The uncomfortable truth about AI in Canada's wealth management industry

Senior leaders at wealth management firms across Canada have spent the past year building what they believe is an AI‑capable workforce.

Policies now sit in compliance manuals. Enterprise software licences have been purchased. Employees have completed online modules on responsible AI use and basic prompt engineering. Adoption dashboards show encouraging numbers.

But a new research study by AI company Section surveying 5,000 knowledge workers at large organizations in Canada, the United States and Britain reveals a stark disconnect between what executives think is happening with AI and what is actually taking place on the ground.

Workers are logging into AI tools. What they are not doing, in most cases, is using those tools to fundamentally change how financial advice gets researched, documented or delivered.

For Canadian financial planners and the firms that support them, this represents a strategic vulnerability that goes well beyond technology adoption. It is a capability gap that touches skills development, work design, management practice and, ultimately, competitive positioning.

What proficiency means now – and why most firms are not there

Twelve months ago, demonstrating AI competence meant answering a few straightforward questions: Do your people understand what these tools are? Can they use them without creating data breaches? Can they write a functional prompt?

Many advisory practices and dealer networks responded with exactly that kind of training. Planners learned not to upload sensitive client information to public chatbots. They practiced asking AI to polish an email or condense a regulatory article. By conventional measures, this counted as progress.

But the research argues that heading into 2026, the definition of proficiency has fundamentally shifted.

Today, being proficient with AI means weaving it into substantive, value‑generating tasks on a weekly basis. Not sporadic experiments with a chatbot, but systematic integration into the workflows that determine practice economics: gathering client information, running scenario analyses, conducting portfolio reviews, supporting tax planning, maintaining compliance documentation and producing client deliverables.

Measured against this new standard, the data suggests most organizations are nowhere close.

Consider the paradox: ChatGPT claims nearly 900 million users each month. More than half of Americans report using AI in some capacity. Yet the research finds that 85 per cent of workers lack what the authors term a "value‑driving" use case, and one quarter never use AI for work purposes at all.

Even within technology companies and roles built around language and analysis, AI usage remains largely cosmetic. For Canadian wealth management firms quietly projecting AI‑enabled productivity improvements into their business plans, that should prompt some uncomfortable questions.

Most workers remain stuck at the beginner stage

Three years after ChatGPT entered public consciousness, the report paints a picture of widespread but shallow adoption.

Approximately 70 per cent of surveyed workers fit the profile of "AI experimenters" – individuals who deploy AI for elementary tasks such as condensing meeting summaries, adjusting email tone or retrieving basic information. Another 28 per cent qualify as "AI novices": people who either avoid AI entirely or tested it briefly before abandoning it.

At the other end of the spectrum, meaningful expertise is vanishingly rare. Just 2.7 per cent of workers meet the threshold for "AI practitioners" – those who have genuinely integrated AI into their workflows and report substantial productivity improvements. A mere 0.08 per cent qualify as "AI experts."

Taken together, fewer than 3 per cent of the workforce operates at a level where AI is materially altering how work gets accomplished.

The research summarizes the situation plainly: 97 per cent of workers are either using AI ineffectively or not using it at all. One quarter report saving zero time through AI use. Forty per cent indicate they would be perfectly content if they never used AI again.

For Canadian planning practices enthusiastic about "AI‑enhanced financial advice," these baseline statistics should serve as a reality check. The probability that your team sits within that high‑performing 2.7 per cent is statistically low.

Why workers cannot identify meaningful applications

The core obstacle is not technical incompetence. Workers generally understand how to interact with a large language model. The challenge lies in recognizing where and how AI can be applied within the specific context of their roles.

The researchers describe this as a "use case desert."

Among survey respondents:

• 26 per cent report having no work‑related AI application whatsoever.
• 60 per cent characterize their existing applications as beginner‑level.
• After examining 4,500 reported work use cases, the research team judged only 15 per cent likely to produce genuine return on investment for employers.

Altogether, 85 per cent of knowledge workers either have no AI use cases or only rudimentary ones. One quarter never use AI at work. Forty per cent would be indifferent if AI disappeared tomorrow.

In financial planning terms, this typically manifests as advisers using AI to clean up client correspondence, draft a paragraph for a quarterly newsletter or digest a Canadian tax bulletin – while continuing to execute their planning and compliance workflows exactly as they did three years ago.

This distinction carries real weight. The time and cost drivers in a planning practice typically cluster around complex, recurring workflows: assembling client data, constructing and revising financial plans, stress‑testing retirement scenarios, documenting recommendations, preparing review presentations and satisfying regulatory evidence requirements. When AI does not penetrate those processes, it will not materially affect the metrics that determine practice profitability or scalability.

How workers are actually deploying AI tools

The study's catalogue of "most valuable" workplace applications helps explain why aggregate impact has been underwhelming.

