AI timekeeping can help law firms make unbilled effort more visible, but the firm first needs to distinguish several different kinds of “unbilled time.”
A lawyer may work without the effort appearing on an invoice because the time was:
Never captured
Captured but not entered
Entered late
Classified as non-billable
Marked no charge
Included in a flat fee
Written down
Held in work in progress
Rejected by a client
Spent on internal firm activity
Those categories have very different meanings.
Unbilled Time Is Not One Metric
A useful framework separates:
Category | Meaning |
Missing time | Work occurred but no time record exists |
Draft time | Work is captured but not approved |
Non-billable time | Internal or otherwise non-chargeable work |
No-charge time | Client-related work intentionally not billed |
Flat-fee effort | Work delivered under a non-hourly fee |
WIP | Approved time not yet invoiced |
Write-down | Time reduced before invoice |
Write-off | Amount not collected or removed after billing |
Client rejection | Invoice amount rejected or reduced |
AI timekeeping mainly helps at the early stages: visibility, capture, classification, and review.
Where Time Disappears in the Workflow
The workflow can be viewed as:
Work performed → Activity captured → Time entry created → Entry approved → Time released → Prebill reviewed → Invoice sent → Payment collected
Loss can occur at each stage.
Examples:
Work performed but not captured
Captured activity ignored
Draft time left unreleased
Entry written down
Invoice delayed
Client rejects the charge
Payment is not collected
A timekeeping system cannot fix every stage, but it can make the beginning of the funnel more reliable.
Missing Time
Missing time is work that never becomes a usable time entry.
Common causes include:
Forgotten short tasks
End-of-day reconstruction
Calendar-only tracking
Timer mistakes
Work across several matters
Offline activity
Late entry
Administrative friction
ABA guidance emphasizes contemporaneous timekeeping because delayed reconstruction can reduce accuracy and completeness.
Captured but Unreleased Time
Automatic capture can create another category:
The system found activity, but the lawyer never reviewed it.
This can create a draft backlog.
Track:
Draft entries by user
Age of drafts
Unreviewed days
Entries with low matter confidence
Entries requiring duration correction
Entries excluded
Entries released
The goal is not maximum capture. It is timely review.
Non-Billable and Internal Time
Non-billable time may include:
Administration
Training
Business development
Firm meetings
Recruiting
Internal management
Technology
Professional development
This time is not automatically a problem.
It becomes useful management data when the firm can see:
Who performs it
Which activities consume the most capacity
Which work could be delegated
Which processes could be automated
Whether high-value lawyers are doing low-value internal work
No-Charge Time
No-charge time is client-related work that the firm intentionally does not bill.
Reasons may include:
Relationship decision
Efficiency adjustment
Learning curve
Scope issue
Client expectation
Billing judgment
Courtesy
Internal error
Tracking no-charge time separately from internal non-billable work gives a better picture of client effort.
Flat-Fee and Alternative-Fee Matters
Flat-fee work may never appear as hourly invoice lines, but it is still important operational data.
Time records can help the firm understand:
Actual effort
Staffing
Scope
Effective rate
Capacity
Pricing accuracy
Repeated tasks
Process improvement
Unbilled does not mean unimportant.
Write-Downs and Prebill Adjustments
A time entry can be captured correctly and still be reduced before invoicing.
Possible reasons include:
Poor narrative
Excessive time
Wrong staffing
Client guideline issue
Duplicate work
Matter budget
Billing judgment
Scope problem
Timekeeping analytics should therefore connect, where possible, to downstream billing data.
MIRA itself is upstream of final invoice review.
What AI Can Reveal
AI-assisted timekeeping can help reveal:
Frequent missing-time windows
Short tasks that are often omitted
Users with late review
High administrative-time categories
Matter-classification problems
Draft backlog
Repeated no-charge activities
Work outside scheduled calendar events
Work patterns across matters
Possible duplicate or overlapping entries
These patterns are useful even when the system does not own the final invoice.
What AI Cannot Infer Reliably on Its Own
AI should not assume that:
Unbilled time is lost revenue
Every non-billable hour should be billed
Every captured activity should become a time entry
Every write-down is a mistake
Every flat-fee hour should have been invoiced hourly
Every low-utilization lawyer is underperforming
Management context matters.
The firm should distinguish commercial decisions from process failures.
Dashboard Metrics
A useful dashboard can include:
Same-day entry rate
Missing-time days
Draft backlog
Draft age
Billable hours
Non-billable hours
No-charge hours
Flat-fee effort
Administrative time
Submission lag
Prebill write-downs
Billing realization
Collection realization
No single metric tells the full story.
Frequently Asked Questions
Is all unbilled time lost revenue?
No. Non-billable work, flat-fee effort, no-charge work, and internal activity may be intentionally unbilled.
Can AI find work that never became a time entry?
Activity-based systems can surface likely work signals, but the lawyer should review whether the activity should become an entry.
What is the most important unbilled-time metric?
There is no single best metric. Firms should separate missing time, draft backlog, non-billable time, no-charge time, flat-fee effort, write-downs, and WIP.
Can AI show why time is written down?
Only if the system has access to downstream prebill or billing-review data. Timekeeping data alone cannot explain every adjustment.
How does MIRA help?
MIRA improves visibility into captured billable and non-billable activity before reviewed time moves toward billing and finance systems.
