Should Retained Search Fees Change Now That AI Does the Sourcing?

TL;DR

The classic retained search fee (a third at signing, a third at shortlist, a third at placement, totalling 25-33% of first-year salary) was priced around sourcing being the expensive, time-consuming part of the job. Recruiters were spending 13-15 hours a week per role just searching for candidates [2], and AI sourcing has compressed a large share of that into minutes [3]. That does not mean the fee should collapse to match the saved hours. Retained search was never really priced on hours anyway - the data shows retained work converts client intros to interviews at 77.9%, against 44.3% for contingent work [4], because of judgment and filtering, not search volume. This article gives agency owners a defensible position: keep charging for outcomes and risk reduction, be explicit about which third of the fee now buys speed-to-signal and verification rather than raw sourcing hours, and stop pricing the parts of the job AI has already commoditised as if they were still scarce.

The Fee Model Was Built for a World Where Sourcing Was the Bottleneck

The standard retained search structure has barely changed in decades: a third of the fee at engagement, a third at qualified shortlist delivery (typically 3-5 candidates), and a third at offer acceptance, totalling 25-33% of first-year salary at large firms and 15-20% at boutiques, often with a $100,000-$150,000 minimum floor at the top end [1]. That structure was not arbitrary. It mirrored where the actual labour sat:

• Sourcing ate the week. Recruiters spend an average of 13 hours per week per role on candidate searching, and 44% say searching takes up most of their time [2]. Staffing recruiters report even higher figures, around 14.6 hours a week on candidate search alone.
• The upfront third funded that search. Clients paid a non-refundable retainer partly because the firm was about to commit dozens of unpaid hours to market mapping before a single candidate was presented.
• The shortlist third rewarded discrimination, not discovery. Even in the old model, the money was less about "finding people" and more about narrowing a long list to the few worth a client's time - already a judgment task, not a search task.
• The placement third covered close risk. Getting an offer accepted, surviving a counter-offer and holding the relationship together for months of process was always the part no amount of sourcing effort could shortcut.

In other words, even the "sourcing-heavy" old model was already paying most of its fee for judgment and closing. Sourcing hours were expensive, so they got priced in visibly. AI sourcing has now made that visible cost mostly disappear, which is exactly why the fee conversation feels shaky - not because the value disappeared with it, but because the part everyone could point to and measure just did.

What AI Sourcing Actually Compresses (and What It Doesn't)

It is worth being precise here, because vendor marketing overstates this constantly. Retained searches for senior or hard-to-fill roles operate in universes of tens to hundreds of realistic candidates, not millions - the constraint was always discrimination among qualified people, not raw scale [4]. AI genuinely collapses some tasks. It does not touch others.

Task	Manual effort (pre-AI)	AI-assisted effort today	Still requires a human?
Market mapping / long-listing	10-20 hrs/role	Minutes to hours	Spot-check only
Signal detection (funding, exec moves, expansion)	2-3 hrs/day, mostly missed	Continuous, 24/7	No
Contact enrichment / decision-maker ID	20-30 min/lead	Near-instant	Verification of accuracy
Candidate profile / submission write-up	Full write-up per candidate	~70% faster drafting [5]	Yes - fit judgment
Shortlisting from long list	Hours of manual review	AI-assisted triage	Yes - final call
Candidate assessment / cultural and role fit	Calls, references, judgment	Minimal help	Yes - fully human
Negotiation, counter-offer handling, close	Fully human	No change	Yes - fully human

A widely repeated vendor claim is that AI cuts time-to-fill by 40-50%. The compression is real but the number is inflated: most timeline delays in a retained search sit on the client side (board scheduling, stakeholder alignment), which no sourcing tool touches [4]. AI shortens the part of the search you control. It does not shorten the part the client controls, and it does not do the part that actually differentiates a good search consultant from a mediocre one: knowing which three of forty plausible candidates are worth a client's calendar.

The Honest Question: Does Cheaper Sourcing Mean a Cheaper Fee?

Some agency owners are answering "yes" pre-emptively. A new entrant, ExactSearch.AI, has launched specifically to compete on this basis - positioning itself as a faster, AI-driven retained search offering at roughly half the fee of traditional firms [9]. That is a live test of the "sourcing was most of the cost, so most of the cost should fall" argument, and agency owners should watch it rather than dismiss it.

But the evidence on where retained search actually earns its premium does not support a proportional fee cut. The clearest data point: client-intro-to-interview conversion runs at 77.9% for retained work versus 44.3% for contingent work [4]. That 33-point gap is not explained by sourcing effort - contingent recruiters source hard too. It is explained by judgment filtering before the client ever sees a name. Lengthen the list because sourcing got cheap, and you dilute the exact mechanism that justifies the fee in the first place.

