Smartflow ROI Model — Assumptions and Methodology
Use this model to quantify the efficiency and risk impact of Smartflow for a specific prospect. Fill in the input variables from discovery conversations. Present the output as a directional estimate — not a guarantee.
How to Use This Model
- Complete the discovery questions in a scoping call to gather input variables.
- Fill in the inputs below using the prospect's actual data where possible; use industry benchmarks where the prospect cannot confirm.
- Calculate the outputs using the formulas in each section.
- Present the estimated impact as a range (conservative to optimistic), not a single number.
- Always caveat: "These are projections based on your inputs and comparable deployments. Actual outcomes depend on document complexity, team adoption, and integration scope."
Input Variables
Loan Onboarding Volume and Effort
| Variable | Description | Prospect Value | Benchmark |
|---|---|---|---|
monthly_volume | New loan agreements processed per month | [Enter] | 20–200 for mid-size APAC commercial banks |
avg_pages | Average pages per credit agreement | [Enter] | 200–500 pages |
current_hours_per_loan | Hours of manual extraction and re-entry per loan | [Enter] | 8–20 hours |
smartflow_hours_per_loan | Hours required with Smartflow (extraction + HITL review) | 0.5–2 | Typical: ~0.5 hrs extraction + ~1.5 hrs review |
working_hours_per_fte_year | Annual working hours per FTE (net of leave) | [Enter] | 1,800 hours |
fte_cost_fully_loaded | Annual fully-loaded cost per FTE (salary, benefits, overhead) | [Enter] | USD 80,000–200,000 depending on market |
Accuracy and Error Costs
| Variable | Description | Prospect Value | Benchmark |
|---|---|---|---|
current_error_rate | Estimated error rate in manual extraction (%) | [Enter] | 10–15% |
avg_error_cost | Average cost per data error (rework, correction, downstream impact) | [Enter] | USD 500–5,000 per error depending on severity |
smartflow_error_rate | Post-HITL error rate with Smartflow (%) | 5% | Conservative estimate; 95%+ accuracy achieved |
error_reduction_pct | Percentage of errors eliminated by Smartflow | 50–70% | Conservative range for HITL workflows |
Covenant Monitoring (Complete if prospect is a target for covenant monitoring)
| Variable | Description | Prospect Value | Benchmark |
|---|---|---|---|
covenant_monitoring_approach | Current approach: manual / spreadsheet / none | [Enter] | Most APAC commercial banks: spreadsheet |
avg_breach_detection_lag_days | Days after detection window before breach is identified | [Enter] | 30–45 days typical |
portfolio_size | Number of active credit facilities being monitored | [Enter] | — |
avg_breach_cost | Estimated cost of a covenant breach missed or actioned late | [Enter] | USD 50,000–500,000+ depending on exposure |
smartflow_prediction_lead_days | Days of advance warning Smartflow provides | 30–90 | Planned Q3 2026 |
Output Calculations
1. Annual Hours Saved
hours_saved_per_loan = current_hours_per_loan - smartflow_hours_per_loan
annual_hours_saved = monthly_volume × hours_saved_per_loan × 12
Example: 50 loans/month × (12 hrs − 2 hrs) × 12 = 6,000 hours/year
2. FTE Equivalent Freed
fte_equivalent_freed = annual_hours_saved ÷ working_hours_per_fte_year
Example: 6,000 hrs ÷ 1,800 hrs/FTE = 3.3 FTEs equivalent
3. Efficiency Value (FTE Cost Basis)
efficiency_value = fte_equivalent_freed × fte_cost_fully_loaded
Example: 3.3 FTEs × USD 150,000 = USD 495,000/year
Note: This does not assume headcount reduction. It reflects capacity freed for reallocation to higher-value work.
4. Error Cost Reduction
annual_errors_current = monthly_volume × (current_error_rate ÷ 100) × 12
errors_avoided = annual_errors_current × (error_reduction_pct ÷ 100)
error_cost_reduction = errors_avoided × avg_error_cost
Example: 50 × 12% × 12 = 72 errors/year × 60% reduction × USD 2,000 = USD 86,400/year
5. Covenant Monitoring Value (Q3 2026 capability)
breaches_at_risk_per_year = portfolio_size × estimated_breach_frequency
avoided_late_breach_cost = breaches_at_risk_per_year × avg_breach_cost × prediction_accuracy
Use conservatively. Prediction accuracy is directional — breach avoidance value depends heavily on borrower action after early warning.
Example Scenario: 50-Loan/Month APAC Commercial Bank
| Input | Value |
|---|---|
| Monthly loan volume | 50 loans/month |
| Average pages per agreement | 300 pages |
| Current hours per loan (extraction + entry) | 12 hours |
| FTEs dedicated to document processing | 4 FTEs |
| Fully-loaded FTE cost | USD 150,000/year |
| Current error rate | 12% |
| Average error cost | USD 1,500 |
| Output | Value |
|---|---|
| Annual hours saved | 6,000 hours |
| FTE equivalent freed | 3.3 FTEs |
| Efficiency value (capacity, not headcount reduction) | USD 495,000/year |
| Error cost reduction | USD 64,800/year |
| Total projected annual value | ~USD 560,000/year |
Presenting the ROI
- Always present as a range (conservative / base / optimistic).
- Always state the assumptions clearly — do not hide inputs.
- Do not promise specific numbers in a contract without Smartflow commercial team involvement.
- Use the ROI model to open a conversation, not to close a deal. The prospect's actual data matters more than benchmarks.
Recommended framing:
"Based on what you've told us — 50 loans per month, 12 hours each, 4 FTEs — and using conservative assumptions, we estimate Smartflow could free the equivalent of 3 FTEs of capacity per year and reduce error-related rework costs by over USD 60,000 annually. That's before any covenant monitoring value. Would you like to work through the numbers with your own figures?"
Related Documents
one-pagers/loan-onboarding.md— Onboarding-focused value narrativeone-pagers/covenant-monitoring.md— Covenant monitoring value narrative- Proposal template:
04-sales-collateral/_TEMPLATE-proposal.md(used for internal drafting)
Internal use only. Not for distribution to prospects in this raw format. ROI projections are directional estimates based on comparable deployments and stated assumptions.