An AI workforce for brokerages
A 100-agent brokerage over one year. Every number is cited in chapter 07 or computed from an assumption you can edit in chapter 05, and the page recomputes when you change one. The totals count company dollar, the share of each commission the brokerage keeps after the agent’s split. Your agents keep the rest, which is most of it.
Modeled Annual Value Created, in company dollar
$765K
A typical side grosses about $10,700 in commission[2][3]. The default assumes $2,500 of that stays with the brokerage and the agent keeps the rest.
25,000
hours returned across the roster per year (modeled)
+100
added sides modeled across the roster
16×
modeled value vs workforce cost
6.2×
the modeled floor: return with zero added closings
Added sides and recruiting drive the headline number, and both are assumptions. The fourth tile strips them out. With closings, recruiting, deferred hires, and leakage all set to zero, the modeled returned time would still cover the workforce cost 6.2×.
Every row ships today. Each outbound send waits for a person to approve it, and each action lands in the audit log.
| Workflow | Runs | What it does | Feeds |
|---|---|---|---|
| Morning lead follow-up · Lucia | Daily | Sweeps each inbox and contact database for leads that need a touch, then drafts the check-ins | Revenue |
| Daily briefing · Ren | Daily | Each seat opens to a ranked queue of pipeline, tasks, and approvals | Operating |
| Auto-prospect · Ivy | Weekly | Claims the likeliest sellers in each territory, proposes next steps | Revenue |
| Cold leads → warm · Scout | Weekly | Re-engages contacts across every office’s book once they sit quiet for 90 days | Revenue |
| Listing health check · Alex | Weekly | Scores every active company listing against the market | Risk |
| Buyer watch · Ivy | Weekly | Re-runs each buyer search, surfaces new listing matches | Revenue |
What leadership controls
The recruiting platform (separate surface)
A roster onboards in cohorts. The monthly series assumes 20% of seats live in month 1 and all of them by month 6. Closings follow the cohorts with a lag: a seat that comes live in month 5 books most of its added closings in year two, so year one captures less than the annual figures in chapter 01. Figure 03 puts a number on the gap. The months below play out the model’s assumptions, and a real rollout can run slower.
Months 1–3
Cohorts & coverage
Offices onboard in cohorts while inbox scans rebuild each database. Daily follow-up and briefings run on every connected seat from week one. By month 3, half the roster works under one playbook.
Months 4–8
Conversion & recruiting
In this scenario the first added closings land around month 5. The inside-sales and coordinator openings stay unfilled while coverage holds. Recruiters work a ranked list of verified candidates, and the first recruits sign.
Months 9–12
Compounding & governance
Listings get checked weekly, past clients hear from their agent each quarter, and each close triggers a referral ask. Budget year two on these fully onboarded months, since the early months ran light.
What turns on when
Months after activation →
Figure 01 · Rollout timeline
Each row is a workflow from chapter 02, placed at the month a typical cohort rollout turns it on. The dot marks the start.
Modeled value created per month
Figure 02 · Monthly value by bucket
The cohort ramp (20% → 100% of seats by month 6) scales every bucket, closings included: a closing in month 6 comes from the seats that were live in month 3. Revenue starts in month 4 because a worked lead sits through a median 10-week search [8] and then contract-to-close.
Cumulative value vs cumulative workforce cost
Figure 03 · Cumulative modeled value vs workforce cost
Value = Figure 02, accumulated. Cost = $48K/yr, spread monthly; edit it to your quote in chapter 05. Year one captures 77% of the modeled annual total ($765K). The gap is the cohorts that came live late. Year two starts with the full roster on.
| Month | Revenue | Operating | Cost | Risk | Month total | Cumulative value | Cumulative cost |
|---|---|---|---|---|---|---|---|
| 1 | $0 | $1,488 | $0 | $0 | $1,488 | $1,488 | $4,000 |
| 2 | $0 | $4,772 | $0 | $0 | $4,772 | $6,260 | $8,000 |
| 3 | $0 | $9,297 | $5,292 | $2,083 | $17K | $23K | $12K |
| 4 | $2,320 | $15K | $6,879 | $2,708 | $26K | $49K | $16K |
| 5 | $7,105 | $20K | $8,467 | $3,333 | $39K | $88K | $20K |
| 6 | $15K | $25K | $11K | $4,167 | $54K | $142K | $24K |
| 7 | $21K | $25K | $11K | $4,167 | $60K | $202K | $28K |
| 8 | $28K | $25K | $11K | $4,167 | $67K | $270K | $32K |
| 9 | $38K | $25K | $11K | $4,167 | $77K | $347K | $36K |
| 10 | $38K | $25K | $11K | $4,167 | $77K | $424K | $40K |
| 11 | $41K | $25K | $11K | $4,167 | $80K | $504K | $44K |
| 12 | $46K | $25K | $11K | $4,167 | $86K | $590K | $48K |
The defaults rest on published figures, cited by number in chapter 07. Company dollar is the piece of each commission the brokerage keeps after the agent’s split, and it is yours to set. The $2,500 default sits against roughly $10,700 of gross commission on a typical side[2][3], with the agent keeping the difference.
