AI Roadmap for HR: 4 Paradoxes and 3 Adoption Profiles
How do you build an AI roadmap for the HR function?
The 1st Barometer of AI in HR in Canada (NexaRH and Parlons RH, 2026) highlights four paradoxes, three adoption profiles and marked differences with France. This final chapter offers concrete guidance to move from individual use to an organizational strategy.
AI roadmap for HR: the four paradoxes at a glance
A declared knowledge, but still superficial: HR professionals feel they know AI, but practices remain individual and fragmented.
Sensitivity to risks, but late governance: confidentiality is the top risk, while AI is mostly used through individual tools.
High expectations, but a lack of training: budgets and support mechanisms do not yet match expectations.
Already transformed, but with a lesser intensity: HR sees its own function as less transformed than most others.
6.1 The four paradoxes of AI in HR
A word from HR: Marc-André Lafrenière, Director, People and Culture Digital Strategy and Enablement, TELUS
The results of this barometer reveal an important insight: professionals are optimistic, yet organizations are at varying states of readiness. Technology is evolving rapidly, making it essential to evolve our business strategies in parallel.
For HR leaders, the pitfall is wanting to automate everything too quickly. The approach must be pragmatic, respecting organizational maturity. For example, instead of aiming for 100 per cent autonomous recruitment when neither internal capabilities nor candidates are ready, it is wiser to start with tools that assist humans (such as note-taking and scheduling). This gradual approach creates value without destabilizing the organization but requires a long-term roadmap that accounts for a gradual transformation.
We must also realize that AI does not just save time; it transforms the very nature of work. Take customer service as an example. If AI handles all simple issues, team members are left managing only complex situations all day long. The mental load increases, so organizations must proactively support the mental health and well-being of their team. Furthermore, it is no longer the same job; this requires different skills and an adapted work organization.
This is where the role of HR shifts. We must become work architects. Our mission is to deconstruct roles, task by task, to decide what should be automated and what must absolutely remain human, while aligning work design with fundamental human needs. This allows us to redesign work structures, adjust required skills, and drive the change needed for success.
Three years into the era of generative AI, we are still in an exciting phase of transformation. Success will not come from technology alone, but from our ability to evolve our processes, our culture, our human skills, and our tools in parallel.
Expert insights: Virginie Bernal, Organizational sociologist and founder of VIBE Momentüm
AI and HR: An Identity Paradox
Canada’s first AI in HR Barometer highlights a striking paradox. HR professionals are largely optimistic: over 91% view AI as an opportunity, and nearly 70% say they are confident about their organization’s future. Yet fewer than 20% report having a clear AI strategy, governance framework, or dedicated funding, suggesting that HR may not yet fully see themselves as part of the challenge.
This paradox points to a potential turning point in the evolution of the HR role.
Historically, HR has been built as the guardian of the organizational framework, rules, standards, equity, and compliance, playing a critical role in protecting people within systems that could otherwise become dehumanizing. That legacy continues to shape both practice and education, where the focus remains largely on securing and structuring organizations, while strategic positioning often remains secondary.
In the context of AI, this posture persists. The Barometer shows that HR functions are primarily mobilized around data privacy and skills development, far more than around governance, value creation, or major technology decisions. They equip. They support. They frame. But AI is not simply another issue to manage. It is transforming how work is organized, redefining roles, and reshaping power dynamics within organizations.
As a result, the HR function finds itself in a singular contradiction: being accountable for people, without being fully co-accountable for the decisions reshaping work. AI requires rapid trade-offs, technological, human, ethical, and economic, yet HR professionals remain largely engaged downstream, rather than at the point where these decisions are made.
This paradox calls for a collective rethinking of HR’s role, posture, and place in the decisions that are reshaping work itself.
The question is no longer whether AI represents an opportunity, but whether HR leaders are ready to fully assume their strategic responsibility in the transformations it is driving.
As systems continue to gain intelligence, the value of organizations will increasingly depend on their ability to redefine what humans create, influence, and transform.
The question is no longer: Are we in favor of AI? It has become: Are we, as HR leaders, ready to redefine our strategic role?
Continuing to equip the company without transforming our own approach means accepting that the transformation will happen without us, or despite us. AI forces HR to choose: to stay the guardians of the existing framework or to become architects of human transformation.
In this revolution, the challenge is not to adopt AI. The challenge is to dare to become the role that AI now makes indispensable.
