Video summary
Claude for Recruiters: The Skill That Formats Your Candidates CVs in 5 Minutes | Vanessa Raath
Main summary
Key takeaways
Summary
Vanessa Raath argues that recruiters should use generative AI to remove repetitive administration—not simply treat it as a search engine. One of the most immediate opportunities is turning raw CVs, interview notes, and transcripts into client-ready candidate documents. She also recommends using AI to improve sourcing research, email handling, and business development, while investing in the human relationships and professional reputation that AI cannot reliably provide.
Practical playbooks and workflows
Automate repeated CV formatting
Build a Claude Skill that accepts a raw CV, interview notes or transcript, and sometimes a LinkedIn profile PDF. It can produce a consistently formatted CV and an executive summary suitable for sending to a client.
Test the Skill on several examples, correct mistakes, and refine its instructions. Once it works reliably, recruiters can reuse it without restating the process each time. Raath says this can take about five minutes, compared with turnaround times of 24–48 hours in some cases.
Treat AI adoption as workflow redesign
Start by asking: What repeatedly slows down the recruitment process, and where is human effort adding little value? Prioritize tasks that are both repetitive and time-consuming, then build and test an AI-supported process around them.
Keep recruiters focused on human work such as building trust, advising, and maintaining client and candidate relationships.
Use AI to improve prompts
Raath cautions that prompts of 15 words or fewer often lack enough context to produce useful recruitment results. Instead of immediately asking an LLM to find candidates, provide the job details and constraints, then ask it to:
- Identify missing information.
- Ask clarifying questions.
- Turn the answers into a detailed, reusable prompt.
Example criteria include location boundaries, candidate priorities, and relevant job-title synonyms. Review and compare the resulting information before using it for outreach or decision-making.
Distinguish between reusable Skills and client-specific Projects
A Skill is useful for repeatable workflows, such as formatting CVs or producing training notes. A Project can hold documents and context for a particular client—such as meeting transcripts, terms, and proposals—so the recruiter can refer to them together.
The speakers note that Skills and Projects can be organized differently depending on the user’s workflow; they are not mutually exclusive.
Build a recruiter’s context files
Raath recommends giving AI four concise reference files, often stored as Markdown files:
- Voice: Tone, style, and communication preferences.
- Services and differentiator: What the recruiter offers and what makes it distinctive.
- Ideal Client Profile (ICP): Who the recruiter serves.
- Proof: Recommendations, case studies, testimonials, and success stories.
These files can support more consistent outputs across tasks and tools. Raath says they can be transferred between LLMs rather than recreated from scratch for each platform.
She suggests using real examples—such as LinkedIn posts and messages—to establish a communication style more effectively than trying to describe it from a blank page.
Capture conversations once and reuse the information
Record and transcribe appropriate calls and meetings, then use the material for summaries, notes, proposals, blogs, or follow-up. This can reduce repeated note-taking and help prevent useful information from being lost.
For candidate information, Raath advises removing sensitive personal data—such as Social Security numbers, phone numbers, and home addresses—before uploading it to generative AI. She also recommends informing candidates when AI is used in the recruitment process and following applicable governance and privacy rules.
Use tools selectively
Raath recommends starting with one paid LLM subscription, choosing the tool that best fits the work rather than buying multiple services before identifying a need. She cites Claude as especially strong for preparing and editing documents in her experience.
She describes Exa as useful for finding information about people and potential candidates, and Tavily as useful for company and market intelligence. She recommends connecting such tools only when they add value to a specific workflow.
For email, she prefers dictating replies in Copilot in Outlook, then reviewing and refining the draft rather than accepting generic, robotic-sounding automated responses.
Strategy, marketing, and leadership
- Position AI as an assistant, not a replacement for recruiters. The goal is to let AI handle administrative work so people can spend more time building trust and advising clients and candidates.
- Make personal brand and credibility part of the business strategy. Raath recommends being visible on LinkedIn, sharing useful industry views, commenting thoughtfully, collecting client recommendations, and meeting people in person where practical.
- Treat in-person activity as business development. Raath describes international training and conference appearances as a source of new relationships and training engagements, while noting that online delivery makes global reach possible.
- Keep learning through experimentation. Start with a practical use case, test it, and improve the workflow rather than waiting to become an AI expert. Short workshops can help teams see what is possible, but employees need time to apply the tools themselves.
Metrics and figures mentioned
- CV workflow: Around five minutes to prepare a client-ready version, versus 24–48 hours in some cases.
- LLM cost: A general subscription was described as roughly US$20 per month. This was an approximate cost, not a guaranteed price.
- Prompt guidance: Prompts of 15 words or fewer were described as usually too short for complex recruiting tasks.
- Training: Raath says she typically runs workshops of about 1.5–2 hours.
- Usage: She reports using Claude roughly two or three times a day and says token limits can affect how much work she can do in a given period.
- Business mix: Raath describes her 2026 work as a roughly 50/50 split between talent sourcing and training, including recruiter AI training.
- Professional proof: She says she has about 600 LinkedIn recommendations. She attributes this in part to asking for recommendations and delivering value; this is an example of social proof, not a general benchmark.
Presenters and sources
- Vanessa Raath — Recruitment and AI trainer; shared the workflows, tools, and business recommendations summarized above.
- Mark Whitby — Host and interviewer.
Rate this summary
Your feedback will help improve summaries.
Improve this summary
Reprocess with a stronger model when the summary feels incomplete or inaccurate.
Translate summary in another language
Ask questions to this video
Chat for follow-up questions, clarifications, and source-backed answers.