Article Summary
Use the Automation ROI Report to prove how much AI drafting is saving your team in time and cost, and to see exactly where drafts still need heavy editing.
Estimated Time
5 minutes for ongoing reviews
Prerequisites
Admin permissions
Approved, answer-required responses in the reporting period
Familiarity with your organization's cost and time assumptions (used in Configure ROI)
Step-by-Step Instructions
Step 1: Access the Automation ROI Report
Navigate to Insights from the main menu.
Click Automation.
The Automation ROI Report dashboard opens, scoped to All Users & Teams by default.
Step 2: Filter and Configure
Select your preferred Time Period and Users & Teams dropdowns, plus Filters, in the top right to scope the report.
Click Configure ROI to open your organization's time and cost assumptions.
Adjust the inputs to match your team's actual labor model, then save. All ROI tiles recalculate automatically.
Step 3: Review the Headline Data Points
The top of the report shows four headline metrics for the selected time period:
Name | What it is |
Automation Rate | Percentage of AI-drafted, approved responses that needed 0-1 words edited. |
Time Saved | Estimated hours (and working days) saved vs. writing responses from scratch. This show how much time has been freed for higher value work. |
Cost Savings | Time saved × hourly cost × opportunity factor. |
Responses Analysed | Count of approved responses in scope (plus total words reviewed). The sample size behind the ROI numbers. |
Step 4: Understand the Change Categories
Every AI drafted response is scored into one of five change categories based on how much of the draft an editor had to change:
Category | Threshold | What it means |
Perfect | 0–1 words changed | Fully automated, highest automation credit. |
Small Tweaks | 1–50 words changed | Light touch-up, still strong automation. |
Minor Changes | 50–100 words changed | Meaningful edits, partial automation. |
Major Changes | 100+ words changed | Heavy rewrite, draft was a starting point only. |
Not Drafted | No AI draft was generated | Excluded from the automation-rate denominator. |
Step 5: Review the Detail Sections
Scroll down to Details for a breakdown of automation performance:
Section | What it shows | Why it matters |
Overall Automation Rate | Stacked bar of category share across all responses. | One glance at how well AI is drafting responses. |
Project Automation Details | Top 15 projects by responses, with change breakdown, estimated time spent, and time saved. | Which deals got the most AI leverage. |
Automation Accuracy Trend | Monthly stacked automation categories. | Is draft quality improving over time? |
Editor Workload Analysis | Top 15 editors by responses, with per-editor automation %. | Who benefits most from AI, and who still rewrites heavily? |
Configure ROI | Configure the statistics that determine you ROI statistics, ie. Fact Retrieval Time, Editing Speed, ect. | Tune savings math to your situation. The configured rates apply to all ROI tiles. |
Step 6: Export Your Report
Click the Download button in the top right corner.
Select your desired export format.
Download the report for sharing with stakeholders.
💡 Tips & Best Practices
Data Accuracy:
Always approve responses to ensure they're included in calculations
Mark projects as completed when finished to capture accurate data
Update ROI settings when your cost structure or processes change
Review and adjust settings quarterly for accuracy
Configuration:
Use realistic time estimates based on actual team performance
Include fully loaded hourly costs (salary, benefits, overhead)
Reporting:
Export reports before quarterly business reviews
Filter by time period to show progress over specific intervals
✋🏼 Common Mistakes to Avoid
Leaving Configure ROI at default values - produces Cost Savings and Time Saved figures that don't reflect your real labor costs.
Comparing Automation Accuracy Trend across periods with different Configure ROI settings - the categories stay consistent, but the savings math won't be comparable.
Ignoring the Not Drafted category - a rising share usually signals a change in project types or missing source content, not a drop in automation quality.
Need Help?
💬 Live Chat: Available in-app
📚 Learning Centre: learn.autorfp.ai/en