6.2 What is your adoption profile?
The Barometer allowed us to outline three distinct profiles. Their behaviors and decisions regarding the adoption of AI in HR differ.
Which profile do you identify with the most?
The three adoption profiles at a glance
The Québécois, optimistic but cautious: confident about the future (70%), clear-eyed about deployment difficulty (71%), moving forward through local tests, held back by the lack of a clear strategy.
The Canadian, pragmatic and structured: better equipped (56% governance, 40% trained teams), executing quickly with limited resources, at the risk of staying isolated in regional initiatives.
The International, innovative and strategic: 65% have a designated AI leader and 48% established governance; the risk is imposing standardized solutions poorly suited to local contexts.
6.3 Ten indicative data points to complete this profile
1. The “forgotten third”: 33% of Canadian organizations have neither a strategic plan, nor governance, nor a team trained in AI. A third of companies are still at the starting line!
2. “Skinny knowledge”: 55% of HR professionals in Canada believe they only have a basic understanding of AI. Expertise remains scarce.
3. Recruitment objective: 52% of organizations use AI to write job descriptions, making it the most mature application area in HR.
4. Payroll… lagging behind: 64% of respondents are unaware of AI applications for payroll and benefits. Payroll is the poor relation of AI in HR.
5. The guardians of the temple: 61% of HR departments play an active role in protecting personal data related to AI, confirming their ethical stance.
6. Prompting, the key skill: 58% of respondents identified "prompting techniques" as a priority skill to develop, highlighting the importance of knowing how to interact with AI.
7. The number one obstacle isn't money: the lack of a clear strategy is cited as the main obstacle to AI deployment (41%), well ahead of budget constraints, except in certain sectors such as hospitality, catering, and events, and personal services, where budget constraints are a greater obstacle (29% and 40% respectively).
8. Budgets under pressure in two sectors: 28% of organizations in the hospitality, catering, and events sectors anticipate a reduction in AI investments, a level well above average, compared to 14% for HR technologies and 28% for AI skills development. In the public and parapublic sectors, this anticipated decrease would affect 15% of organizations, reflecting tighter budget constraints due to the economic climate and public spending reduction policies.
9. Less confidence in the future for these sectors: 47% of respondents in the public and parapublic sectors also express greater uncertainty about the future, stating that they lack confidence (18%) or do not know what to expect (29%). In the industrial and construction sectors, 35% of respondents are pessimistic or undecided about future prospects (15% and 20% respectively).
10. A tie between Canada and France: When assessing individual AI use, Canada is ahead of France, both in terms of simple tools and automation and advanced tools (generally between 5% and 9% difference). However, collective use, meaning integration into organizational practices, is less widespread or the same as in France, indicating that organizational AI is not as integrated here as it is in France. It's worth noting that the HR Barometer in France was conducted six months earlier than the Canadian version.
6.4 The main differences with France
While the French and Quebecers respond very similarly to many questions, some major differences stand out:
Attitude (Q25)
Canadians see AI as an opportunity at 91% compared to 82% for the French.
Training (Q18)
Overall, French HR respondents reported offering more AI training than Canadian respondents. Depending on the topics, the French declared 10% to 25% more training sessions than Canadians, even though the Canadian survey was conducted six months later.
Risk factors (Q21)
Among the risk factors, Canadians cite thirdly the multiplication of AI uses and tools without overall coherence, which could complicate AI governance (sources, data quality and security, complex architecture, confusion of uses). The early grey area of AI usage could give way to an abundance and confusion of uses, requiring choices, rationalization, mapping, and clarity.
For the French, the third risk to consider was the dehumanization of HR relationships between managers and employees. Judged as very risky by 21% of the French, only 14% mention it in Canada, placing it fourth on the risk scale.
Barrier (Q22)
With 76% of respondents, the absence of a clear strategy to deploy AI in HR is by far the primary barrier identified. It is also the top barrier for the French, but at a level 24 points lower. HR teams on both sides of the Atlantic feel the need to build an AI roadmap in HR. This seems urgent for Canadian HR.
Canadians ranked budget constraints third among barriers (32%), while the French only placed it fourth. For the latter, resistance to change from their peers would be a more significant barrier (44%).
