What each number is
Every cost line and total is arithmetic on your inputs. The industry profiles (apart from the BLS wages in the cross-industry, insurance, retail and BPO profiles), the salaried load and the license uplift are internal planning values, labelled as heuristics; a field still at its profile grades Directional, and Planning-grade needs every graded field set off the profile. The wage escalator is a sourced market figure. The optimization figures are a conditional forecast at the realization stance you pick.
Formulas
Agents × hourly wage × (1 + benefits load) × 173 paid hours
Paid hours, because you pay for shrinkage. The benefits load opens at the shared 1.3.
Headcount × hourly rate × 1.25 × 173 hours, for supervisors, QA, WFM, trainers and IT
The salaried load is this tool's own planning value.
round(agents × annual attrition ÷ 12) hires × (recruiting cost + training days × 8 hours × loaded wage)
Carried in overhead.
Contacts × voice share × (AHT − after-call work) ÷ 60 × price per minute
Billed on line-open minutes; the agent is still paid for after-call work.
Seats × (CCaaS + WEM + CRM seat prices) + AI usage + analytics + iPaaS + recording + knowledge + security + telephony
Seats are agents, supervisors, QA and WFM.
Cloud infrastructure + amortized professional services + facilities + attrition
Cost per contact = monthly TCO ÷ contacts. Cost per resolution = cost per contact × (2 − FCR). Marginal per contact = AHT minutes × wage × the marginal load (1.18, never above the loaded rate entered) per minute + voice share × line-open minutes × telephony price
Cost per resolution uses the one-plus-repeat model. Unit costs stay on the loaded rate; deflection, repeat and handle-time savings are valued at the marginal load. Capturing them by not backfilling seats removes benefits too, about 10% more at the opening loads.
Year one = annual. Years two and three escalate labor and attrition at the wage rate and contracted software at the license rate; telephony and facilities stay flat. Plus any one-time implementation, once
A single blended rate is offered but misstates a labor-heavy base.
Containment: deflected contacts × marginal. FCR: avoided repeats on the handled pool × marginal. AHT: seconds saved × remaining handled contacts × loaded per minute. Attrition: fewer hires × cost per hire. Each × the stance
Applied in order so no contact is counted twice; rounded to the nearest $1,000. Stances: none 0%, conservative 50%, expected 70%, aggressive 100%.
Rules and bands
The range printed around annual and three-year cost by cost basis.
Each is disclosed and holds completeness Directional; AI usage is exempt from the software share check.
The read points to the Occupancy Risk Simulator before assuming a saving is free.
The optimization total and the stance never cap a confidence axis.
Every constant and where it comes from
Supervisors, QA, WFM, trainers and IT are salaried and carry a lighter employer burden than an hourly agent, so they load at less than the agent benefits multiple. A fourth load concept because it prices a different population; it is no second opinion about the agent load.
The 2,080 hour full-time year over twelve months. Labor cost is computed on paid hours, shrinkage included, because that time is paid even when no contact is handled.
Annual labor escalation in the three year view. Applied only to the labor bucket, because wages and contracted license inflate at different rates. Source: US Bureau of Labor Statistics, Employment Cost Index, wages and salaries, private industry, twelve month change.
Annual uplift on contracted recurring software at renewal. The middle of the 3 to 10 percent band enterprise renewal clauses commonly carry. Replace it with the uplift in your own contract.
Single blended escalator, offered only when the user opts out of the split rates. A blended rate misstates a labor heavy base, so the split is the default.
Plausibility guard on input coding. A cost per agent per month above this sits outside any real operation and almost always means an annual or total figure was entered as monthly. Holds completeness Directional.
A single software line above this share of software cost is the signature of a miscategorized or mis-scaled input. Holds completeness Directional. AI usage is exempt, because usage pricing legitimately dominates.
Span of control above this understates supervision cost and usually means supervisors were undercounted. Holds completeness Directional.
Occupancy below this is noted as idle capacity already on hand, so freed capacity is redeployment before it is cash. Framing only.
Cost per resolution above cost per contact by more than this reads as rework cost worth naming. Framing only.
