Redesigning hiring's discovery layer for knowledge work
- Jan 27
- 8 min read
The hiring process for knowledge workers is fundamentally broken - not because there isn't enough talent, but because the systems we use to connect talent with opportunity can't see what actually matters. This exploration asks: what would it look like to make invisible capabilities visible, and to democratize the "witness function" that networking provides without requiring insider connections?
The problem
Hiring for knowledge work is in crisis. Time-to-hire has climbed to 44 days on average, up consistently over the past four years. Candidates face 250+ applicants per corporate job posting, with only 3-4% getting interviews. Meanwhile, employee referrals account for 30-50% of all hires despite being a tiny fraction of applications.
The real issue for candidates isn't volume - it's visibility.
The core mismatch
Hiring systems filter for exact credentials: specific titles, years of experience, company brand names, particular tools or methodologies. Success in modern knowledge work depends on something else entirely.
What actually predicts success doesn't fit neatly into résumé format:
Transferable skills that don't match keywords - You've facilitated complex stakeholder discussions but your title was "project manager." You've done deep technical research, just not "market research" with the exact tools the job posting lists. You've built relationships that unlocked key opportunities, but "relationship building" doesn't appear anywhere on your résumé.
Capabilities that don't have labels - Skills that exist but don't follow linear career paths or match expected progression.
How you work, not just what you've done - Your approach to ambiguous problems. How you synthesize information across contexts. The way you learn new domains.
How you show up - Collaboration style. Communication patterns. Whether you're someone who asks good questions, handles complexity well, brings clarity to messy situations.
Judgment under uncertainty - How you make decisions when there's no clear answer. Your ability to operate across different environments and constraints.
Actual interest versus credential-collecting - Whether you genuinely care about this specific work or you're just mass-applying to anything that matches your keywords.
Here's the problem: if these capabilities don't match credential templates, they're invisible. People who have exactly what's needed can't show it. Hiring managers can't assess it even when they're looking for it.
The current landscape
→ Traditional résumés and applications
What they do: List credentials in standardized format - titles, companies, education, years. The point is to be quickly scannable.
Why this fails:
Complex work gets flattened into bullet points
Your thinking process disappears
Non-linear paths look scattered rather than adaptive
There's no way to show how you work, only where you've worked
Transferable skills become keyword-matching exercises
The result: If your experience doesn't fit the template exactly, you might as well be invisible.
→ LinkedIn and portfolio platforms
What they do: Expand on the résumé - you can add recommendations, skill endorsements, work samples, show your network.
'Why this still fails:
Still organized around credentials and past titles
Endorsements are social gestures, not meaningful validation ("endorsed for leadership" from someone you worked with once)
Work samples show outputs but rarely reveal process or thinking
Platforms favor people with traditional paths and existing networks
The result: You can tell someone has done certain work, but not whether they have the underlying capabilities that matter.
→ Applicant tracking systems (ATS)
What they do: Automate screening by filtering résumés based on keywords, requirement matches, and algorithmic scoring.
Why this makes things worse:
Optimize for exact matches, which eliminates transferable skills by design
Candidates never get feedback on why they were filtered out
Both sides get forced into narrow templates
Every mismatch is treated as disqualifying rather than as potentially valuable diversity of experience
The result: Hiring managers are overwhelmed by volume but can't find good fits. Candidates get algorithmically screened out before any human sees their actual potential.
→ Networking and referrals
What they do: Personal connections vouch for candidates, providing context résumés can't capture.
Why networking works: Referrals translate invisible qualities into legible signals.
Someone can say:
"Incredibly creative problem-solver, finds angles nobody else sees"
"Thrives in ambiguity even though their role was structured"
"Great at building relationships across teams, gets things unstuck"
"Deep technical expertise even though title doesn't reflect it"
"Their work got killed by business priorities but the quality was excellent"
The data backs this up: Referred candidates are hired at 4x the rate of other applicants. They stay 70% longer in roles and perform better. Companies save $3,000+ per referral hire.
Why this doesn't solve the problem:
Profoundly inequitable - requires existing social capital and insider access
Reinforces homogeneity and systematically excludes outsiders
Most people don't have the networks needed to benefit from this
The result: Networking works brilliantly for those with access. For everyone else, it's a locked door.
What's missing: There's no system that makes invisible capabilities visible with evidence. Nothing that shows how someone thinks and works beyond credentials. No way to get the "witness function" without insider connections. No path for people with non-traditional backgrounds or unlabeled skills. No tools to help hiring managers assess what actually predicts success.
That gap is the opening.
Who gets hurt by this
The broken discovery layer creates problems for two groups with completely misaligned struggles:
Knowledge workers:
Have capabilities that don't match their titles (deep expertise, creative problem-solving, relationship building, etc.)
Followed paths that look "scattered" on paper but actually demonstrate adaptability
Possess transferable skills that don't match keyword templates
Can show how they think and work, but have no platform for it
Don't have extensive networks or insider connections
Hiring managers:
Need to assess thinking style and collaboration from résumés that only show credentials
Drowning in 250+ applications, default to risk-averse shortcuts (brand names, exact title matches)
Know their job descriptions don't capture what they actually need
Want diverse talent but lack tools to identify capabilities outside traditional templates
Spend significant interview time only to discover basic misalignment that should have been obvious earlier
Right now: Neither side can solve their problem. Candidates with relevant capabilities can't make them visible. Hiring managers can't assess them even when actively looking.
