Are your books ready for AI-Enabled bookkeeping?

Why this assessment exists

Every bookkeeping firm now pitches AI-powered workflows, but few finance leaders have a reliable way to tell whether their own books are ready for that shift before they buy the tooling. Automation applied to disorganized data does not fix the disorganization. It accelerates whatever problems already exist: unreconciled accounts spread faster, miscategorized transactions compound, and control gaps that used to surface slowly now surface at machine speed. This assessment gives $5M to $75M companies a structured way to answer the readiness question honestly, before committing budget to a platform change. It scores where the books actually stand today, not where leadership intends them to be, across the operating dimensions that determine whether automation helps or hurts.

How the scoring works

The assessment asks 30 questions across six operating dimensions, each scored from 0 to 3. A 0 means the control or process is not in place, an acknowledged gap with no plan yet. A 1 means it exists in partial or informal form. A 2 means it is mostly in place, with some gaps or exceptions. A 3 means it is fully in place and consistently maintained. Each dimension totals 15 points, for a maximum score of 90 across the full assessment. The scoring is self-reported, which makes it fast to complete, but the standard is deliberately literal: score where the organization stands today, not where a policy document says it should stand. The total maps directly to one of four readiness tiers that determine the recommended next move.

The six readiness dimensions

The 30 questions are organized into six dimensions that together determine whether automation will help or compound existing problems: data infrastructure, process standardization, control environment, technology readiness, talent and organizational readiness, and data quality. Data infrastructure covers how transactions enter the system, from live bank feeds to digital bill capture. Process standardization asks whether close, approval, and categorization workflows are written down and repeatable rather than held in one person's head. Control environment covers reconciliation timing, segregation of duties, and audit trail integrity. Technology readiness covers integration, access security, and backup procedures. Talent and organizational readiness asks whether a named owner and human oversight capacity exist. Data quality asks whether the inputs are clean enough for a rules-based system to work.

• Data Infrastructure
• Process Standardization
• Control Environment
• Technology Readiness
• Talent and Org Readiness
• Data Quality

The four readiness tiers

The total score sorts into four tiers, each with a different recommended posture toward automation. A score of 0 to 30 is Not Ready: foundational cleanup has to happen before any AI layer, because automating unreliable data compounds errors rather than reducing them. A score of 31 to 55 is Foundational: targeted AI tooling can start in narrow domains, but core workflows are not yet standardized enough to automate safely at scale. A score of 56 to 75 is Transition-Ready: the company is a strong candidate for AI-enabled bookkeeping, and automation will multiply discipline that already exists. A score of 76 to 90 is AI-Native: the foundation supports continuous close, autonomous classification with human review, and real-time anomaly detection. Each tier carries its own recommended first move in the full checklist.

• 0 to 30: Not Ready
• 31 to 55: Foundational
• 56 to 75: Transition-Ready
• 76 to 90: AI-Native

The operating principle

The assessment is built around one governing rule: AI proposes, human disposes. Artificial intelligence is well suited to ingestion, categorization, optical character recognition, reconciliation matching, and anomaly flagging, tasks that are high volume and low judgment. It is not well suited to the calls that require context: revenue recognition, accrual judgment, and month-end attestation. Every AI output should be reviewed by a human before it affects the general ledger. Companies that treat automation as a replacement for oversight, rather than a way to redirect it toward higher-judgment work, tend to end up with faster errors instead of fewer errors. This is the standard the assessment holds every dimension against.

What the full assessment adds

This page covers the framework: six dimensions, a 0 to 3 scoring method, and four readiness tiers. The full assessment adds the complete 30-item instrument, five items per dimension, each written as a scoreable statement rather than an open question, so there is no ambiguity about what a 3 looks like. It includes the per-dimension scoring table that totals each of the six dimensions out of 15 and rolls up to the full score out of 90, plus the tier-specific recommended first move for each of the four outcomes, from foundational cleanup to agentic workflow evaluation. This is the same standard applied across Catalyst's finance-function readiness reviews for companies in the $5M to $75M range.