Across a range of projects, I have worked alongside Human Resources professionals. Whether establishing recruitment and performance management practices in smaller organizations, transforming employee benefits, or redesigning recruitment processes and technology, I have applied a culture-informed design lens.
A Recruitment Challenge
While it is difficult to quantify how many Canadian organizations use artificial intelligence specifically for applicant screening, adoption may be increasing. A 2025 Statistics Canada report found that text analytics was the most widely adopted AI application, increasing by nearly nine percent year over year.1
As AI-powered applicant screening tools become more common, it is critical to recognize that technology alone does not resolve recruitment challenges. Without human-centred design, clear process ownership, and trust in the system, AI-enabled hiring can reinforce inefficiencies and inequities rather than improve outcomes.
Most AI-powered applicant screening tools rely on text analytics, using keyword matching and linguistic analysis to compare resumes against job requirements. Some tools also incorporate analytics to validate employment timelines or identify patterns associated with specific skills. When designed and used well, these tools create efficiencies, reduce administrative effort, and support consistent screening practices.
Observation
“We need AI to help screen and shortlist applications.”
A high volume of resumes may seem preferable to a lack of qualified candidates. However, when recruiters can spend only a few moments scanning each resume, it becomes difficult to review applications thoroughly. At the same time, a lengthy time to hire can cause strong candidates to accept offers elsewhere. Efficient recruitment saves time and money.
In this case, the organization was experiencing an influx of resumes per posting. The existing hiring tools relied on screening questions embedded in the applicant tracking system, but hiring leaders and recruiters often second-guessed the automated results, suggesting a potential design problem. Some are reviewing all resumes, including rejected ones, resulting in inefficiency and inconsistent screening practices.
A broader structural issue was that hiring leaders were often responsible for screening and shortlisting candidates using these tools. While this approach may work in smaller organizations, larger organizations typically benefit from a centralized recruitment service model. These models improve consistency, efficiency, and the application of specialized recruitment expertise. Poor hiring decisions – particularly when candidates lack the required qualifications or experience – carry significant costs in both time and resources.
“We Need AI” Is Only Part Of The Solution
“We need AI” is an understandable request today. Business culture is saturated with promotion for artificial intelligence. Have you been on LinkedIn lately? Major consulting firms, IT services groups, and educational institutions are heavily marketing their products and services to meet the demand generated by large language models. Some executives suggest that teams will fall behind if they do not use AI in some way.2 The pressure – and opportunity – to adopt automation and AI is real.
In this case, “We need AI” led to problem definition and workflow mapping. This design-focused approach is important and will help the organization avoid a common pitfall of buying the first promising solution. Off-the-shelf products do not always fit complex enterprise environments, nor are they always designed for unique organizational structures.
Effective AI enablement should start with fundamental questions:
- Problem definition: What problem is being solved? Is it systemic, such as consistently high application volumes, or localized to a specific business unit? Is there an opportunity to re-imagine the entire workflow or even function?
- End-users and ownership: Who is the solution designed for? Recruiters, hiring leaders, or applicants? How will process ownership and accountability be defined? Where is the “human-in-the-loop” most imperative in the work?
- Change and adoption: What organizational changes are required to support adoption? How will trust in new tools be built? Will users receive training, and how will adoption and value be measured? For instance, do hiring leaders or recruiters still review applicants who failed screening?
These questions reflect human-centred design: prioritizing human needs, not the technology. Organizations should observe how tools are actually used in practice and incorporate design-led inquiry into AI initiatives from the outset. Without addressing these and other questions, organizations risk implementing technology-driven solutions that fail to deliver value.
If your hiring process (or any business process) seems broken, adding AI alone won’t fix it.
Design will.