The top ten work‑related applications by prevalence:

1. Replacing Google searches – 14.1 per cent
2. Generating initial drafts – 9.6 per cent
3. Correcting grammar and adjusting tone – 5.7 per cent
4. Conducting elementary data analysis – 3.8 per cent
5. Generating code – 3.3 per cent
6. Supporting ideation and brainstorming – 3.2 per cent
7. Assisting with meetings (such as transcription) – 2.7 per cent
8. Condensing documents – 2.0 per cent
9. Supporting learning and skill development – 1.6 per cent
10. Automating tasks and processes – 1.6 per cent

The researchers' assessment of these applications is unflattering:

• 59 per cent constitute elementary task assistance.
• More than 25 per cent play no meaningful role in broader processes or workflows.
• Only 2 per cent qualify as sophisticated applications.

Categorically, research and writing activities predominate, accounting for 19.6 per cent and 18.1 per cent of users respectively. Both categories skew heavily toward novice‑level deployment: isolated copy suggestions and superficial information retrieval.

For financial planners, this represents the gulf between asking AI to "rephrase this explanation of TFSA contribution limits for my client" and architecting a planning process where AI routinely extracts data from scanned documents, pre‑populates planning software fields, identifies inconsistencies, and produces initial drafts of meeting documentation and action items.

Modest time savings reveal limited transformation

Because AI usage remains superficial for most workers, the research documents only marginal productivity improvements.

Self‑reported weekly time savings break down as follows:

• 24 per cent report no time savings.
• 21 per cent save less than two hours.
• 23 per cent save two to four hours.
• 18 per cent save four to eight hours.
• 8 per cent save eight to twelve hours.
• 6 per cent save more than twelve hours.

Greater proficiency correlates with larger gains. "AI practitioners" prove 1.8 times more likely than "experimenters" to save four or more hours weekly, and 20 times more likely than "novices" to reach that benchmark.

The study notes that fewer than one third of knowledge workers save four or more hours per week, despite the fact that most organizations should be targeting savings of ten or more hours per employee to justify their AI investments.

For Canadian financial planning practices navigating fee compression, escalating regulatory burdens and rising client service expectations, trimming a couple of hours per week from email management may fall considerably short of the transformation they anticipated when they embedded "AI strategy" in their business plans.

Organizations are investing. Skills are not improving.

Credit where due: employers have not been passive. The research documents clear increases in organizational support:

• 63 per cent of respondents indicate their employer has established an AI policy.
• 50 per cent report access to an AI tool.
• 44 per cent have received AI training from their organization.

These investments demonstrate measurable effects:

• Workers at organizations with formal AI strategies score 1.6 times higher on proficiency measures than those at organizations without strategies.
• Workers with access to AI tools score 1.5 times higher than those without access.
• Workers who have received training score 1.5 times higher than untrained workers.
• Workers whose managers actively expect AI usage score 2.6 times higher than those whose managers discourage it.

Since March 2025, the proportion of workers reporting access to formal AI policies has climbed 17 per cent. Clear usage guidelines have increased by 16 per cent. Investment in AI tools and platforms has risen by 2 per cent.

Despite these positive trends, most employees remain clustered at the "experimenter" level. Workers who have completed AI training score an average of 40 out of 100 on proficiency assessments.

The explanation appears straightforward: corporate initiatives continue to emphasize access, risk mitigation and elementary prompting – providing employees with a language model, establishing guardrails and teaching basic prompt structure. These elements are foundational but insufficient for transforming how work gets executed.

For Canadian financial planners, this pattern explains how a firm can credibly announce an "AI‑ready workforce" while individual advisers continue constructing plans and documents using virtually identical methods to those employed three years earlier.

Leadership optimism sharply diverges from frontline reality

Among the study's most politically sensitive findings is the pronounced gap between executive perception and employee experience.

C‑suite respondents express overwhelmingly positive assessments:

• Strong majorities report their organizations possess clear AI strategies, transparent tool access and effective usage policies.
• Many indicate employees are actively encouraged to experiment and develop their own AI applications.
• Nearly half perceive widespread adoption and robust sharing of best practices.

Individual contributors – employees without direct reports who perform much of the organization's operational work – paint a markedly different picture.

On critical questions, perception gaps are dramatic:

• 81 per cent of C‑suite leaders affirm, "We have a clear, actionable policy that effectively guides AI use," compared with 28 per cent of individual contributors – a chasm of 53 percentage points.
• 80 per cent of executives report that tools exist with transparent access processes, versus 39 per cent of ICs.
• 71 per cent of leaders confirm a formal AI strategy exists, versus 32 per cent of ICs.
• 66 per cent of executives feel encouraged to experiment and build AI solutions, versus 25 per cent of ICs.
• 48 per cent of leaders perceive widespread adoption with open sharing of practices, versus 8 per cent of ICs.

Simultaneously, 75 per cent of C‑suite members report excitement about AI's implications and 94 per cent express confidence in its contributions. Most executives (57 per cent) use AI daily for work; only 2 per cent avoid it entirely.

For Canadian wealth management firms where principals and senior partners may be personally enthusiastic AI adopters, this optimism is comprehensible. However, when success gets measured primarily through licence counts, login statistics and training completion rates, leaders can easily overlook the reality that planners and support personnel still find AI confusing, peripheral or irrelevant to their substantive responsibilities.