So the honest answer sits between the two extremes agency owners keep hearing:

• Not "charge the same, change nothing." If a third of your fee was explicitly sold as funding weeks of sourcing labour, and that labour now takes hours, keeping that framing invites exactly the client pushback ExactSearch.AI is betting on.
• Not "charge less because AI does it now." The 77.9%/44.3% gap shows the premium was never really about sourcing hours. Cutting the fee to match saved hours discounts a service you were barely charging for in the first place.
• The defensible middle: reprice against what actually changed. Rebalance which milestone the money attaches to, and be explicit with clients about what each third now buys. That is the rest of this article.

Where the Billable Value Actually Lives Now

If sourcing hours are no longer the scarce input, agencies need to be precise about what is. Five things hold up under scrutiny:

1. Signal timing, not search volume

Being first to a genuine hiring signal (a funding round, an executive departure, a restructuring) beats contacting more people. Signal-detection tools that surface hiring intent 48-72 hours before a role is posted publicly turn "we searched hard" into "we called before the client's own team knew they had a mandate." That is a timing advantage, not a labour advantage, and it is billable precisely because it is rare.

2. Verified data over raw output

86% of hiring managers say AI now makes it too easy to embellish a resume, and 80% say candidate resumes do not match real skills at least sometimes [7]. As AI-generated candidate material floods every channel, a consultant who has actually verified a candidate's claims, references and motivation becomes more valuable, not less. Cheap AI sourcing produces more unverified data. It does not produce more trust.

3. Calibrated judgment on fit

AI can surface forty plausible candidates for a role. It cannot sit in a confidential conversation with a sitting CFO, read the political context of why a client actually needs this hire, or know that the last two people this client rejected were too similar to the outgoing hire [4]. That contextual judgment is exactly what does not scale, which is exactly why it is still billable.

4. Negotiation and close

Around half of candidates in a live process receive a counter-offer, and roughly 57% accept it - but of those who accept, a large majority regret it or leave within one to two years [8]. Managing a candidate through that moment, keeping the client calm during a slipping timeline, and closing without losing the placement is a skill AI has not touched at all. It was always the least automatable third of the process, and it still is.

5. Risk transfer via the guarantee

A failed senior hire can cost well over 200% of salary once lost productivity, team disruption and re-search costs are counted, with some failed executive placements documented above $500,000 in total cost [6]. The replacement guarantee attached to a retained search is effectively insurance against that outcome. Insurance is priced on the size of the risk it removes, not on how many hours the search firm spent sourcing. That framing survives AI sourcing completely intact.

A Defensible Fee Framework for the AI-Sourcing Era

Rather than "cut the fee" or "keep the fee and hope nobody asks," agencies need a framework that reweights the milestones to match where the labour and the risk actually sit today.

Milestone	What it historically funded	What it should be framed as funding now
Retainer (traditionally 1/3)	Weeks of manual market mapping and long-listing	Exclusivity, ICP calibration, signal monitoring set-up, and priority access to the consultant's time
Shortlist delivery (traditionally 1/3)	Hours spent narrowing candidates down	Verified, referenced, judgment-filtered shortlist - the 77.9% conversion mechanism itself
Placement (traditionally 1/3)	Reward for a successful outcome	Negotiation, close, and the replacement guarantee - unchanged, arguably underpriced

In practice this means three concrete moves for agency owners deciding how to reprice:

• Shrink the retainer's implicit "sourcing hours" story, not necessarily its size. Keep the retainer if it still buys exclusivity and calibration, but stop describing it to clients as funding weeks of search labour they can now see AI doing in an afternoon.
• Grow the shortlist milestone's justification, even if the percentage stays flat. This is where the 77.9%/44.3% conversion gap lives. Make the verification and filtering work visible - reference checks completed, signals acted on, candidates screened out and why.
• Tie the placement milestone explicitly to guarantee terms. A 90-day-to-12-month replacement guarantee is worth more when a bad senior hire costs 200%+ of salary [6]. Price and market it as risk transfer, not as a finder's fee.

