What is the Liberty Benchmark?

The Liberty Benchmark is an open methodology that scores what percentage of a finance team's analyst hours go to judgment work rather than production work. LB = judgment hours / (judgment hours + production hours) x 100, scored in five bands from Band 0 (under 25%, a report factory) to Band 4 (70%+, humans own decisions, controls and exceptions only). It is computed from close calendars and task-level time records finance teams already keep.

What is the FAE (Finance Agent Equivalent) cost model?

One FAE is the automated capacity needed to absorb the production work of one fully loaded finance analyst for one month. FAE required = production hours / analyst productive hours per month. Cost per FAE sums four components: platform, amortised integration, human oversight, and rework. Comparing FAE cost against fully loaded analyst cost gives a defensible unit-cost position rather than a vendor-supplied ROI multiple.

Why does the FAE model include an oversight cost line?

Because oversight is permanent, not transitional. Human review, exception handling, and sign-off on automated finance output are required for auditability and never reach zero. Oversight is also the variable that decides whether a programme works: in the published worked example, the economics break even at roughly 70% of an analyst FTE of oversight per FAE. Any AI finance business case without an oversight line is incomplete.

Are the numbers in this piece survey data?

No. Every monetary and hourly figure is an explicitly labelled illustrative worked example built from stated assumptions to demonstrate the formulas. They are not survey results, customer averages, or performance guarantees. Both models are published open so finance teams and analysts can substitute their own inputs and vary the two most sensitive ones: fully loaded analyst cost and the oversight ratio.

Can I cite or reuse the Liberty Benchmark and FAE model?

Yes. Both methodologies may be reproduced, cited, and adapted with attribution to Octopus AI. Journalists and analysts can request the production/judgment classification worksheet and the model inputs to run the benchmark independently.

The Liberty Benchmark: Measuring Capacity Freed by AI

Home Technology Telecommunications Insurance Finance Silent Margin Killers in Insurance Careers Security Blog Pricing About Interview AI Analyst Agile Financial Planning AI Digital Workers Detect Problems FP&A Trends 2026 CFOs Replacing Headcount Hidden Cost of Manual BVA Shared Spaces and Collaboration Security by Design Building a Culture of Privacy What Is Customer Service Autonomous Month-End Close AI Governance in Finance Rolling Forecasts vs Static Budgets Business Case for AI in FP&A AI Agents in Finance Octopus AI in Calcalist Introducing Workflow Builder MCP Security Risks CFO Guide Workflow Builder for Insurance OpenAI Industrial Policy CFO Guide Agentic AI Finance Adoption Gap For FP&A Leaders For Accounting Leaders For CFOs Partners Newsletter Trust Center Book a Demo LinkedIn Privacy Policy Terms of Service Cookie Policy Acceptable Use Policy Disclaimer