Early Adoption of AI Agents in Recruitment

AI agents are emerging as powerful tools in recruitment, automating tasks, enhancing efficiency, and improving the candidate experience. Although a relatively new solution category, we are seeing strong adoption of recruiting agents among earlier stage, high-growth firms that have adopted such solutions in lieu of a more traditional recruiting function. That said, we are also seeing significant interest from larger, complex, global organizations that are looking to either pilot or invest in recruiting agents to automate administrative tasks, create more efficient processes, or to reimagine the TA process entirely with new approaches that are now only possible because of agentic recruiting solutions.

We have spoken with many providers and early adopters of these solutions, and these are the most common use cases currently for AI agents in recruitment:

AI Agents for Sourcing and Administrative Support

The most accessible AI Recruiter vertical are agents designed to automate manual or administrative tasks and functions within talent acquisition, such as sourcing, invite-to-apply, prescreening, and interview scheduling. These AI workers typically don’t perform candidate-facing tasks, such as voice/video screening or interviewing, or chat-based conversations. Instead, they orchestrate and help generate communication sent on the recruiter’s behalf.

A great example of an AI recruitment assistant is Lucy, the first digital worker from Cykel AI. Launched in November 2024, Lucy is specifically designed to support candidate sourcing, screening, and outreach. Cykel AI is making waves in the fast-growing AI worker sector, after having shifted from a general AI assistant platform to creating AI workers tailored for specific business needs. Soon after Lucy came available for hire, Cykel released Eve, an AI sales agent designed for B2B sales, and Samson, an AI agent designed for research analysis.

Cykel has been on a rapid trajectory for growth and development, with an oversubscribed equity raise in February that has funded a number of recent enhancements. In May, the firm integrated Deep People Search, a tool that harnesses neural search technology to identify optimal candidates from over 300 million profiles. In late summer 2025, Cykel plans to release new preboarding capabilities for Lucy to help coordinate administrative tasks for new hires.

As of June 2025, Cykel was trading on an impressive 68% month-on-month growth illustrating the emerging appetite for digital workers. Cykel’s growth fits with the broader AI agent market momentum, and it also fits with our own recent survey findings: roughly two-thirds of organizations have or plan to pilot agentic AI recruiting capabilities in the next twelve months. Cykel’s acceleration from demo activity to revenue conversion even further suggests the market is transitioning from experimentation to production.

When Lucy was first launched, Cykel expected to see the most interest from the SMB segment, but the product is gaining traction across a range of organization sizes and types: from scale-ups with under 1,000 employees to large corporations with up to 20,000 employees. Cykel is also working with some of the largest staffing firms in the UK who have Lucy screening thousands of candidates per week. This is not surprising, as we find that larger organizations with high applicant volumes are often looking for agile recruitment tools that can augment their existing tech with minimal friction.

Balancing Innovation with Compliance and Brand Experience

On the other hand, large enterprise organizations also tend to have greater concerns about AI use, especially when it pertains to talent acquisition. Organizations that want to adopt AI recruiting agents often wonder how they can maintain their “human recruiter” standard, which may be a key strategic pillar in talent brand and experience. Like most AI recruiting agents on the market, Lucy is designed to handle certain parts of the talent acquisition process (e.g. sourcing, outreach, prescreening, and interview scheduling) but does not handle end-to-end recruitment. Cykel made a conscious choice to develop an agent that could take on manual and administrative tasks of recruiting but would not speak directly with candidates or conduct interviews and assessments. Therefore, Lucy would not replace the human recruiters’ role in building relationships and evaluating talent. Other builders of AI recruiting agents, such as Symbal.ai, include AI phone screening capabilities, as well as what Cykel offers.

Another area of concern for many organizations regarding AI recruiting agents is of course the legal and compliance risk. It’s a highly nuanced and complex subject that we cover a great deal in our research and advisory work, but for the purpose of this article, we find that solutions like Lucy have a lower AI risk profile because:

• It does not use advanced AI to evaluate candidate-job fit

• It does not communicate directly with candidates

• It has no autonomous learning or decision-making capabilities

The agent can be configured to move candidates through a process in the same way that the human would, because it takes orders and directions from the recruiter and automates these tasks on their behalf. As a result, Cykel has successfully demonstrated in the most risk-averse industries and compliance-heavy jurisdictions that it is possible to have a compliant process using an AI recruiter.

AI Interviewers: Expanding into Candidate Assessment

Another set of early adopters is using video and voice AI interviewers – highly specialized agents capable of handling comprehensive first-round interviews (e.g., ~35 minutes, 8-10 questions). For example, Apriora’s AI Interviewer, Alex, can ask, understand, and evaluate behavioral questions, competency-based questions, and questions about a candidate’s background, skills, and motivations. It can also conduct technical assessments and behavioral assessments.

Apriora was founded in 2023, initially planning to sell into the staffing industry but soon expanded into the corporate enterprise segment. Similar to other agentic recruiting solutions, the interviewer is able to adapt its responses based on how an interviewee answers a question. The solution is predominantly used for technical interviews and early career recruitment, but the firm is currently investing in better calibrating its interview scoring, which may help to expand into other types of hiring.

Prudential recently introduced Apriora to its summer internship hiring process, having previously used HireVue’s video interviewing platform for these roles. Candidates complete their first-round video interview with Apriora’s AI, where they are asked a combination of behavioral, skill-based, and experience-based questions. For a large financial organization such as Prudential to adopt an autonomous AI video interview is a significant milestone. We’re seeing other large enterprise organizations leaning toward AI video interviews to replace their current pre-recorded video interview solutions. Video AI interviewers are a logical progression of the traditional pre-recorded interview method with the advantage of a more interactive, two-way format that may elicit a more authentic response from candidates.

Navigating Risk and Reputation in AI-Based Hiring

What makes interviewing agents a higher-risk use case than sourcing agents is that the former uses AI to evaluate and score candidates based on grading criteria assigned to each interview question. For example, Apriora’s AI evaluates candidate responses based on the customer’s pre-defined, customizable metrics aligned with the job requirements. The customer can also provide instructions to explain how the AI should evaluate a candidate’s response. Although this approach gives the customer more control and transparency over the scoring algorithm, there is some degree of AI bias risk whenever the technology is used in the candidate evaluation process. To minimize compliance risk, some organizations using AI interviews chose not to enable the AI evaluation feature. Providers such as HeyMilo give their customers the option to disable AI interview scoring and use its platform to store interview data for review by human recruiters.

Beyond the risks of using AI-based selection tools, there are also some inherent risks to adopting a cutting-edge product in terms of performance and consistency that can affect employer brand and reputation. In May, a video of a glitching AI Interviewer went viral on TikTok which exposed some poor experiences that candidates have with these products. The malfunctioning interview was hosted by Apriora, but the negative sentiment is generally directed toward the employer rather than the tech provider.

At Talent Tech Labs, we are still seeing that risk-averse organizations are more hesitant to adopt AI Interviewers compared to other types of recruiting agents, but as demonstrated by companies like Prudential, the adoption curve for advanced AI products is getting shorter with every major release.

If you would like to learn more about how Talent Tech Labs can support your talent technology strategy, join the Talent Tech Labs community or contact us at hello@talenttechlabs.com. Also follow us on LinkedIn to stay informed on the latest talent technology insights and updates.