Practical Steps to Re-Price (or Re-Justify) Your Retained Fee

Whether an agency ends up holding its percentage flat or adjusting it, the process should look the same:

1. Audit where your consultants' hours actually go today. Time-log a live search for two weeks. Most agencies are surprised how few hours now go to raw sourcing versus verification, client management and candidate care.
2. Separate "AI did this" from "a person judged this" in your own reporting. If your shortlist report cannot say which candidates were AI-surfaced versus human-vetted, you cannot defend the fee that sits on top of the vetting.
3. Rewrite your engagement letter's value language. Replace "dedicated research resource" with concrete deliverables: signal monitoring, verified shortlist, reference-checked finalists, guarantee terms.
4. Benchmark against low-cost AI-native entrants openly with clients. Do not pretend ExactSearch.AI-style competitors do not exist. Name the trade-off directly: faster and cheaper versus a verified, judgment-filtered, guaranteed outcome.
5. Consider a modest retainer reduction paired with a placement/guarantee premium. This mirrors where the value has actually moved, without collapsing total fee income if your close rate and retention numbers back it up.
6. Track the KPIs below for two full quarters before changing anything permanently. Repricing off a hunch is as risky as refusing to reprice at all.

KPIs to Track If You're Repricing

Metric	Why it matters here	Target
Hours per search on sourcing vs verification	Shows whether your fee story matches reality	Verification > sourcing hours
Client-intro-to-interview conversion	The judgment metric that justifies the premium over contingent	>60%
Time from signal to first outreach	Where AI-driven speed actually creates client-visible value	<72 hours
Placement retention past 12 months	Proof the guarantee/close-quality third is earning its fee	>80%
Client-reported reason for choosing you over an AI-native competitor	Direct evidence of what clients are actually paying for	Tracked qualitatively every mandate

How boilr Fits Into This Shift

This shift is exactly what boilr is built around: let AI do the sourcing and signal-detection labour that used to justify part of the fee on hours alone, and free the consultant to spend that reclaimed time on the parts that still justify a premium - verification, judgment and closing.

• Signals monitors funding rounds, executive moves, expansions and job-posting velocity across thousands of sources, often surfacing hiring intent before a role is posted publicly - the timing advantage described above, delivered continuously rather than manually chased.
• Companies matches and enriches target accounts against your ICP automatically, replacing the market-mapping hours the old retainer used to fund.
• Candidates sources and shortlists candidates against role requirements, cutting the long-listing time that used to be billed as pure sourcing labour.
• Tasks turns researched signals and candidates into ready-to-review outreach, so the consultant's time goes to verification and judgment rather than manual research assembly.
• Company Brain retains what worked (winning ICPs, objection handling, signal patterns) as institutional memory that survives consultant turnover - the calibration knowledge clients are increasingly paying the retainer for.
• Integrations with Bullhorn, RecruiterFlow and Spott mean the automation sits inside the workflow consultants already use, rather than becoming another disconnected tool.

What boilr deliberately does not automate: candidate reference verification, fit assessment on a confidential call, negotiation through a counter-offer, or the final decision on which three names actually go in front of a client. Every signal carries a traceable source, and every outreach task is reviewed and sent by the consultant, not the software. The fee argument only survives if agencies can honestly say a human is still doing the judgment work AI cannot - and that only holds if the tooling is built to keep that boundary visible, not blur it.

Want to see how much of your team's sourcing week is reclaimable? Book a walkthrough and find out what your consultants could bill for instead.

Fee-Model Mistakes Agencies Are Making Right Now

Mistake #1: Discounting the whole fee because sourcing got cheap

Why it fails: Sourcing hours were never the majority of what retained search actually charged for, so cutting the fee to match saved hours discounts work you were not billing separately for either.

Fix: Isolate which part of your pricing was ever explicitly tied to sourcing hours, and only adjust that part.

Mistake #2: Keeping the old pitch language unchanged

Why it fails: Clients who can see AI sourcing tools themselves will not accept "dedicated research team" as a justification for a retainer they know now takes hours, not weeks.

Fix: Rewrite the engagement letter around verification, signal timing and guarantee terms instead of research hours.

Mistake #3: Ignoring AI-native low-cost entrants

Why it fails: Firms like ExactSearch.AI are actively pitching clients on the "sourcing got cheap, so should the fee" argument [9]. Pretending they do not exist does not make the client conversation go away.

Fix: Address the comparison directly with evidence - conversion rates, retention data, reference depth - rather than avoiding the topic.

Mistake #4: Letting AI shortlist without a visible human filter

Why it fails: If a client cannot tell the difference between your shortlist and an AI-generated long list, you have no basis for the fee that sits on the shortlist milestone.

Fix: Document verification steps (references checked, calls held, signals acted on) as part of every shortlist deliverable.

Mistake #5: Treating the guarantee as boilerplate

Why it fails: A vague 90-day guarantee clause undersells the actual risk transfer a client is buying, especially now that a failed senior hire is documented at 200%+ of salary in lost cost [6].

