The hiring manager who fired her external recruiter

A hiring manager at a 40-person SaaS company told us a story we want to share (anonymized).

She'd been working with an external technical recruiter on a $25K retainer + 18% placement fee. The recruiter was good — well-known in the local market, decent network, professional process.

After 6 months of using Avery NXR's Marcus template internally, she ended the recruiter contract.

This isn't a recruiter-hate piece. The recruiter does a real job. The story is about which parts of that job are now done better by AI agents — and which parts still aren't.

What she was paying the recruiter for

The retainer covered:

→ Sourcing candidates (cold outreach to passive talent) → Screening inbound resumes → Initial phone screen conversations → Coordinating interview logistics → Reference checks → Negotiation support on offers → General market intelligence

For roles where she'd pay placement (senior engineers, etc.), the 18% fee on a $180K salary was $32,400 per hire.

Combined with 2-3 placement fees per year + the retainer, she was paying $80K-150K annually for the external recruiting function.

What Marcus replaced

Marcus is Avery NXR's resume screening agent. Configured for her use case:

→ Reads incoming resumes (inbound applications + recruiter forwards + cold inbounds via website) → Scores against role-specific criteria → Auto-sends interview invitations to strong matches → Flags borderline candidates for her review → Logs everything in the audit ledger

After 6 months of Marcus running, she had data:

→ Marcus processed ~2,400 resumes → Marcus invited ~180 candidates to phone screen → Of those, ~95 took the call → Of those, ~28 moved to onsite → Of those, 11 received offers → 8 accepted (3 didn't)

Result: 8 hires in 6 months. The recruiter, in the previous 6 months, had delivered 5 hires.

Marcus's hires were:

→ Cheaper (no placement fee) → Faster to source (auto-invitations within minutes of application) → Equally well-qualified (per her assessment in interviews)

The 8 hires from Marcus didn't have a recruiter intermediary.

What Marcus DIDN'T replace

Honest about what Marcus can't do:

Sourcing passive talent. Marcus reads INBOUND resumes. It can't reach into LinkedIn and start conversations with engineers who aren't job-hunting. For passive talent, you still need a human (internal recruiter or external).

The hiring manager's solution: hired an internal recruiter at $90K + benefits to do sourcing + relationship work. Cheaper than external retainer + placement fees. The internal recruiter focuses on the parts Marcus can't do.

Negotiation support. Marcus doesn't help close offers. The hiring manager handles negotiations herself, with the internal recruiter providing market context.

Network access. External recruiters have networks that take years to build. Marcus has access to whoever applies. For senior, specialized, or rare roles, network access still matters. For most engineering roles at her company, inbound was enough.

Market intelligence. External recruiters know what other companies are paying, what candidates are looking for, what's trending in the market. Marcus doesn't have this knowledge. Her internal recruiter + competitive research now covers it.

The new staffing model

After 6 months of Marcus + observation, her recruiting function looks like:

→ Marcus (Avery NXR agent): Inbound resume screening, interview invitations, audit logging → Internal recruiter (human, $90K + benefits): Passive sourcing, relationship building, market intelligence, candidate experience, negotiation support → Hiring managers (each team): Interview, decide, sell

Total annual cost: ~$112K (internal recruiter + benefits + Avery NXR Pro for the team).

Previous model annual cost: ~$80K-150K (retainer + placement fees, depending on hire volume).

For her hire volume, Marcus model is cheaper AND produced more hires.

What changed beyond cost

Speed. Inbound candidate to interview invitation went from 5-10 days (recruiter triage cycle) to <60 minutes (Marcus auto-invite). Top candidates were closer to offer stage before they accepted competing offers elsewhere.

Consistency. External recruiter's screening criteria evolved based on whoever they talked to last. Marcus applied the same criteria every time. Selection consistency went up.

Documentation. Every Marcus decision was logged in the audit ledger. When she wanted to understand why a candidate wasn't invited, the answer was queryable. With the external recruiter, the rationale was "their judgment."

Internal upskill. Her internal recruiter learned to use Marcus as a tool. The combination of human relationship work + AI screening became more effective than either alone.

What she'd tell other hiring managers

She gave us a few specific recommendations to pass on:

→ Don't fire your recruiter month 1. Run Marcus in parallel for 2-3 months. Compare outcomes. The data will tell you whether the model works for your specific hiring shape.

→ Internal recruiter is still valuable. The "AI replaces all recruiters" framing is wrong. AI replaces the SCREENING + INVITATION work. The relationship + sourcing + negotiation work is still human.

→ Configure Marcus carefully. First-pass scoring criteria will be off. Iterate weekly for the first 6 weeks. After that, the criteria stabilize.

→ Watch for bias. Audit ledger lets you see Marcus's decisions by candidate cohort. If patterns suggest bias, adjust the criteria. Don't assume Marcus is neutral — verify.

→ Communicate internally. Some candidates ask "did a human read my resume?" Honest answer is "no, an AI agent did first-pass screening based on the criteria I set, then I personally reviewed the ones it invited." That answer is acceptable to most candidates if the process is otherwise good.

The bigger pattern

Recruiting is one example of a function where AI agents handle a SUBSET of the work better than humans, while the rest of the work still requires humans.

The mistake is "AI replaces recruiters" or "AI can't replace recruiters." Both are wrong.

The right framing is "AI does the parts of recruiting that are pattern-matching + scaling well. Humans do the parts that are relationship + judgment + negotiation."

When you decompose a function this way, AI augments rather than replaces. The COST structure changes (less human time on the AI-handled parts). The capability stays.

This pattern shows up across functions:

→ Sales: AI handles lead qualification + initial research. Humans handle relationships. → Support: AI handles FAQ-shaped tickets. Humans handle nuanced cases. → Marketing: AI handles content drafting + competitor monitoring. Humans handle strategy. → Finance: AI handles invoice processing + reconciliation. Humans handle judgment calls.

Every function has an AI-handleable subset and a human-required subset. The teams figuring out the boundary correctly are operating at a structural advantage.

What this means for buyers

If you're hiring and spending more than $50K/year on external recruiting:

Configure Marcus on Avery NXR Free Desktop. Run it for 3 months on your inbound applications. Compare to your external recruiter outcomes.

If the math works the way it did for the hiring manager in this story (it often does), you'll have data to renegotiate your external recruiting contract or restructure your recruiting function.

If the math doesn't work for your specific case (rare roles, all passive talent, no inbound), you've learned that for $0 investment.

→ avery.software — Free Desktop tier. Marcus is pre-loaded.