Testing Innovative AI for Career and Skills Navigation

The Issue

Jobseekers are navigating a labour market characterized by technological change, shifting skill demands, and a growing expectation to be able to identify, explain, and update their skills across different roles and sectors. On the other hand, many organizations that provide career development services face constraints in their ability to offer the level of individualized, repeated, and timely support that this context requires. 

This issue is important because career navigation is not only a matter of finding available job postings. It also involves interpreting labour market information, understanding how a worker’s prior experience can be presented to employers, identifying skills gaps, preparing applications, and building confidence for interviews. For equity-deserving groups such as newcomers, youth, Indigenous people, or persons with disabilities, these tasks may involve additional challenges including limited networks, uncertainty about Canadian workplace norms, credential recognition issues, or less confidence in translating past experience into employer-facing language. 

Career development practitioners occupy an important position within this system. Their work often requires judgment, local labour market knowledge, understanding of clients’ circumstances, and the ability to help jobseekers reflect on career goals that may not be obvious in a resume or job posting. However, when practitioners have high caseloads or limited time, some forms of support, such as repeated mock interviews, individualized resume tailoring, or exploring alternative pathways, may be difficult to provide at the level required. 

Generative AI adds another layer to this issue. Jobseekers and practitioners are already using generic AI tools for resumes, cover letters, interview preparation, and career planning. Yet these tools are not necessarily designed for career development practice. Nor do they always provide transparent, accurate, localized, or context-sensitive guidance. As a result, the central question around AI use in career services is not simply whether it can be used, but under what conditions it can be used responsibly to expand access to support while preserving the human expertise and trust on which career development depends. 

PathPilot was developed to explore that question through a staged approach. The first phase sought to validate whether GenAI-powered career navigation could provide useful and tailored support to diverse users. The second phase aimed to scale that work by expanding the user base, engaging more partners, and examining how an AI career navigation tool could fit within the broader career development ecosystem.

What We Investigated

This project investigated the feasibility, value, and implementation conditions of using generative AI to support career navigation. It did so across two related phases. 

The first phase ran from October 2024 to January 2025 and was designed as a validation project to assess whether GenAI-powered tools could provide effective career development support for equity-deserving groups, including newcomers, youth, Indigenous people, and others. This phase aimed to engage 500 users, provide tailored career development support, and assess the impact of AI tools on user satisfaction, time to employment, and job/interview readiness. It also explored, in consultation with employment service organizations, opportunities and barriers associated with deploying AI-powered enhancements for skills and resume analysis, coaching and preparation, and career navigation. 

Building on this work, the second phase sought to expand PathPilot into a broader career navigation platform serving jobseekers and Career Development Practitioners across Canada. The platform included functions such as job discovery, resume optimization, cover letter generation, upskilling recommendations, interview preparation, and mock interview practice. It also included practitioner-facing activities. These included AI literacy and onboarding supports, as well as feedback sessions designed to understand how practitioners might use the tool with clients or to prepare for client meetings. 

The evaluation focused on both individual and system-level learning. It examined:

• The conditions under which jobseekers, including equity-deserving groups, found value in AI-supported career planning
• How users could access career and job discovery information that was trustworthy, accurate, and reliable
• How the needs of Career Development Practitioners differed from those of individual users
• How PathPilot could create value and efficiency in employment, education, and training service delivery. 

The evaluation used a mixed-methods and developmental approach. Data were collected through user surveys, platform analytics, interviews, focus groups with jobseekers and career development practitioners, user testing observations, and review of the platform’s outputs. The project tested how users liked the tool and also how it functioned within a broader system of jobseekers, practitioners, partner organizations, employers, and technology developers.

What We’re Learning

The validation phase established the rationale for a larger scaling effort
The initial project tested whether there was enough promise, demand, and user feedback to justify further development. By focusing on 500 users, consultations with employment service organizations, user acceptance testing, and recommendations for future phases, the validation phase helped position PathPilot as a tool that could expand toward a more structured exploration of AI in career services. The scaling phase was a deliberate attempt to build on this feasibility exercise. 

PathPilot was perceived as a useful additional resource for career exploration
In the scaling evaluation, 73% of jobseekers who responded to a project survey agreed that PathPilot was a helpful resource to explore career options. No respondents disagreed. While this finding does not show labour market impact, it does suggest that users saw value in having access to an AI-supported tool as part of their career planning process. In that respect, the project points to a broader need for tools to help jobseekers move from accessing general labour market information to more specific actions, such as identifying possible roles, understanding required skills, and preparing for applications.

Users valued concrete supports for the job-search process
Platform users responded particularly well to features that helped them complete practical tasks more quickly, including resume optimization, cover letter drafting, job matching, and interview preparation. These features reduced some of the friction associated with modern job search, where applicants are often expected to tailor materials for multiple postings and map their experiences onto employer language. Several users also valued the platform’s ability to support structured interview preparation, especially because mock interview practice is not always available through traditional service models. 

AI’s strength is its complementary support
One of the most important findings is that PathPilot was most promising when used alongside human expertise. AI and career development practitioners should not be seen as competing sources of support. A practitioner may help a jobseeker develop a strong general resume, clarify goals, reflect on fit, and interpret labour market options. The AI tool supports additional practice, targeted revisions, and support between sessions. This distinction matters because it reframes AI as a tool that can increase the amount and continuity of preparation available to jobseekers, and not as a replacement for human support. 

Newcomers and early-career jobseekers appeared to benefit most
Both jobseeker and practitioner feedback suggested that newcomers and early-career users may benefit most from PathPilot. These groups often have more to gain from support with resume conventions, interview preparation, job-search norms, and translating experience into language that fits employer expectations. However, the findings also point to limits for users in highly skilled, specialized, or niche fields. These jobseekers may require more precise labour market information, more localized credential guidance, and stronger sector-specific content than a general AI-supported career navigation tool might provide.

