There's a number that haunts every professional services firm: the gap between hours worked and hours billed.
Partners know it exists. Associates feel it daily. Finance teams see it in utilisation reports that never quite add up. The industry calls it "leakage" — as if billable time were a liquid slowly draining through cracks in the floor.
But leakage implies a plumbing problem. The reality is more fundamental than that.
The Anatomy of Lost Time
Time doesn't disappear in large, obvious chunks. It vanishes in the spaces between tasks — the two-minute email that turns into fifteen minutes of research, the hallway conversation that surfaces a critical insight for a matter, the context-switching tax of moving between seven different applications.
The problem isn't laziness or forgetfulness. It's that the work most worth billing is often the hardest to categorise in the moment. Deep thinking doesn't fit neatly into six-minute increments. A flash of insight during a shower doesn't have a matter code.
Why Timer-Based Tracking Fails
Traditional time tracking tools all share the same assumption: that the professional will remember to press a button at the exact moment work begins and ends.
Timer-Based Tracking
- Requires conscious interruption of work
- Fails during context switches (the most common state)
- Punishes deep work — flow states don't pause for buttons
- Creates guilt loops: forgot to start → reconstruct later → inaccurate
- Assumes work has clean start/stop boundaries
AI-Observed Tracking
- Captures work passively from application context
- Handles overlapping activities naturally
- Rewards deep work — longer uninterrupted sessions are captured fully
- Eliminates reconstruction: time entries draft themselves
- Understands that work is continuous, not discrete
The fundamental mismatch is between how tracking tools model time (discrete, sequential blocks) and how knowledge work actually happens (overlapping, recursive, non-linear).
The Reconstruction Problem
When professionals reconstruct their day from memory, they're not just inaccurate — they're systematically biased in specific directions.
Short tasks get rounded up. Long research sessions get compressed. Work done outside business hours gets forgotten entirely. The result isn't random noise — it's a consistent understatement of the complex, high-value work that firms most want to bill for.
Legal billing in six-minute increments was designed for a world of dictation and physical files. In a digital workplace where context-switching happens every 3 minutes on average, it creates more fiction than precision.
The professional who spent 45 minutes deeply researching a novel legal question, interrupted by three emails and a Teams message, will often log it as "30 mins — research" because the interruptions felt like they consumed time that wasn't really theirs to bill.
What AI Changes
AI-powered time capture doesn't ask professionals to change their behaviour. It observes the digital exhaust they already produce — application switches, document opens, email sends, calendar events — and drafts time entries from the pattern.
The shift is from recall to recognition. Instead of asking "what did I do for the last hour?", the professional reviews a draft entry and says "yes, that's right" or "no, that was personal."
The Trust Equation
The objection is always the same: "I don't want software watching everything I do."
It's a valid concern. And the answer isn't to dismiss it — it's to build trust through transparency and control.
The time capture runs locally. The data stays on the professional's machine. The firm never sees raw activity data — only the time entries that the professional explicitly approves and submits. The AI is a drafting assistant, not a surveillance tool.
Curious how it works?
Good Minutes captures time without changing how you work. Local-first, private by design, and built for professionals who value their hours.
See the productThe professionals who lose the most billable time aren't the ones who need more discipline. They're the ones doing the most complex, high-value work — the kind that doesn't fit into neat boxes. They deserve tools that understand that.