The Recruiter BD Time Audit: Where the Hours Go and Which to Hand to AI (2026)

TL;DR

Nobody can tell you how many BD hours your desk loses, because the honest answer lives in your own diary. The published research agrees on the direction, not on your number: in Salesforce's 2026 State of Sales survey of 4,050 sellers, the average seller spends just 40% of their time selling [1], and Salesforce's own summary puts non-selling work at 60% [2]. This guide gives you a one-week audit a desk or agency owner can run without any software: log BD time into seven buckets, split each bucket into gathering, recording and judging, and hand only the gathering and recording work to an AI sales employee. The consultant keeps verifying, sending, calling and meeting. Use the research percentages as a planning haircut (sellers expect agents to cut prospect research time by 34% and email drafting by 36% [1]), not as a promise, and let your own log set the target. boilr is built for exactly this split: it researches, enriches, scores and drafts, and you verify and send.

Why Audit BD Time Before You Buy Anything

Most agencies decide on AI tooling by feel. Someone sees a demo, a licence gets bought, and six weeks later nobody can say whether a single hour moved. An audit fixes the order of operations: measure first, automate second, re-measure third. It also gives you the one number that matters in a board or partner conversation, which is hours per week available for live client conversations.

• Selling time is a minority of the week: Salesforce's 2026 report finds the average seller spends 40% of their time selling [1], and its sales statistics page states that reps spend 60% of their time on non-selling tasks [2]. These are cross-industry sales figures, not recruitment-specific ones, which is exactly why you should measure your own desk.
• This is not a new problem: an earlier study of more than 720 reps by XANT (formerly InsideSales) found only 35.2% of time went on selling, with roughly 14% on administrative tasks and 17.9% spent in the CRM [3]. The study dates from around 2018 and covers inside and field sales broadly, so treat it as history, not a benchmark for your desk.
• Admin is heavy on the delivery side too: a 2025 Totaljobs survey of 748 HR leaders reported 17.7 hours of manual admin per vacancy, including 3.6 hours reviewing applications, 2.5 hours scheduling interviews and 3 hours on post-interview notes [4]. That is hiring-side admin, not BD, but it shows how much recruiter time disappears into low-judgement work.
• Your own diary beats any average: a desk that lives on retained searches has a completely different time profile from a high-volume contract desk. The audit below turns the average into your number.

What this guide deliberately does not claim

You will find "recruiters waste X hours a week on BD" headlines everywhere. Most trace back to a vendor blog with no method attached. One widely repeated line says individual recruiters spend over 100 workdays a year seeking sales opportunities; it comes from an investor article that cites no underlying study [5]. We do not build the case on it. We give you a template instead, and the numbers you get from it will be yours.

The Seven-Bucket Framework

Seven buckets are enough to cover BD for a typical 360 or niche desk. Fewer and you hide the problem; more and consultants stop logging by Wednesday. A similar audit by Recruiterflow uses five broader categories and suggests that admin landing near 30% of the week is typical [6]; that is a vendor rule of thumb, so use it as a sanity check only.

Bucket	What goes in it	Typical recruiter example
1. Account research	Reading company news, funding, leadership changes, job posts, finding the hiring manager	Checking whether a Series B fintech has a new CTO before you ring
2. List building	Assembling target-company or contact lists, ICP filtering, exporting and cleaning	Building 60 logistics firms in the Midlands with 50-500 staff
3. CRM admin	Logging notes, updating stages, creating contacts, deduplicating, tagging	Re-keying a call outcome into Bullhorn or RecruiterFlow
4. Drafting outreach	Writing and personalising emails, LinkedIn messages and call scripts	Writing a signal-led email to a COO who just opened a second site
5. Follow-up	Chasing non-replies, scheduling next touches, sending CVs or market updates after a call	A day-four nudge to a client who opened your last email twice
6. Calls	Live conversations with prospects and clients about roles and mandates	A 12-minute intro call with a head of talent
7. Meetings	Scheduled client meetings, intake calls, internal BD reviews	A 45-minute briefing with a hiring manager on a new search

Resist an eighth "other" bucket for the first week. If a consultant cannot place an activity, make them write it down in plain words next to the entry; the strays usually reveal a seventh-bucket habit you did not know existed.

