Why Monks built a chatbot recruiter based on its founder — and how it signals an agency shift

Forget the standard job posting, Monks is auditioning its next AI leaders through an AI of its own creation: an AI recruiter personifying the agency’s co-founder.
The S4-owned agency recently debuted a new chatbot called “WesleyBot” that serves as both recruiter and marketer. The prototype is based on the mind and personality of MediaMonks co-founder Wesley ter Haar, who recently took on a new role as the chief AI officer.
Designed to attract top AI talent, WesleyBot operates as a conversational interface within ChatGPT for the AI-ambitious who feel like “lone wolves” in a traditional agency or consultancy. Part of the plans has meant creating a paid campaign on LinkedIn targeting mid- to senior-level talent who have media, data, or AI tooling skills across rival companies — especially in cities where Monks also has an office. So far, more than 600 people have chatted with WesleyBot, according to Monks.
Beyond just targeting engineers, ter Haar said his agency wants to reach people who feel stuck in what they’re allowed to do — and what AI tools they’re allowed to use.
“I talk to many of them and there’s one or two people carrying the whole AI road map for a global network and they don’t get the support that they should be getting,” ter Haar told Digiday.
The strategy behind WesleyBot
WesleyBot is fine-tuned for ChatGPT, but users can have a similar conversation on other platforms like Gemini and DeepSeek using a 1,200-word system prompt crafted by Monks to mirror the chatbot’s tone and behavior.
Each interview involves a six-question conversation where candidates can show their AI knowledge, creativity, adaptability and thought process. Questions are curated from nearly two dozen that range from ice-breakers to work scenarios to test a person’s creativity, adaptability and thought process. Examples include:
- What’s an AI prediction that might age terribly?
- Where is AI close to a breakthrough?
- What’s the first project you’d launch with the ultimate AI toolkit?
- If you could use AI to flip the creative process on its head, what would you tweak, break or rebuild to make it better?
The personality of WesleyBot is designed to be “sharp, witty, and insightful,” via a direct but friendly Dutch tone, that challenges candidates “in a fun, engaging way while keeping the conversation dynamic and welcoming”, according to the Monks system prompt reviewed by Digiday. The bot also evaluates people for their “Monk-ness,” which is described as curious, adaptable, collaborative, resilient and proactive.
Behind the scenes, Monks used Google’s Gemini and Workspace tools to design a backend evaluation workflow that analyzes candidate resumes, transcripts and email notes. A scoring agent then ranks applicants from 0 to 100 based on the aforementioned criteria with results reviewed via Google Sheet dashboard to help human recruiters prioritize outreach. WesleyBot is also a proof of concept that could be scaled across the agency to help recruit for a range of roles.
Monks’ AI evolution
WesleyBot is part of the third year of Monks evolving to adapt in the AI era, which also includes ter Haar in February adding chief AI officer to his other role as chief revenue officer. The agency is also in the process of breaking down siloed teams and instead creating “cross-functional squads” to replace isolated expertise with collaborative teams.
Each squad is anchored by a technologist, who helps accelerate creative output by turning ideas into execution in days instead of weeks. Squads include roles that reflect a broader talent shift toward building a “brand model practice” as AI becomes central to the agency’s future work — roles focused on model-tuning, reinforcement learning, knowledge base development, and brand voice training,
Monks is also exploring ways to move beyond the traditional agency model of hourly billing with the goal of shifting to an output-based model that focuses on the quality and impact of work.
As clients gradually adopt these new models, ter Haar sees it reshaping the link between revenue and talent. He also thinks it could move agencies away from the traditional “unhealthy” cycle of scaling headcount up or down with revenue that undermines agency stability, culture and quality.
“A huge opportunity here is to decouple the revenue and cost,” ter Haar said. “You could have revenue growth without the same growth in talent. … We’ve been talking a lot about a software and service company, which is also about moving away from this very body-shop-like nature of our industry at scale, which I don’t think is good for anyone.”
Companies often mistake LLMs for products rather than infrastructure, ter Haar said, adding that reasoning models are more powerful. Reasoning models differ from earlier LLMs by going beyond pattern recognition to perform structured, multi-step human-like thinking to follow logic and solve problems. However, he also noted that text-based LLMs don’t scale.
“Where a lot of people have taken the wrong turn is they’re expecting these crazy results by just having some off-the-shelf software in this space that does almost nothing for people,” ter Haar said. “If you use off-the-shelf, people [say] ‘this is crazy that this is possible, but I also don’t know how to use it in a meaningful way.'”
As more agency work becomes computable and budgets shift to strategic priorities, Monks is guiding clients toward agentic systems like its Monks.Flow platform. ter Haar recalled several recent meetings where he told clients they won’t likely need Monks in a few years once the work agencies now pitch are no longer needed.
“The main difference I would say is we’re being very clear and transparent about changing the economics,” he said. “My personal take is that an AI story that doesn’t change economics isn’t anything more than a story, so we’re being more aggressive around changing the model for our clients.”
A recent report by Gartner predicted 30% of business roles will require no-code skills by 2028 while 20% of digital workers will build their own AI agents. However, the report also warned of potential “agent anarchy” without proper planning and supervision if unmanaged agents clash across enterprise applications and automated systems.
As agentic AI becomes embedded in service businesses, agencies must rethink their models, said Gartner analyst Nicole Greene. She thinks the shift will help streamline workflows, speed up ideation, and meet growing demands for faster, cheaper better and better content. Recent acquisitions of major data companies by major holding companies like WPP and Publicis Groupe also show how fast the landscape is changing, according to Greene.
“Monks recognizes that the agency model is changing, and they’ll need to find creative ways to use AI to drive growth,” Greene said. “Clients are going line by line to reduce costs and bring capabilities in house because their internal capabilities are improving due to the impact of AI. The unique position of agencies will be focused on strategic creative thinking, a big shift from the reality that a lot of dollars come from operational creative tasks.”
More in Media Buying

Media buyers look for ways to stay ahead of clients’ tariff budget pressures
Some media agencies are reconsidering the ways they invest media budgets, and focusing on direct deals and curation to drive efficiencies.

Ad Tech Briefing: Is Google’s curation tool a play for second-mover advantage or an antitrust swerve?
Is (part of) Google preparing to ‘go it alone?’

WPP ups the data-driven arms race with the purchase of InfoSum
The deal represents a leveling-up on WPP’s part in the data-driven arms race among the agency holding companies.