The workers who could benefit most receive the least support

The research also documents a structural inequity with significant implications. Individual contributors – employees without supervisory responsibilities – are least likely to benefit from organizational AI resources, despite frequently performing the most repetitive and automatable work.

The disparities are pronounced:

• Only 32 per cent of individual contributors report clear access to AI tools, compared with 57 per cent of managers, 72 per cent of directors and vice presidents, and 80 per cent of C‑suite executives.
• Just 27 per cent of ICs have received organizational AI training, versus 48 per cent of managers, 68 per cent of directors, 70 per cent of vice presidents, and 81 per cent of executives.
• Only 7 per cent of ICs receive reimbursement for AI tools, compared with 26 per cent of managers, 44 per cent of directors, 33 per cent of vice presidents and 63 per cent of C‑suite leaders.

Consequently, individual contributors report higher rates of anxiety and feeling overwhelmed by AI, lower trust levels, and minimal perception of transformative impact. Managerial support for AI usage among ICs has actually declined 11 per cent since May 2025. Only 7 per cent report that their managers expect daily AI usage, and roughly one third receive any encouragement to use it.

Within financial planning organizations, many individual contributors occupy roles ripe for AI transformation: paraplanners, client service coordinators, operations personnel and compliance staff. When these workers receive the least support and training, the opportunity cost becomes substantial.

Where financial services sits in the proficiency landscape

The study ranks industries on a 100‑point proficiency scale. Finance scores 36 points, trailing technology at 42 but ahead of consulting (35), manufacturing (34), media (33), real estate (32), food and beverage and education (both 29), healthcare (28) and retail (27).

Leading this particular pack does not equate to readiness. A score of 36 out of 100 signals that even apparent frontrunners remain in early stages of capability development.

Across business functions, engineering or technology roles achieve the highest proficiency score of 41, followed by strategy at 39, business development and sales at 37, and human resources at 37. Finance and legal functions score 35, product scores 34, operations 32, and customer service and support trails at 27.

Notably, even in functions with transparent high‑value applications, many workers fail to deploy AI for obvious tasks. The research documents that 54 per cent of engineers do not use AI for writing or debugging code, 56 per cent of marketers do not use it for creating initial content drafts, and 87 per cent of product managers do not use AI to develop prototypes.

The financial planning parallel is direct: advisers who do not use AI to analyze household cash flow patterns or model scenario outcomes, paraplanners who do not use it to synthesize lengthy regulatory documents, and compliance teams who do not use it to review advice documentation.

A practical agenda for Canadian planning leaders

The report concludes with leadership mandates that translate readily into actionable priorities for Canadian financial planning firms.

Abandon adoption metrics as primary success indicators.

When 55 per cent of your workforce uses AI weekly but only 15 per cent possess value‑generating use cases, your adoption statistics obscure reality. For planning organizations, meaningful metrics resemble time required per financial plan, days to complete new client onboarding, hours needed for review preparation, and the proportion of compliance tasks that have been partially or fully automated.

Elevate use case development to a managed competency.

The workforce is not paralyzed by technical inability; it is paralyzed by not knowing which problems AI can solve within their specific roles. Organizations can develop function‑specific use case repositories – distinct libraries for advisers, paraplanners, client service teams and compliance personnel – alongside role‑based implementation guides that articulate what "foundational", "developing" and "sophisticated" AI usage looks like for a Canadian financial planner.

Correct the support disparity affecting individual contributors.

The employees performing the most repetitive work receive the least access to tools, training and encouragement. This allocation is counterproductive. Prioritizing IC capability development, standardizing approved tool access across organizational levels, and requiring every manager to identify and monitor at least three AI applications per direct report would begin addressing this imbalance.

Acknowledge that training has merely established a foundation.

When trained workers score 40 out of 100 on proficiency measures, existing programs are teaching inadequate content. Training must evolve from generic prompt structure and risk awareness toward mapping actual workflows – from initial data gathering through final documentation – and demonstrating where AI can safely eliminate steps, accelerate analysis and improve consistency.

Bridge the awareness gap at the executive level.

When partners and senior leaders believe AI deployments are succeeding while frontline staff report negligible impact, both a data problem and a cultural problem exist. Regular skip‑level discussions focused explicitly on AI obstacles, combined with structured opportunities for leadership to observe how planners and support staff actually use (or avoid) AI in their daily responsibilities, would inject necessary realism.

Prepare for continuously rising standards.

The distance between casual experimenters and genuine practitioners will expand as AI capabilities advance. Establishing continuous learning infrastructure now – communities of practice, internal credential programs, peer coaching systems – will prove more valuable than additional generic vendor presentations.

The test that matters

For Canadian financial planners and the firms that employ them, AI has moved beyond theoretical consideration. The technology exists. Policies are appearing. Training sessions populate calendars.

The more difficult question is whether any of this is actually changing how financial plans get constructed, documented and delivered – and whether working planners believe they possess the use cases, tools and support needed to make AI more than an occasional convenience at the periphery of their practice.

An industry built on evidence and long‑term thinking may need to apply identical discipline to its own AI narrative: measure what genuinely matters, confront what is not functioning, and redesign the work itself rather than merely the presentation materials.