Use case (Q12 to Q17)
Out of 31 use cases surveyed, Canadians report using AI more often than the French:
6.5 Trends to keep on your radar for 2026
1. Sovereign AI
Sovereign AI encompasses AI architectures, models, and data hosted, governed, and controlled locally in accordance with Canadian regulatory frameworks. It aims to reduce dependence on global platforms, secure sensitive data, and ensure control over critical algorithmic decisions.
HR handles some of the most sensitive human data. In 2026, AI sovereignty becomes an issue of compliance, social trust, and business continuity. HR decision-makers will need to balance model performance, dependency on global digital infrastructure providers, and legal requirements. Sovereign AI is not an ideological choice but a risk management lever.
2. AI-Ification of Everything
AI-Ification refers to the systemic integration of AI across all business processes, tools, and roles, including in civic, cultural, and private life. AI is no longer limited to isolated use cases but becomes a transversal layer of automation, augmentation, and orchestration of HR decisions, integrated into daily operations.
In 2026, HR will no longer choose “whether to use AI” or not. They will have to manage an HR function fully augmented by default. Without an overall vision, AI-Ification creates an accumulation of tools, decision debts, and operational risks. The real challenge is to govern the coherence, value, and accountability of this omnipresent AI.
3. Agentic AI
Agentic AI relies on autonomous agents capable of performing complex tasks, collaborating with each other, and acting directly within HR systems. It transforms HRIS from Systems of Record into Systems of Action, capable of orchestrating end-to-end HR processes.
Agentic AI enhances HR value by executing complex processes. In 2026, the question is no longer “what to recommend?” but “what to allow to act?”. HR will need to define clear boundaries between AI agent autonomy, human control, and decision-making responsibility. With a clear framework, productivity gains should not become a major organizational risk. HR will activate Agents in “AgentStores” as they currently do with apps on their iPhones. Tomorrow, we will speak of “Agentic Organization”.
4. Explainability and Traceability
Explainability refers to the ability to understand, document, and justify decisions made or influenced by AI. It includes traceability of data, models, rules, and arbitrations to make HR decisions auditable, defendable, and understandable for stakeholders.
HR decisions involve fairness, reputation, and legal accountability. In 2026, a high-performing but opaque AI becomes a liability. HR will need to demonstrate why a decision was made, not just that it is “optimal.” Explainability becomes a cornerstone of the managerial and social legitimacy of AI in HR.
5. The “first and last mile” of human involvement
The first human mile corresponds to complex tasks of choice, assembly, and business intent that AIs cannot fully assume. The last human mile corresponds, on the other hand, to contextual judgment and final decisions that AIs cannot fully assume. These are ambiguous, political, human, or ethical situations where responsibility, judgment, and intent remain fundamentally human.
As AI consolidates, prioritizes, and recommends, the human role shifts upstream (architecture) and downstream (decision-making and risk acceptance). In 2026, HR’s value will lie less in execution and more in the ability to carry out this last mile. Neglecting this space is to dehumanize the HR function and undermine its strategic credibility.
The 7 parts of the AI in HR Barometer 2026
Chapter 6. Thoughts for an AI roadmap in HR (you are here)
Frequently asked questions
What are the four paradoxes of AI in HR?
A declared but superficial knowledge, a strong sensitivity to risks but late governance, high expectations but a lack of training, and a function already transformed that sees itself as less exposed than others.
How do you start an AI roadmap for HR?
Start from the problems to solve rather than the tools: clarify strategy and governance, train teams, launch measurable low-risk pilot projects, then scale the use cases that prove their value.
What are the differences between Canada and France on AI in HR?
The Barometer compares its results with the French survey by Parlons RH. Canadian HR teams use AI more for pre-screening candidates and producing training content, with gaps of 15 to 20 points on these uses.
Source and methodology
Source: 1st Barometer of AI in HR in Canada, 2026 edition, published by NexaRH in collaboration with Parlons RH. Pan-Canadian online survey conducted from October 1 to November 30, 2025 among 336 respondents (80% HR professionals within organizations, 20% HR service providers), in French and English, complemented by a discussion workshop with HR vice-presidents on November 12, 2025.
Read the full overview: 1st Barometer of AI in HR in Canada 2026.
How to cite this barometer
NexaRH and Parlons RH (2026). 1st Barometer of AI in HR in Canada, 2026 edition. Survey of 336 HR professionals, October and November 2025. Available at: www.nexarh.com/en/post/ai-hr-barometer-canada-2026
Comments