Case Study
How a Solo Founder Hired a Marketing Lead in 8 Days
No recruiter, no HR team, no $500/month ATS subscription. Just a kanban board and an AI agent.
The Solo Founder Hiring Problem
Enterprise ATS platforms like Greenhouse and Lever cost $200–$500+ per month, take weeks to configure, and are built for HR teams with dedicated recruiters. A solo founder hiring one role doesn't need pipeline analytics dashboards. They need to not lose track of candidates.
The alternative is worse: candidate emails pile up in Gmail, outreach messages get copy-pasted across tabs, and there's no pipeline view. You end up ghosting good candidates because you forgot to reply.
AI recruiting tools promise to fix this, but they remove the human from decisions that matter — who to reach out to, what to say, and whether someone is actually a fit. What a solo founder needs is a visual pipeline where an AI agent handles the grunt work while the founder makes every call.
One Board, Five Lists
A kardbrd board with five lists becomes a full recruiting pipeline:
Each candidate is a card that accumulates everything: structured profile, Djinni links, CV attachments, outreach drafts, interview questions, debrief notes, and the full comment history between founder and agent. The board is the pipeline. Each card is a candidate's full history.
The agent does the grunt work.
You make every hire/reject call.
From JD to Shortlist in 8 Days
Write the JD
The founder asks the agent to write a job description. The agent reads everything on the board — existing candidate research, target audience notes, marketing copy, the product website — and distills it into a tight 420-word JD. A slop check catches 7 em dashes, 2 unicode arrows, and generic JD phrases. All fixed. Done in a day.
Post and Wire Up Auto-Import
The agent posts the job to Djinni and sets the recruiter email to the board's inbound address. From this point, every application that comes in automatically becomes a card in the Candidates list. Zero manual data entry.
Candidates Flow In
20+ candidates arrive as cards on the board. Each one is auto-formatted into a structured profile: standardized title, salary and experience table, cover letter summary, skills, strengths, open questions, and links. The agent enriches cards with full Djinni profile data, runs slop checks on cover letters to spot AI-generated applications, and stack-ranks everyone against the JD with star ratings. Day 1 stats: 2x views and 3.7x applications versus the category average.
Outreach and Scheduling
For strong candidates, the agent drafts personalized outreach in the founder's voice. Each message references the candidate's actual achievements — no templates, no corporate fluff. Quality gates enforce the rules: no salary talk (founder handles that), must include a scheduling link, slop score of zero. The founder reviews every message before it goes out.
Interview to Decision
The agent generates tailored interview questions based on each candidate's background, the gaps in their profile, and the JD requirements. Each question comes with a “why to ask” rationale and what to listen for. The founder interviews. The agent writes a structured debrief from the founder's notes. Card moves to Shortlist or Rejected. Result: 5 scheduled, 3 interviewed, 1 top candidate shortlisted.
The Agent Learns From Corrections
When the agent gets something wrong, the founder corrects it once. The agent encodes the fix into its skills permanently. The same mistake doesn't happen twice.
“Don't mention salary in outreach”
Now a permanent quality gate. Every outreach draft is checked automatically.
“Always include the scheduling link”
Enforced on every message. Missing link = draft rejected before the founder sees it.
“Rewrite card titles, don't just comment”
Formatting behavior updated across all future candidates. No reminder needed.
Correct once, fixed forever. The agent gets better at the job the longer you work with it.
The Entire Automation: 21 Lines
Two rules. That's the entire automation config. Everything else happens through natural language conversation in card comments.
board_id: YOUR_BOARD_ID
agent: HRBot
rules:
# Stop active agent session
- name: Stop agent
event: reaction_added
emoji: "🛑"
action: __stop__
# Auto-format email imports
- name: Format new candidate card
event: card_created
list: Candidates
model: sonnet
action: Run skill against this card /hr-format
New card in Candidates triggers auto-formatting. Everything else — outreach,
enrichment, interview prep, debriefs — happens through @HRBot mentions in card comments. No code. No integration setup.
How It Compares
| Kardbrd + HRBot | Manual (Gmail + Sheets) | Enterprise ATS | |
|---|---|---|---|
| Setup time | Minutes (2 YAML rules) | None (no structure either) | Weeks |
| Cost | Board plan + AI API costs | Free (but time-expensive) | $200–$500+/month |
| Auto-import applications | Email to card, automatic | Manual copy-paste | Yes |
| Personalized outreach | AI drafts with quality gates | Manual writing | Templates only |
| AI writing detection | Slop check on candidates and outreach | No | No |
| Human decision gates | Every action, every candidate | Everything is manual | Configurable |
| Agent learns from feedback | Corrections become permanent rules | N/A | No |
From “write me a JD” to a shortlisted candidate.
Eight days.
Build your hiring pipeline today
Create a board. Define your lists. Wire up auto-import. The agent handles the rest.
Get Started