Fix: Price and present the guarantee as a named, quantified benefit, not small print.

A 90-Day Plan to Re-Position Your Fee Story

• Days 1-30: Time-log two live searches to see the real hours-per-task split. Draft new engagement-letter language around verification, signal timing and guarantee terms.
• Days 31-60: Pilot the reweighted milestone framing (smaller "sourcing" story on the retainer, bigger emphasis on shortlist verification and guarantee) with two or three live mandates.
• Days 61-90: Compare client-intro-to-interview conversion and placement retention against prior mandates. Roll the new framing out agency-wide if the numbers hold; adjust if they don't.

Frequently Asked Questions

Should retained search agencies lower their fee because AI does the sourcing now?

Not automatically. The evidence shows retained search's premium over contingent work comes from judgment filtering (77.9% versus 44.3% client-intro-to-interview conversion) rather than sourcing hours [4]. A blanket fee cut discounts work that was never really being charged for by the hour. The defensible move is to reweight what each part of the fee is framed as buying, not to cut it across the board.

What percentage of a retained search fee historically went to sourcing?

There is no single published breakdown, because retained fees have always been milestone-based (a third at signing, a third at shortlist, a third at placement) rather than itemised by task [1]. What is known is that sourcing consumed the most visible hours (13-15 hours a week per role [2]), which is why it felt like the main cost driver even though shortlisting and closing were arguably where more of the actual value sat.

Is AI going to replace retained search consultants?

The current evidence points to augmentation, not replacement. AI performs well at market mapping, call transcription and assembling deliverables, but assessing executive fit requires confidential conversations and organisational context no model has access to [4]. The roles most exposed are the sourcing-only, low-judgment parts of the job, not the full consultant function.

How do I explain a fee to a client who says "AI can find candidates for free"?

Separate finding from verifying. AI can surface names; it cannot verify references, assess fit in a confidential conversation, or navigate a counter-offer. With 86% of hiring managers now saying AI makes it too easy to embellish a resume [7], verified, human-checked candidate data is becoming more valuable, not less, even as raw sourcing gets cheaper.

What is a fair fee structure for a retained search in 2026?

Most firms still use the milestone model (a third at signing, a third at shortlist, a third at placement, totalling 25-33% of first-year salary at large firms or 15-20% at boutiques) [1]. What is changing is not necessarily the percentage but the justification behind each milestone: framing the retainer around exclusivity and calibration rather than sourcing hours, and the shortlist and placement milestones around verification, judgment and guaranteed outcomes.

Should the guarantee period change now that AI speeds up sourcing?

Faster sourcing does not reduce the risk a guarantee protects against - a bad senior hire still costs upwards of 200% of salary once disruption and re-search costs are counted [6]. If anything, agencies should price and market the guarantee more explicitly as the industry's fee conversation shifts away from sourcing hours.

Are low-cost AI-native search firms a real threat to traditional retained search fees?

They are a real and growing signal to watch. Entrants like ExactSearch.AI are already pitching AI-driven retained search at roughly half the traditional fee [9]. Whether they succeed will depend on whether they can match the verification depth and conversion quality that currently justifies the premium, not just the speed of sourcing.

How can boilr help my agency defend or reprice its retained search fee?

boilr automates the sourcing and signal-detection work that used to be billed on hours (via Signals, Companies and Candidates), freeing consultant time for the verification, judgment and closing work that still justifies a premium fee. The Company Brain also retains the calibration knowledge (winning ICPs, objection handling) that clients are increasingly paying the retainer for, even as consultants change.

Sources

Information sourced from public industry reports and research publications as of September 2026.

1. Pin - Retained vs Contingent Search: How to Choose in 2026
2. Pin - Recruitment Statistics 2026: 50 Data Points Recruiters Need
3. Pin - The 9 Best AI Sourcing Tools for Recruiters in 2026
4. Recruiterflow - Will AI Replace Executive Search? Hype vs Reality
5. Recruiterflow - Retained Search Process: The Model & How AI Is Changing It
6. Talentfoot - The Real Cost of a Senior Leadership Mis-Hire: A 2026 Benchmark
7. Express Employment Professionals-Harris Poll (via Yahoo Finance) - 86% of US Hiring Managers Say AI Makes It Too Easy to Exaggerate Skills
8. Eclipse Software - 7 Counter Offer Statistics Every Recruiter Needs To Know
9. ExactSearch.AI Launches to Disrupt Executive Search with Transparent Pricing and AI-Driven Precision