The solution: A comprehensive signal engine
The core idea: Build a system that replicates what networking provides - visibility into invisible qualities - but makes it available to everyone, not just people with insider connections.
How it works
1. Pattern recognition + signal bank
Build a rich profile of experience across all contexts, not just formal employment. The system recognizes patterns and transferable capabilities regardless of labels.
Instead of listing "Product Manager at Company X," you'd capture: deep technical expertise demonstrated through open-source contributions, creative problem-solving shown through community projects, facilitation skills developed in volunteer leadership, cross-domain synthesis applied in varied contexts.
The platform identifies these patterns even when they don't have formal titles attached.
2. Role-specific matching
For each opportunity, analyze what the role actually requires (not just stated credentials) and generate a tailored profile showing relevant strengths with supporting evidence.
Rather than "do you have 5 years as a [title]," ask role-appropriate questions like:
Can you navigate ambiguous technical problems with incomplete information?
Do you build relationships that unlock opportunities across organizational boundaries?
Have you brought creative approaches to established processes?
Can you synthesize insights across different domains or contexts?
Then surface evidence from the signal bank demonstrating these capabilities.
3. Evidence layer
Work samples and documented thinking that prove your approach and process, not just final outputs.
Show:
How you reframed a problem
Your research and synthesis process
Decision-making under constraints
How you facilitated alignment across stakeholders
This reveals thinking and judgment in ways credentials never could.
4. Validation layer
Specific, contextualized observations from collaborators - not generic endorsements.
Not "great teammate!" but rather:
"Identified the root issue three meetings before anyone else"
"Brought clarity to an ambiguous situation through structured facilitation"
"Adapted their approach when the initial strategy didn't land"
These perform the "witness function" that referrals provide, without requiring insider connections.
5. Progressive disclosure
Structure information so hiring managers can scan quickly (60-90 seconds) while having depth available when needed.
Surface level: Core capabilities, key patterns, strength of evidence
Deeper level: Specific examples, artifacts, validation, full context
This respects time constraints while enabling thorough assessment for serious candidates.
6. Two-way transparency
Companies share the reality of the role - actual challenges, team dynamics, what success looks like beyond the job description. Candidates understand what matters and can self-assess fit before applying.
Reduces misalignment and wasted effort on both sides.
How this addresses what's broken
For candidates whose capabilities are invisible:
→ Current problem: Relevant skills disappear when they don't match credential templates
The fix: Pattern recognition identifies transferable capabilities regardless of labels; evidence layer proves thinking and process
→ Current problem: Non-linear paths look "scattered"
The fix: System synthesizes patterns across contexts, showing adaptability and diverse experience as strengths
→ Current problem: No feedback on rejections, just algorithmic screening
The fix: Clear visibility into what matters; you can see gaps and address them rather than being silently filtered out
For hiring managers drowning in applications:
→ Current problem: Can't assess what predicts success from résumés
The fix: Progressive disclosure surfaces invisible qualities (thinking style, collaboration, adaptability) with evidence; validation layer provides credible signal
→ Current problem: Default to risk-averse proxies (brand names, exact matches)
The fix: Direct evidence of capabilities reduces need for credential proxies; discover talent you're currently missing
→ Current problem: Job descriptions don't capture actual needs
The fix: Two-way transparency shares role reality; role-specific matching surfaces relevant capabilities even when unlabeled
For the equity problem:
→ Current problem: Networking works but requires social capital
The fix: Democratizes the "witness function" through validation and evidence; makes discovery equitable and scalable
What makes this different
Most hiring tools are variations on the same question: "Do your credentials match our requirements?"
This asks something else: "Can you actually do the work, and how would we know?"
The real innovation isn't any single feature. It's rethinking what hiring's discovery layer should surface in the first place. Instead of optimizing credential-matching, we're making invisible capabilities visible with proof.
The central insight: Networking works because it translates invisible qualities into legible information. Someone can say "incredibly creative problem-solver" or "builds relationships that get things unstuck" or "deep technical expertise beyond their title" or "thrives in ambiguity." That translation only happens for people with networks.
What if we built those translation mechanisms into the discovery system itself? What if evidence, validation, and pattern recognition could perform the witness function at scale?
That's the opportunity - taking what works about referrals and making it accessible to everyone, not by replacing human judgment, but by surfacing the signals human judgment actually relies on.
Reflections
This started from watching people with exactly the right capabilities get screened out because their experience doesn't fit credential templates. The problem isn't lack of talent - it's that the discovery layer can't see it.
The "witness function" feels like the key insight. When we say "networking works," what we mean is: someone who knows you can vouch for invisible qualities that matter but don't show up on résumés. That's what's missing from formal hiring - not the personal connection itself, but the translation of invisible signal into legible information.
What would need to be figured out:
How do you actually capture "pattern recognition" in a way that's credible? What prevents it from just becoming keyword matching 2.0?
What motivates people to provide specific, contextualized observations rather than generic endorsements?
How do you prevent gaming - people optimizing their profiles the same way they optimize résumés?
If the goal is democratizing access, the solution can't require significant effort, social capital, or insider knowledge to participate. How do you keep barriers low while capturing nuanced signal?
That tension - between capturing what actually matters and making it accessible to everyone - is where the real design challenge lives.