AI-generated materials need review and reflection
The project also surfaced important cautions. While AI can help generate or revise resumes and cover letters quickly, both users and practitioners noted that outputs may include inaccuracies, formatting issues, or language that may not fully reflect the jobseeker’s experience. Some practitioners raised the concern that if AI produces application materials too quickly, jobseekers may lose the learning that comes from articulating their own skills and achievements. The implication is that AI-supported career tools should not simply produce finished documents for users. They should also prompt reflection, encourage revision, and help jobseekers understand why particular changes are being made. 

Scaling AI tools requires trust, structure, and implementation support
The project reinforced the elements of successful implementation. Career service organizations require clear use cases, structured onboarding, privacy explanations, practitioner training, communications materials, and shared expectations about how the tool will be used. This became especially clear as PathPilot further developed with practitioner, organization, and employer-facing interfaces. The lesson is that responsible scaling is not only a technical process. It is also an organizational and relational process that depends on trust, workflow integration, and change management.

Why It Matters

The PathPilot project provides a concrete example of how AI can enter the career development ecosystem, and the conditions that make it useful. The findings suggest that AI can help jobseekers complete certain tasks more quickly, access more opportunities for practice, and engage with career planning in a more personalized way. However, these findings also make clear that AI’s value depends on how the tool is designed and introduced, how its outputs are reviewed, and how it is connected to the work of human practitioners. 

The staged structure of the project is also important. The initial validation phase allowed GuruLink to explore whether there was sufficient promise in GenAI-powered career navigation before moving into a more complex scaling phase. This sequence reflects a useful model for innovation funding in the skills ecosystem. Rather than treating technology as ready for scale simply because it exists, the project moved from feasibility to user feedback, to scaling, and then to deeper learning about implementation. In that respect, PathPilot illustrates the value of using early-stage funding to test assumptions before broader deployment.

For career service providers, the project highlights an important tension. Many organizations need ways to extend their capacity, especially when practitioners have limited time and clients require repeated support. AI tools can help by supporting preparation between sessions, generating first drafts, offering interview practice, and helping users explore job pathways. Yet, these same tools may introduce new risks if they are used without sufficient guidance. A resume that is generated quickly but contains inaccuracies, or that a jobseeker cannot confidently explain in an interview, may create more problems than it solves. Therefore, the practical implication is that organizations should identify where AI can responsibly support the parts of career development that are repetitive, time-intensive, or difficult to provide at scale.

State of Skills:
Unleashing AI into the Skills Development Ecosystem

FSC-supported AI tools have bolstered outcomes in skills matching, career development guidance, and recruitment. The overall effectiveness of these tools was underpinned by recognizing and mitigating the inherent bias and discrimination embedded into these technologies.

Read Thematic Report

For career development practitioners, the project suggests that AI can become a capacity-building tool when it is framed appropriately. If AI is presented as a substitute for human judgment, it is likely to generate concern and resistance. If, instead, it is framed as a tool that helps practitioners prepare for meetings, support clients between sessions, provide additional practice, and reduce some administrative burden, then it can strengthen rather than weaken the practitioner role. This distinction is important in a field where trust, listening, contextual interpretation, and values-based career exploration remain central to effective practice. 

For jobseekers, PathPilot shows the potential value of AI tools that make career planning more actionable. Many jobseekers need help connecting their job applications to their own background, goals, skills, and constraints. This is particularly relevant for newcomers and youth, who may be navigating career systems in which expectations are not always explicit. AI can help lower the initial barrier by generating prompts, examples, options, and practice opportunities. At the same time, the project reinforces that jobseekers must remain active participants, because career confidence is developed by learning to assess, adapt, and communicate one’s own experience, not just through AI outputs. 

For policymakers and funders, the project provides a useful caution about technology-oriented innovation. Interest from partners should not be confused with sustained use. Access to a platform should not be confused with integration into practice. Digital tools can be scalable, but their social and organizational adoption needs support. AI adoption in career services requires investment in training, onboarding, privacy and governance practices, evaluation infrastructure, and implementation supports. It also requires attention to equity, because tools intended to support equity-deserving groups must be tested for relevance, accessibility, trustworthiness, and the quality of guidance as well as usability. 

Overall, PathPilot points to a cautious but important conclusion. AI-enabled career navigation can be useful, especially for jobseekers who need timely and personalized support, but its strongest contribution may be in strengthening the relationship between jobseekers and practitioners. The project therefore provides an evidence-informed basis for thinking about responsible AI adoption as part of a broader human-centred career development system.

What’s Next

GuruLink is continuing to develop PathPilot v2, with interfaces for jobseekers, practitioners and organizations, and employers. Future development is expected to focus on several areas, including the employer interface, AI-supported note-taking, cross-platform integrations, practitioner-facing workflows, skills assessment tools, interest-based and psychometric assessment features, and more mobile-friendly forms of interaction. Continued work on trust and safety testing is also expected, including more robust and persona-specific testing of AI outputs. 

At the same time, the project team is prioritizing partner activation and implementation. This will require clearer partner launch plans, more structured onboarding, co-branded webinars, defined engagement targets, practitioner training, and stronger implementation data. The next phase will continue to assess if users find the platform helpful, as well as which partners, outreach strategies, practitioner workflows, and implementation conditions lead to sustained and responsible use. 

An investment from the Canadian Venture Capital and Private Equity Association is expected to support employer-facing aspects of the platform. Continued SCALE work with FSC is also expected to help onboard more organizational and individual users.