk and materiality; document who reviewed and when.
  • Use playbooks for repeat engagements (templates, prompts, and checklists).
  • Periodically test AI-enabled workflows and remediate gaps.
     

Claim scenario: AI-generated recommendations overlook governing requirements or embedded bias, contributing to noncompliance or foreseeable adverse outcomes.

  • Identify the governing standards and confirm conclusions against primary sources.
  • Use structured intake to capture facts, assumptions, and constraints before drafting.
  • Apply enhanced review for high-risk topics (e.g. tax, employment, benefits, lending, sanctions).
  • Make and document key judgments in plain language (not just AI-generated phrasing).
     

Claim scenario: The firm cannot demonstrate a reasonable process because prompts, outputs, sources checked, and review steps were not retained.

  • Retain key prompts and outputs relied upon, sources checked, and reviewer sign-off.
  • Store AI artifacts with the matter file / workpapers under existing retention rules.
  • Record the tool and version used where it may matter to reproduce the work.
  • Ensure deliverables reflect the firm’s final professional judgment.
     

What professional services firms should discuss with their broker

By partnering with experienced risk advisors such as the Howden Professional Services team, professional services firms can understand how these technologies may affect their professional liability, cyber, management liability, employment practices liability and contractual risk profile. A proactive conversation can help identify potential coverage gaps, assess emerging exposures, and ensure insurance programs keep pace with evolving business practices.

For professional services firms, the issue is no longer whether AI will shape client advice, but whether it will be subject to appropriate governance. Firms that combine AI-enabled efficiency with clear oversight, defensible processes, and professional judgment will be better positioned to manage professional liability risk, protect client trust, and stand out in a rapidly changing market.
 

Speak to our Professional Services Insurance team today

Contact Us