Run the Audit in One Week

The audit works because it is short and forces a decision on Friday. Pick a normal week: no bank holiday, no quarter-end rush, no consultant on leave. Include every consultant who does BD, including the owner.

Before Monday: set the rules (30 minutes)

• One shared sheet: columns for date, start time, duration in 15-minute blocks, bucket (1-7), target account or contact, and a one-line outcome.
• Log as you go: a quick note at the end of each block beats a Friday reconstruction, which always inflates calls and shrinks admin.
• BD only: candidate-side delivery (screening, interviews, offers) is out of scope. If a task serves both, log it where it started.
• No judgement: tell the team the audit measures the process, not the person. Honest logs need that promise.

Monday to Thursday: log and tag

• Add one tag per entry: G for gathering (finding information), R for recording (writing it into a system), J for judging (deciding, persuading, relating). Tag at the moment of logging, not later.
• Note the trigger: was the work prompted by a signal (funding, exec move, job post) or by a cold list? This feeds the signal-led BD comparison later.
• Flag rework: mark any entry where you looked up something you had already looked up. Repeat research is the cleanest AI target and the clearest symptom of missing shared memory.

Friday: total and classify (60 minutes, whole desk)

1. Sum hours per bucket for each consultant and for the desk.
2. Convert to share: bucket hours divided by total BD hours logged.
3. Count G, R and J hours inside every bucket.
4. List the five biggest G and R entries by total hours across the desk.
5. Mark each as hand off, assist or keep using the table below.
6. Write the target: hours per week you want back and what they will be spent on, in numbers.

Sort Every Hour: Hand Off, Assist or Keep

The principle is simple. AI takes gathering and recording; humans own judging. The consultant verifies and sends every message, because trust with a client is your moat and a bad send costs more than a slow one.

Bucket	Gathering / recording share	Verdict	Why
1. Account research	Mostly G	Hand off, consultant reviews	Public information, repeatable, easy to check against a source link
2. List building	G and R	Hand off, consultant sets the ICP	The ICP is a judgement; matching companies to it is not
3. CRM admin	Mostly R	Hand off where an integration exists	Value depends on the sync actually working with your CRM
4. Drafting outreach	G plus first-draft writing	Assist: AI drafts, consultant edits and sends	Tone and timing are relationship calls
5. Follow-up	Scheduling G, content J	Assist: AI queues and drafts, consultant sends	Knowing when to push and when to leave a client alone is judgement
6. Calls	Prep G, conversation J	Keep, hand off only the prep	This is where mandates are won
7. Meetings	Prep G, meeting J	Keep, hand off only the briefing pack	Face time is the product

Notice what the table does: it never moves a call or a meeting. The goal is more conversations per consultant, not fewer humans in them. Our 20-minute weekly BD pipeline review is the right place to look at those numbers every Friday once the audit is done.

Worked Example: An Illustrative Consultant Week

The figures below are invented for illustration, not benchmarks. They show how to turn a log into a hand-off estimate. Assume a consultant logs 20 BD hours in the week. To avoid overstating the gain, the planning column uses the only published absorption figures we can cite, Salesforce's seller expectations of a 34% cut in prospect research time and a 36% cut in email drafting [1], applied to research, list building and drafting. We assume zero for CRM admin until your integration is proven in a pilot.

Bucket	Logged hours (illustrative)	Share of BD time	Planning absorption	Hours back
1. Account research	4.5	22.5%	34%	1.5
2. List building	3.0	15%	34%	1.0
3. CRM admin	2.5	12.5%	0% until piloted	0
4. Drafting outreach	3.0	15%	36%	1.1
5. Follow-up	2.0	10%	0% until piloted	0
6. Calls	3.0	15%	Keep	0
7. Meetings	2.0	10%	Keep	0
Total	20.0	100%		about 3.6

In this made-up week, 13 of 20 hours (65%) sit in buckets 1 to 4, which are the hand-off candidates, but the cautious plan recovers about 3.6 hours. That is deliberate. A planning figure you beat is a good story; one you miss is a failed project. The pilot then tells you whether the real absorption rate is higher.

