AI for professional services in Northwestern Ontario: what accountants, lawyers, and insurance brokers can use right now

Each of the three major professional services regulators in Ontario — the Law Society, RIBO, and CPA Ontario — has published specific AI guidance. The tools that earn their keep in a professional services firm are narrow but real: drafting, research, document review, and client communication. Here is what applies, what is off-limits, and where the funding is.

For an accounting firm, law practice, or insurance brokerage in Northwestern Ontario, AI tools are genuinely useful for a narrow but high-value set of tasks: drafting correspondence, summarizing documents, structuring research, and generating first-draft templates. The practical ceiling is real, and it is defined not by the technology but by the regulatory obligations that govern each profession. The Law Society of Ontario published guidance on lawyers' use of generative AI in April 2024. The Registered Insurance Brokers of Ontario published responsible AI use guidance in May 2025. CPA Ontario has addressed AI as a professional conduct obligation under its own framework. All three start from the same position: professional obligations do not change when AI is involved, and professional judgment cannot be delegated to a machine. For a Thunder Bay firm that wants to start using these tools without creating a regulatory problem, the sequence matters — obligations first, then tools, then funding.

What each regulator says

The Law Society of Ontario white paper on generative AI (April 25, 2024) identifies six areas where existing professional conduct rules directly apply to AI use: competence, confidentiality, supervision, client relationships, billing practices, and non-discrimination. Lawyers must have enough understanding of an AI tool to assess its outputs — technical expertise is not required, but the capacity to critically evaluate results is. Lawyers cannot bill for time they did not spend; efficiency gains cannot inflate invoices. Client data must be protected before it enters any AI system, and every output must be reviewed before it is used or provided to a client. The LSO does not mandate blanket disclosure of AI use but identifies four factors a lawyer should weigh: whether the use will be public in a proceeding context, what the client reasonably expects about who prepared materials, whether client data entered the system, and any reputational risk to the client. From January 2026, Legal Aid Ontario requires roster lawyers to confirm annually in the Lawyer Self-Report that they have read and comply with LSO generative AI guidance. The relevant resources — a quick-start checklist, best practice tips, and a professional obligations summary — are available through the LSO Technology Resource Centre at lso.ca.

The Registered Insurance Brokers of Ontario issued "Responsible AI Use Among RIBO Licensees" on May 29, 2025. Its core position: using AI technology does not alter a licensee's existing obligations under the RIBO Code of Conduct, and AI use cannot replace a licensee's professional judgment or conduct. RIBO sets out four governance expectations. First, licensees need to be trained to identify when AI is being used and understand the associated risks — technical expertise is not required, but informed oversight is. Second, licensed brokers must oversee all AI-generated or AI-modified outputs before they reach a client; for underwriting applications, firms should audit outputs to prevent systemic bias. Third, transparency: "a customer should know when they are engaging with AI instead of a human." Customer-facing AI tools must disclose that they are AI, and human oversight must be in place to catch and correct errors. Fourth, data protection: brokers must vet vendors to ensure client data is not transferred to third parties or used for AI training; individuals should not process client information through open-access AI systems. RIBO expects firms to develop a written AI governance policy before any tool goes live. The guidance is at ribo.com/responsible-ai-use-among-ribo-licensees.

CPA Ontario has addressed AI as a matter of professional conduct rather than issuing standalone AI-specific rules. The firm-level framing comes through CSQM 1 — the Canadian Standard on Quality Management for accounting firms — which has been effective since December 2022 and requires documented controls for any technology resource that affects engagement quality, including AI systems. For practical purposes this means: if a firm's accountants are using AI in the course of engagement work, the firm's quality management framework must address it. CPA Ontario's published guidance treats AI use as a CPA Code obligation — the professional is responsible for the output regardless of how it was generated. Confirm current CPA Ontario AI guidance and any updates at cpaontario.ca.

Where AI actually earns its keep

The clearest fit in all three professions is the same pattern: AI drafts, the professional verifies, the professional is responsible. That pattern holds whether it is a lawyer drafting a demand letter, an accountant preparing a tax memo, or an insurance broker drafting a coverage summary. The tasks where this works reliably are structured writing from known facts — the professional supplies the information, the AI produces a draft, the professional reads and edits it before anything leaves the office.

• Legal: first-draft correspondence, contract clause summaries, research memos on defined questions, intake questionnaire responses, meeting notes from recorded client calls (with PIPEDA consent obligations — see the AI meeting notes post on this site).
• Accounting: tax position summaries, draft client letters, structured working paper narratives, payroll communication templates, note-taking during review meetings.
• Insurance: coverage summary drafts from policy documents, initial claims intake notes, client communication templates, policy comparison summaries for standard lines.
• Across all three: internal training materials, HR-facing policy drafts, marketing copy for firm websites and service descriptions, and responses to standard FAQ-style client inquiries where a professional reviews and sends.

