The playbook for building high-talent-density teams | Adam Ward, Head of Talent at Cursor

Ward calls the conventional recruiting funnel a “funnel of doom.” Companies contact a broad pool, treat response rates as if they reveal quality, then eliminate candidates at each stage and hire whoever remains. A 20% reply rate from a list of 100 does not identify the top 20%; it identifies the 20 people who happened to respond. He proposes beginning instead with an explicit thesis about who the strongest candidates are, then using assessment to test that thesis.

That shift starts with scoping: define the skills and experiences that matter, rank them, and state what success in the role means before speaking with candidates. The specification should describe the company’s needs rather than borrow another company’s definition of excellence or rely on résumé logos. Ward then pairs the scope with market mapping, using trusted people and specific questions to identify candidates whose work matches the criteria. He favors asking who is especially collaborative with designers or unusually good at turning frameworks into products over asking who is simply “the best.”

The intended result is a small, reasoned target list—perhaps 50 people for a leadership search—rather than a large pool whose composition is determined by who applied or replied. Ward says this can work at scale when the scope is clear and the team concentrates its effort on those people. The approach changes where the work happens: less indiscriminate top-of-funnel volume, more deliberate research and ongoing outreach.

Ward argues that recruiting cannot be delegated in the same way a company might delegate a transactional process. Hiring managers and leaders make the decisions, so recruiters should build confidence and provide useful evidence rather than carry sole accountability for the hiring outcome. He recommends keeping hiring visible through measurement and accountability, and involving the people who will feel the effects of a good or bad hire.

At Cursor, employees can post promising people in a shared “Hiring Ideas” Slack channel. A post can trigger colleagues to explain what they know, suggest how to approach the person, and help bring them into the process. The ideas are not always right, Ward cautions, but this creates an additional source of candidate leads and makes talent discovery part of the company’s shared work rather than an activity initiated only by recruiters.

For candidates who are not ready to move, Ward recommends staying in contact over weeks, months, or sometimes years. A “not now” can lead to a coffee, a visit, or a conversation with a teammate rather than ending the relationship. He also stresses individualized outreach: learn what might interest the person, find a credible warm introduction when possible, and keep showing care. His claim is that candidate experience improves when people feel wanted and recognized as individuals, instead of being treated as units in a volume funnel.

Ward describes work samples and side-by-side projects as strong ways to gather evidence about performance. He also wants candidates to learn from the process: carefully chosen teammates, conversations, meals, and practical work can help a candidate decide whether the company and its people are a fit. In that sense, evaluation should support self-selection as well as the employer’s assessment.

The format should fit the role. Cursor uses versions of practical challenges for go-to-market and product candidates, and Ward says the company gives candidates choices, adjusts the scope where possible, and accommodates scheduling constraints. He acknowledges the substantial time commitment. When Cursor experimented with removing work trials, the team found it harder to feel confident in its decisions and brought them back. Ward also notes that the process can surface collaboration and values over time, not only technical skill.

The hiring manager’s first conversation deserves particular attention. Ward sees it as a pivotal moment to ask important questions and understand the candidate’s motivations, rather than treating it as a brief screen before the “real” onsite. If the candidate is one of a small number of carefully identified prospects, the hiring manager may meet them several times before a formal interview process begins.