Turn recovered hours into billings (use your own numbers)

• Hours back per consultant per week: from your audit.
• Conversations per hour: from your call log, bucket 6.
• Conversation to meeting rate: from your CRM history.
• Meeting to mandate rate: from the last two quarters.
• Average fee per placement and fill rate: from your finance data.

Multiply down the chain. If any link is unknown, that is your next audit, not a reason to guess. Our guide to client tiering by lifetime gross profit shows how to decide which accounts deserve the freed hours first.

What an AI Sales Employee Can Absorb, and What It Cannot

boilr is an AI sales employee for recruitment, one per consultant. It runs the top of the BD motion and hands you finished work to check. Here is how each audit bucket maps to a real module on boilr.ai [7].

Audit bucket	boilr module	What it does	Consultant still does
Account research	Companies and Signals	Finds and enriches target clients, reads hiring and buying signals such as funding, moves and hiring activity	Decides which accounts are worth a call
List building	Companies and the ICP	Matches companies to your ICP and scores them	Defines and reviews the ICP
CRM admin	Integrations	Pushes verified tasks into the CRM you already use; Bullhorn, RecruiterFlow and Spott are named on boilr.ai	Confirms the data landed correctly
Drafting outreach	Tasks	Drafts personalised outreach, including multi-step email and LinkedIn sequences	Verifies, tweaks a line and sends
Follow-up	Tasks and Agent	Queues the next touch as a ready-to-send task	Chooses whether to send it
Candidates for open roles	Candidates	Sources candidates against a live role	Screens and presents
Shared memory	Company Brain	Keeps the agency's ICP and winning patterns in one place that survives consultant churn	Feeds it with what actually works
Calls and meetings	None, by design	Nothing replaces the conversation	All of it

The audit is also how you size a pilot honestly. Salesforce found that top-performing sellers are 1.7 times more likely than underperformers to use prospecting AI agents for outreach [1], and 54% of sellers say they have already used agents [1]. Adoption is moving, but correlation is not a guarantee, so verify the gain on your own desk. For the wider picture on how adoption links to growth, see the AI adoption revenue gap, and for the scaling argument without extra hires, read how agencies use automation to scale BD without hiring more staff. This guide is the measurement step both of those assume you have done.

Where it should not run unattended

• Sending: every message goes out under a consultant's name; a human reads it first.
• Pricing and terms: fee percentages and exclusivity are negotiated by people.
• Sensitive relationships: a client you placed with last month gets a personal note, not a sequence.
• Candidate-confidential detail: keep anything covered by a candidate's confidence out of drafts until you have checked it.
• New-market judgement: an unfamiliar sector needs a consultant's read before the ICP is trusted.

Manual Week vs Audited and Reallocated Week

Dimension	Before the audit	After the audit and a pilot
Where BD hours go	Guessed; remembered as "mostly calls"	Logged by bucket and tagged G, R, J
Research	Repeated by each consultant on the same account	Done once, stored in the Company Brain, reviewed by the consultant
Targeting	Cold lists, gut feel	ICP-scored companies with signals
Outreach	Written from a blank page	Drafted, then verified and sent by a human
Proof of gain	Anecdotes after a demo	Same log, re-run in week 6, compared by bucket
Headcount plan	Hire a BD resource when it hurts	Add capacity only after recovered hours are fully used

KPIs to Track After the Audit

Metric	How to measure	What good looks like
Live-conversation hours	Buckets 6 and 7 as a share of BD time	Rising against your own baseline
Research and list-building hours	Buckets 1 and 2 per consultant per week	Falling while account quality holds
Repeat-lookup flags	Count of entries tagged as rework	Trending to near zero
Edit rate on drafts	Share of drafts where the consultant rewrites more than a line	Falling as the Company Brain learns
Signal-to-conversation time	Days from signal detected to first conversation	Shorter than your pre-pilot average
Mandates per consultant	New exclusive or retained mandates per quarter	Up, with headcount flat

Common Audit Mistakes

• Auditing in a freak week: a month-end or holiday week distorts every bucket. Run it again if so.
• Letting the owner skip it: the owner's BD pattern often differs most, and consultants copy it.
• Logging activity, not time: "sent 30 emails" tells you nothing about hours.
• Treating research as free: one hour on a single account feels harmless; ten of them is a billing line.
• Assuming 100% absorption: use a cautious haircut, then raise it with evidence.
• Skipping the re-measure: without a second log in week 6, you will never know if the pilot worked.
• Automating before ICP clarity: if the desk cannot say who it targets, AI will research the wrong companies faster. Refresh the ICP first; our quarterly ICP refresh covers how.