The area where the value is highest — and where the risk is also highest — is legal and regulatory research. AI tools can surface relevant statutes, summarize published case summaries, and draft structured research memos quickly. The problem is hallucination: AI language models will sometimes generate plausible-sounding citations that do not exist. Canadian courts have already seen consequences from this. In Zhang v. Chen (2024) and Ko v. Li (2025), lawyers submitted AI-generated citations to fictitious cases. In both instances, the fabricated citations were not caught before they reached the court. The professional obligation to verify AI-generated legal research against primary sources is not optional — it is the core of competent practice. Every AI-generated case name, statute reference, or legal proposition must be confirmed against a primary source before it is relied on or filed.

Data and privacy: what client information can go into these tools

PIPEDA — the Personal Information Protection and Electronic Documents Act — applies to AI-assisted workflows in the same way it applies to manual ones. A professional services firm that enters client names, financial records, file details, or correspondence into a general-purpose AI assistant is transferring personal information to a third-party processor, typically on servers in the United States. PIPEDA does not prohibit cross-border transfers but requires appropriate safeguards — typically a Data Processing Agreement with the vendor that governs how data is handled, the retention period, and deletion on request. For a law firm, confidentiality and solicitor-client privilege apply on top of PIPEDA and do not disappear at the platform boundary: client information entered into an AI tool is client information the firm is responsible for. For an insurance brokerage, RIBO's guidance explicitly states that individual licensees should avoid processing client information through open-access AI systems and that vendors must be vetted to ensure client data is not used for AI training. The practical approach across all three professions is to use AI tools on anonymized or synthetic content for drafting and research, and to bring in client-specific data only through verified, governed systems with a signed Data Processing Agreement in place. Regulated professional practices with complex data handling needs should confirm specific requirements with a privacy lawyer and the relevant regulatory body before deploying any AI tool on client data.

Funding available in NWO for professional services firms

The Northwestern Ontario Innovation Centre's Building Blueprints for AI Adoption (BBAA) program is the most directly accessible AI adoption funding for professional services firms in Thunder Bay and Northwestern Ontario. BBAA reimburses up to $20,000 (50% of eligible project costs) for qualifying AI software, consulting, and integration work at growth-oriented for-profit SMEs with a valid Canadian business number in the Kenora, Rainy River, and Thunder Bay districts. BBAA excludes distributors, resellers, and retail-only businesses — accounting, law, and insurance firms are not retail businesses and should confirm eligibility directly with NOIC before applying. The program requires a pre-application consultation and a 30/60/90-day AI Adoption Plan. Contact: hussain@nwoinnovation.ca or 807-768-6682. Confirm current status, available funding, and eligibility at nwoinnovation.ca/programs/bbaa. The NOIC BBAA program has limited funding — contact the program before committing project expenditures.

Sources: Law Society of Ontario — Guidance on Use of Generative AI, April 25, 2024 (six professional conduct areas: competence, confidentiality, supervision, client relationships, billing, non-discrimination; no blanket disclosure mandate; four disclosure factors; quick-start checklist and professional obligations summary available through LSO Technology Resource Centre): lso.ca/lawyers/practising-in-ontario/practice-management-guidelines/technology-and-your-practice/generative-artificial-intelligence | Osler summary of LSO generative AI guidance (April 25, 2024): osler.com/en/insights/updates/law-society-of-ontario-publishes-guidance-for-licensees-with-respect-to-the-use-of-generative-ai | Legal Aid Ontario — 2026 Lawyer Self-Report update requiring annual AI compliance confirmation: legalaid.on.ca/in-briefs/2025-12-12_2026-update-to-lawyer-self-report-ai-compliance-confirmation | Registered Insurance Brokers of Ontario — Responsible AI Use Among RIBO Licensees, May 29, 2025 (four governance recommendations; written AI policy required before deployment; transparency obligation; data protection requirements): ribo.com/responsible-ai-use-among-ribo-licensees | CPA Ontario — AI as professional conduct obligation under CPA Code; CSQM 1 (effective December 2022) requires documented quality management controls for technology resources affecting engagement quality, including AI: cpaontario.ca | Thunder Bay-Northwest CPA Ontario Association (600+ CPAs across Northwestern Ontario): cpaontario.ca/members/community/associations/thunder-bay-northwest | NOIC BBAA — Building Blueprints for AI Adoption, up to $20,000 (50% of eligible costs), for-profit SMEs in Thunder Bay, Kenora, and Rainy River districts, excludes distributors, resellers, and retail-only businesses: nwoinnovation.ca/programs/bbaa | Office of the Privacy Commissioner of Canada — PIPEDA guidance on cross-border transfers and third-party processors: priv.gc.ca | Canadian legal AI hallucination incidents — Zhang v. Chen (2024) and Ko v. Li (2025): cited in Thomson Reuters legal AI guidance (legal.thomsonreuters.com/blog/how-ai-is-transforming-the-legal-profession) and Clio Canada legal AI compliance guide (clio.com/ca/blog/ai-legal-compliance)