Every solution you build should have a clear link to measurable impact. This framework treats impact quantification as a core part of the product lifecycle for automation, not an afterthought.
1. Define a clear "Before State"
Before you build anything, document what the work looks like today. For every automation initiative, capture four things: the manual effort required (tasks, process steps, who does them), the time consumed per execution, how often the task occurs, and the business value at risk if it goes wrong.
Don't waste time converting hours to dollars at this stage. Capture the raw effort first.
Worked example: Invoice data extraction
A finance analyst spends roughly 4 hours per month manually pulling data from invoices and uploading it to an internal dashboard. Three teams do the same thing.
| Step | Calculation | Result |
|---|---|---|
| Total hours per month | 4 hrs/team × 3 teams | 12 hrs/month |
| Total hours per quarter | 12 hrs/month × 3 months | 36 hrs/quarter |
| Total hours per year | 36 hrs/quarter × 4 quarters | 144 hrs/year |
| Convert to FTE months | 144 hrs ÷ 160 hrs (standard FTE month) | 0.9 FTE months/year |
That single task, across three teams, burns almost a full person-month every year.
2. Pick a tangible metric
Each automation maps to at least one measurable outcome. Five categories cover the territory most teams will encounter.
Time Saved — FTE hours per period, task completion time (before vs after), manual handoffs removed. Example: 80 hrs/month saved from automated reporting across all teams.
Cost Saved — SaaS licence reductions, contractor spend reduced. Example: $25K/year saved by replacing outsourced data tagging software.
Revenue Enabled — Conversion rate uplift, faster lead routing, time-to-market improvements. Example: +$100K revenue from leads prospected by AI systems.
Accuracy / Risk — Error rate reductions, compliance coverage, audit trails added. Example: 98% drop in data entry errors via AI form pre-fill.
Velocity / Throughput — More tasks completed in less time, latency reductions, SLA improvements. Example: 3× increase in customer tickets triaged daily.
3. Calculate ROI
One formula, three inputs.
ROI = (Vₐ − Cₘ) / C₀
Vₐ = Annualised value created (time saved, cost saved, or additional revenue)
Cₘ = Annual cost to maintain the system
C₀ = Upfront cost (time spent building + infrastructure costs)
For more granular reporting, expand each variable to account for loaded labour rates (typically 1.4× salary to cover super, benefits, etc.) and separate builder costs from user costs.
ROI = (Hₛ × Rₗ) − (Hₘ × Rᵦ + Cₛ) / (Hᵦ × Rᵦ + C_setup)
Hₛ = Hours saved per year
Rₗ = Loaded hourly rate of the employee saving time
Hₘ = Hours spent on maintenance per year
Rᵦ = Loaded hourly rate of the builder/maintainer
Cₛ = Annual cost of subscriptions, APIs, and infrastructure
Hᵦ = Hours spent to build the initial solution
C_setup = One-time setup fees for software or infrastructure
Assumes standard work hours of 2,080 per person per year.
The five-step process:
Document assumptions and key inputs
Salary rates, hours, frequency, infra costs.
Calculate upfront costs (C₀)
Builder hours × loaded rate, plus any one-time setup fees.
Calculate annual maintenance costs (Cₘ)
Ongoing maintenance hours, subscriptions, API spend.
Calculate value created (Vₐ)
Hours saved × loaded rate, cost reductions, or revenue enabled.
Run the final ROI calculation
Plug the numbers into the formula. If ROI > 1, the initiative pays for itself in year one.
4. Lightweight reporting cadence
Don't create a reporting overhead that eats into the time and money you're trying to save. A single snapshot per initiative is enough to keep stakeholders aligned.
Example reporting snapshot
| Field | Detail |
|---|---|
| Project | Auto-Triage Support Tickets with LLM |
| Stakeholder | Head of CX |
| Before | Manual triage by 2 FTE, ~6 hours/day |
| After | 85% of tickets auto-tagged with >90% accuracy |
| Time saved | 90 hrs/month (~0.6 FTE) |
| Cost saved | $36,000/year (fully loaded salary) |
| SLA improvement | Response time improved by 40% |
| Status | In Production |
| Year 1 ROI | 6.4× |
This framework is designed to be copied and adapted. Start with the simple ROI formula. Graduate to the granular version when you need to justify spend to finance or leadership.