A Time-Boxed Plan: Audit to Pilot in Six Weeks

Week	Action	Output
Week 1	Run the seven-bucket audit across the desk	Hours and G/R/J split per bucket
Week 2	Classify hand off, assist or keep; set the hours-back target; confirm the ICP	One-page audit summary
Week 3	Connect the CRM, calendar and inbox; load the ICP into the Company Brain; start with two consultants	Pilot live, tasks arriving
Weeks 4-5	Consultants verify and send; log edit rate and repeat lookups	Edit and rework data
Week 6	Repeat the audit for the pilot consultants and compare by bucket	Real absorption rate, go or no-go on rollout

Pair this with the earlier piece on the account-research tax if you want a deeper look at what bucket 1 costs per desk, and with our guide to AI deep-research tools for recruiter account research when you compare options.

Want to run the audit with a pilot behind it? Book a demo or start with boilr and let your AI sales employee take the research while you keep the calls.

Frequently Asked Questions

How long should a recruiter time audit for business development take?

One working week of logging plus an hour on Friday to total it. Two weeks gives a steadier picture but most desks stop logging properly after five days, so one clean week beats two patchy ones. Re-run it in week six of a pilot to compare.

Do I need a time-tracking tool?

No. A shared spreadsheet with 15-minute blocks and a bucket column is enough. A tool adds friction, and the point is the classification, not the stopwatch. If your consultants already use a timer, use it, but keep the seven buckets.

What share of BD time can AI realistically take over?

There is no recruitment-specific benchmark we can responsibly cite. Salesforce reports that sellers expect agents to cut prospect research time by 34% and email drafting by 36% once fully implemented [1]; those are expectations, not measured results. Use them as a cautious planning figure and replace them with your own pilot data.

Will a consultant still need to approve every message?

Yes, in the boilr model. The AI sales employee drafts and the consultant verifies and sends, so the message goes out in a human's name. That keeps client trust intact and lets the consultant tweak a line before it lands.

Why not just hire a BD resource or researcher instead?

You can, and some desks should. But a hire adds a fixed cost before you know which hours are the problem. The audit tells you whether you need more capacity or just fewer hours lost to gathering and recording work. It also stops the knowledge from leaving with the person, which a shared Company Brain addresses.

How do I handle consultants who resist logging?

Be clear that the audit measures the process, not performance, and share the Friday totals back to them. Start with the owner's own log. When people see that the output is fewer research hours rather than a headcount review, resistance usually drops.

Can this work for a very small desk or a solo recruiter?

Yes, and the maths often favours it: a solo consultant has no one to share research with. Log a week, find the largest gathering bucket and test an AI sales employee on that alone before widening the scope.

Does the audit apply to candidate-side work as well?

The template does, with different buckets, but this guide scopes BD only. Candidate-side admin is large too, as the Totaljobs figure of 17.7 admin hours per vacancy suggests [4], though that survey measured HR leaders and is not a BD benchmark.

Sources

Information sourced from public industry reports as of October 2026.

1. Salesforce - State of Sales report announcement 2026 (4,050 sales professionals, 22 countries, surveyed August-September 2025)
2. Salesforce - 40 Sales Statistics (60% of rep time on non-selling tasks)
3. XANT (InsideSales) - Why Sales Reps Spend Less Than 36% of Time Selling (study of 720+ reps, published January 2018)
4. HR Review - Recruiters lose two days per hire to admin (Totaljobs survey of 748 HR leaders, August 2025)
5. Antler - Why we invested in Vente AI (the 100 workdays claim, no underlying study cited)
6. Recruiterflow - The Recruiter's Time Audit (five-category audit method, vendor guidance)
7. boilr - AI sales employee for recruitment (product modules and integrations)