Labor above this share reads as a people-cost operation in the analyst read. Framing only.
Labor above this share prompts the read to explain the separate wage and license escalators. Framing only.
Range printed around annual and three-year cost when costs are estimates. Display only.
Range printed when costs come from vendor quotes. Display only.
Range printed when costs come from invoices. Display only.
Tolerance on the channel mix totalling one. The voice share prices telephony, so a mix that does not total 100 percent misprices the usage bucket. Holds completeness Directional.
Opening agent wage for the financial services profile. An internal planning value, not a published median. The platform's sourced wage benchmark is market.wage.agent. A wage still at this value grades cost evidence Directional.
Opening agent wage for the healthcare profile. An internal planning value, not a published median. The platform's sourced wage benchmark is market.wage.agent. A wage still at this value grades cost evidence Directional.
Opening agent wage for the retail and eCommerce profile: the BLS May 2025 median for customer service representatives in Retail trade. A published median rather than the reader's own pay, so a wage still at this value grades cost evidence Directional. Source: US Bureau of Labor Statistics, Occupational Outlook Handbook, Customer Service Representatives, pay by industry: median hourly wage in Retail trade, May 2025 (TB research, 28 Sep 2026).
Opening agent wage for the telecommunications profile. An internal planning value, not a published median. The platform's sourced wage benchmark is market.wage.agent. A wage still at this value grades cost evidence Directional.
Opening agent wage for the insurance profile: the BLS May 2025 median for customer service representatives in Insurance carriers and related activities. A published median rather than the reader's own pay, so a wage still at this value grades cost evidence Directional. Source: US Bureau of Labor Statistics, Occupational Outlook Handbook, Customer Service Representatives, pay by industry: median hourly wage in Insurance carriers and related activities, May 2025 (TB research, 28 Sep 2026).
Opening agent wage for the BPO and outsourcer profile: the BLS May 2025 median for customer service representatives in Business support services (which includes telephone call centers). A published median rather than the reader's own pay, so a wage still at this value grades cost evidence Directional. Source: US Bureau of Labor Statistics, Occupational Outlook Handbook, Customer Service Representatives, pay by industry: median hourly wage in Business support services (which includes telephone call centers), May 2025 (TB research, 28 Sep 2026).
Wage plus benefits and employer payroll burden, and nothing else. The narrowest of the three loads. Use it wherever a wage becomes a loaded hourly rate for unit metrics.
The variable cost that disappears when one contact goes away. The only load a saving may be valued on, because fixed technology and facilities do not fall with volume.
Worked example
Computed by the tool's own engine at its opening case: the cross-industry profile, the expected stance and costs as estimates.
Operation: 200 agents at $21.53 an hour, 120,000 contacts a month, AHT 6:30 with 45 seconds after-call work, FCR 70%, containment 28%, attrition 40%
Staff: 20 supervisors, 5 QA, 4 WFM, 3 trainers, 4 IT
Technology: $150 CCaaS, $45 WEM and $75 CRM a seat; voice 55% at $0.025 a minute; AI, analytics and platform lines at the profile
$21.53 × 1.3 = $27.99 an hour; agent labor $968,419 a month
$1,203,267 with salaried staff
229 seats; telephony $9,488; technology $106,318
7 hires a month at $8202.15 = $57,415; overhead $82,415
$1,391,999 a month, $16,703,994 a year, $6,960 per agent a month
$11.60 per contact, $15.08 per resolution, $2.83 marginal (6.5 handle minutes × $21.53 × 1.18 ÷ 60, plus telephony)
$16,703,994 + $17,312,558 + $17,944,398 = $51,960,949
$74,000 a month gross, $52,000 at the expected 70%
What this tool cannot tell you
- The industry profiles are planning values that let the tool open on a runnable case. Replace every graded field with your own before relying on a total; the grade says how many remain.
- Cost per resolution uses the one-plus-repeat model (2 − FCR). An observed recontact rate is more precise.
- The optimization levers are sized against your targets. They show what a lever is worth, never how to reach it.
- Freed capacity is not cash until an action converts it; the stance is the haircut you choose, and none books $0.