Agency time tracking is a fiction. Not intentionally, but practically. The gap between the time your team actually spends on work and what gets logged is significant, and it is costing you money.
Studies consistently show that professionals fail to log 20-30% of their working time. In agencies, where context-switching is constant and days are fragmented, that number climbs higher. If your team logs 6 hours in an 8-hour day, the missing 2 hours are not breaks. They are client emails, quick Slack responses, internal meetings, and small tasks that feel too minor to log.
Those unlogged hours add up to real revenue leakage. AI is the first practical solution.
The problem with active time tracking
Active time tracking (clicking start, clicking stop, filling in timesheets) fails for three reasons.
1. It relies on discipline. People forget. They get pulled into a meeting, switch tasks, check emails, and suddenly two hours have passed unlogged. Nagging about timesheets does not fix this; it just creates resentment.
2. It is inaccurate. Even diligent time trackers round up or down, misattribute time between projects, and underestimate admin work. The data you collect is better than nothing, but it is not the truth.
3. It creates friction. Every timer start and stop is a context switch. For creative and strategic work, that interruption has a cost. Your team resents it because it genuinely gets in the way of the work you are paying them to do.
The result: most agencies have time tracking data that is directionally useful but precisely wrong. Decisions made on that data (project pricing, team capacity, client profitability) inherit those errors. This is one reason agency profit margins are often worse than owners believe.
Passive tracking: the AI alternative
Passive time tracking works in the background. It monitors which applications, documents, and websites your team uses throughout the day and automatically categorises that activity into projects and task types.
No timers. No timesheets. No discipline required.
How it works
The tool runs quietly on each team member’s device. It detects:
- Which applications are in focus (Figma, Google Docs, Slack, your PM tool)
- Which files and documents are being worked on
- Which websites are being visited
- Calendar events and meetings
AI then categorises this activity: “45 minutes in Figma working on the Acme Corp brand refresh” or “20 minutes in Google Docs on the quarterly content plan for Client X.” The categorisation improves over time as the AI learns your team’s patterns.
The tools
Timely is the standout for agencies. It was built specifically for passive tracking with AI categorisation. It creates draft timesheets from detected activity, which team members can review and adjust in minutes rather than building from scratch. Pricing starts at around £8 per user per month.
Toggl Track has added AI features to its traditional timer-based approach. Its AI suggests time entries based on calendar events and integrations. Not fully passive, but a meaningful step up from manual logging.
Clockify offers AI-assisted categorisation on its paid tiers. It analyses your time entries and suggests project assignments based on patterns. Less sophisticated than Timely but more affordable.
Memory.ai (formerly Timely’s standalone product) focuses specifically on passive tracking with strong privacy controls. Good for teams where monitoring concerns are a barrier.
Privacy considerations
Passive tracking raises legitimate concerns. Your team will want to know what is being monitored and what is not.
Best practice:
- Track application and document usage, not screen content or keystrokes.
- Give team members full access to their own data and the ability to delete or reclassify entries.
- Be transparent about what the tool captures. Hold a team meeting before deployment.
- Exclude personal time. Most tools can be configured to ignore activity outside working hours and on specific applications (personal email, banking).
Handled well, passive tracking actually reduces the burden on your team. Instead of spending 15 minutes a day filling in timesheets, they spend 3 minutes reviewing AI-generated entries. Most people prefer that trade.
AI categorisation of time entries
Whether you use passive tracking or stick with manual entry, AI categorisation improves your data quality.
Project attribution. AI matches time entries to projects based on document names, application context, and client associations. “Working in the Brand Guidelines doc for Acme Corp” automatically maps to the correct project and task category.
Task type classification. AI categorises time into types: strategy, design, development, admin, meetings, communication. This gives you a breakdown of how time is actually distributed across activities, not just projects.
Anomaly detection. AI flags unusual patterns. A designer spending 40% of their time in email is a problem. A developer logging no time on a project that is due next week is a warning. These flags surface issues before they become crises.
Utilisation analysis and recommendations
Raw time data is only useful if you analyse it. AI turns tracking data into actionable insights.
Utilisation rate calculation
The core metric: what percentage of your team’s available time is spent on billable work?
Benchmarks:
- Below 55%: Significant capacity waste. Investigate where time is going.
- 55-65%: Typical for agencies without systematic tracking.
- 65-75%: Good. Healthy balance of billable and operational work.
- 75-85%: Excellent. Be careful not to push beyond this; burnout follows.
- Above 85%: Unsustainable. Your team has no room for development, internal projects, or unexpected work.
AI adds context: Rather than just showing you a percentage, AI breaks down non-billable time into categories. How much is unavoidable admin? How much is internal meetings that could be shorter? How much is context-switching between too many projects?
Capacity recommendations
AI analyses utilisation patterns and recommends adjustments. This feeds directly into resource planning and capacity management:
- “Designer A is at 82% utilisation and trending upward. Consider redistributing the Smith project to Designer B, who is at 58%.”
- “The content team collectively spent 35 hours on internal meetings last month, up from 22 hours the previous month. Review meeting cadence.”
- “Project X has consumed 40% more hours than scoped. Flag for a pricing conversation with the client or scope reduction.”
The gap between logged and actual time
This is the metric most agencies are afraid to measure, and the most important one.
When agencies switch from active to passive tracking, they consistently discover that actual time spent on client work is 15-25% higher than what was being logged. That gap represents revenue you earned but never billed (on hourly projects) or margin you thought you had but did not (on fixed-price work).
What to do with this data:
- Adjust your pricing. If a project type consistently takes 25% longer than your team logs, your pricing is 25% too low (or your scope is too generous).
- Identify efficiency problems. Which tasks take longer than expected? Where does time disappear? AI highlights the patterns.
- Improve scoping. Use real time data (not logged time data) to scope future projects. Your estimates will be more accurate, your margins more predictable.
Project profitability insights
Combining accurate time tracking with financial data gives you true project profitability.
The calculation: Revenue from the project minus (actual hours multiplied by team cost per hour) equals gross profit. Divide by revenue for the margin percentage.
What AI reveals: Projects that looked profitable on paper (based on logged time) that are actually break-even or loss-making when measured on actual time. We regularly see agencies discover that 20-30% of their projects are less profitable than they believed.
The fix is pricing, not time tracking. Better data does not fix margins on its own. But it gives you the evidence to have honest conversations about how to price agency services, scope, and which clients are worth keeping.
The argument for passive over active
If you are still debating which approach to take, here is the summary.
Active tracking gives you data your team remembers to enter, rounded to the nearest 15 minutes, and categorised by their best guess. It costs your team 15-20 minutes per day in admin.
Passive tracking gives you data captured automatically, accurate to the minute, and categorised by AI pattern recognition. It costs your team 3-5 minutes per day in review.
The data from passive tracking is better. The team experience is better. The insights are better. The only reason not to switch is if your team has genuine concerns about monitoring, and those are worth addressing directly rather than avoiding with an inferior system.
Start measuring properly. Then start improving.
This is part of Margin Watch, a series on how AI is reshaping the business of running an agency. Subscribe to the newsletter to get new articles weekly.