9
median sides per agent, 2025 [1]
~16%
of agents switched brokerages in 2025 [16]
$5–10K
cost to recruit one producing agent (2016) [17]
$127K
salary for one inside sales agent [13] plus one transaction coordinator [14], the deferred-hire default
The gaps a brokerage standard aims at
Why one added side per agent is the default
One side on the nine-side median is an 11% lift. Controlled studies of this kind of work measured about 14% on average[11], and closing a deal is harder than finishing a task, so the default stays under the study number.
The mechanisms match the team model: follow-up[5][6], retention[8], and prospecting. A hand-picked team adopts a playbook. A 100-agent roster adopts unevenly, which is why this default sits at a third of the team model’s.
The same studies measured about 34% for the newest workers[11]. At a brokerage, the middle of the roster has the most room.
The build-vs-buy line
Brokerages buy this coverage with headcount and point software: inside sales agents at about $69,400 a year[13], transaction coordination at $300 to $500 a file or a $58K salary[14], plus per-seat licenses. The default defers two hires, one inside sales agent and one transaction coordinator, against a workforce cost of $48K/yr. If those openings were never real, zero the line in chapter 05.
Every assumption sits here. Recruiting enters the total as 2 recruits at 8 first-year sides each, under the median 9[1], valued at your company dollar. Count only recruits the platform wins you beyond your current pace; recruiting you would have done anyway adds nothing.
Editable assumptions
Modeled added company dollar
$250Kfrom the existing roster / yr
Today: 900 sides → $2.25M company dollar
Recruiting production
$40Kincluded in Revenue Growth
Counts recruits won beyond your current pace, in year one only. Chapter 06 shows it as its own line
Return on the workforce
16×modeled return multiple
Modeled zero-closings floor: 6.2× from returned time alone
Bars scale to the total. The three roster rows split one computed figure by assumption; recruiting and the rest are computed.
Composition of the total
$765K / yr
* The 45 / 35 / 20 split of the roster’s added company dollar is an assumption. Leakage recovered counts deals already in the pipeline that would have fallen through; database reactivation counts new business from quiet contacts. Keep the two apart when you enter your own numbers.
Roster revenue
Agents × added sides × company $ / side
Recruiting production
Recruits × first-year sides × company $ / side
Time returned
Agents × hours/week × 50 weeks × loaded $/hour × capture %
Deferred support hires
Deferred hires × salary. One inside sales agent [13] + one transaction coordinator [14]; loaded cost runs higher
Leakage recovered
At-risk sides recovered × company $ / side
Return multiple
Total modeled value ÷ annual workforce cost
The model earns trust when your numbers replace ours. Here is how that happens, and where the model stops.
Before a pilot
The audit
We interview leadership, operations, recruiting, marketing, and top producers, then map the systems the brokerage runs on. The baseline for hours, response times, and production is what every later claim gets checked against.
During a pilot
Measured monthly
Each month we pull response times, sequence coverage, health flags acted on, and the share of drafts approved. We also track adoption by cohort, because chapter 03 holds only if the ramp is real.
At renewal
Realized vs modeled
Leadership reads the realized numbers against this page line by line: added sides, hours, deferred hires, recruits. Where the model ran hot, we lower it.
About these numbers: every dollar figure on this page is a hypothetical estimate computed from the assumptions shown, including the editable defaults. The figures are illustrations. They are not a promise, a projection, or a guarantee of any earnings or outcome, and they do not describe typical results. Actual results depend on your market, your adoption, your effort, and conditions this model ignores, and some customers may see smaller gains or none. Cited third-party studies measured other settings and carry no guarantee here. Nothing on this page is financial, legal, or tax advice.
We run the audit and rebuild this page with your numbers.