What is AI Time Tracking?

AI time tracking — also known as automatic timekeeping or passive time capture — uses artificial intelligence (AI) or machine learning (ML) models to track what tasks someone is working on and helps generate timesheets in order to bill their clients.

AI time tracking applications aim to reduce the time and effort required to track tasks, capture more billable hours, and improve compliance with firm and client billing guidelines.

The concept of passive time capture has been around for several years. Originally, these applications collected metadata and used simple machine learning models to interpret it. More recently, the use of large language models (LLMs) has enabled a new generation of more intelligent and highly automated time tracking applications.

In this article, we’ll walk through the different approaches to AI time tracking, how the software works, and the main players in the space.

What is Time Tracking?

Any professional who bills their clients on an hourly basis has to keep track of what work they’ve done and for how long. This is a common practice across law, accounting, consulting, IT/engineering, and other professional services firms.

Typically, this process involves creating time entries, which list the client and project worked on, the amount of time spent, and a narrative description of the work performed. Depending on the industry and client, there might be additional requirements like task or activity codes, which indicate the category of task being performed.

For the purposes of this article, we will focus primarily on time tracking within the legal industry.

Problems With Manual Time Tracking

Although time tracking sounds simple in concept, it can be challenging in practice. In the legal industry in particular, attorneys may work on many different clients and projects in a given day and constantly switch focus between them. Accurately recording how much time is spent on each task and writing high quality narratives describing the work performed can be a time-consuming and error-prone process.

This leads to several downstream problems. First and most obvious is the actual time spent tracking time and creating timesheets — which is often around 15-30 minutes per day.

Second, the data shows that lawyers who record their time manually are consistently forgetting to record some of the time they work. At PointOne, we’ve found that most firms are losing around 10% of their billable hours this way, depending on their practice area and existing timekeeping practices. Sometimes we see that number as high as 25-30%.

Third, attorneys who track their time manually often produce low quality time entries. It’s common to see narratives that are not descriptive of the value of the work, like “Attention to matter.” Time entries may also be non-compliant with “outside counsel guidelines” (rules provided by client on how they can be billed) which can cause delays in collecting on bills.

Finally, lawyers just really hate timekeeping. PointOne ran a poll of over 700 lawyers, with 87% saying they “hate” time tracking and 21% calling it “the worst part of my job”. We often hear lawyers who work in-house say that the best thing about leaving private practice is no longer having to track their time.

Historical Approaches

Before the age of computers, lawyers used to fill out physical timesheets on specially designed legal paper. This was later digitized, but the overall process remained more or less the same.

In the early 2010s, the first “passive time capture” systems began to emerge — Element55 and RescueTime were early players. These systems would collect data from a user’s computer — for example, the names of documents they had opened, emails they sent, and events on their calendar — and present them to the user, either in a software interface or by email. The theory was this would help users remember what they had worked on in a day, reducing mental burden and capturing otherwise lost billable hours.

The issue with passive time capture systems is that, although they aggregate tasks in one place, they still leave almost all the work to the end-user to turn those tasks into time entries. Further, these systems are not smart enough to aggregate data by task or project, resulting in hundreds of logs to review every day. As a result, usage of these tools within firms tends to be low.

Over time as machine learning models improved, new functionality has been added to these systems. One major improvement is the ability to predict which client and project a task is associated with.

Some systems have also added the ability to check for compliance against firm or client billing guidelines — for example, prohibiting block billing, requiring activity codes, or prohibiting certain keywords. Typically these function in a simple, rules-based way.

These features have helped solve some of the initial barriers to adoption by improving the usefulness for lawyers and firms. However, until recently, the overall quality of the output remained low and still required a lot of effort from attorneys to prepare time entries.

Modern AI Time Tracking

In late 2022 and early 2023, we witnessed the explosion of modern Large Language Models (LLMs), starting with ChatGPT. Since then, we've seen LLMs applied across nearly every sector of the economy, from legal services to general purpose tools for web automation to more specialized niches like the construction industry.

From these, a number of new timekeeping systems have emerged that leverage these new capabilities. PointOne and Laurel are two newer AI-powered timekeeping products that leverage LLMs natively to generate time entries.

The advantage of using LLMs to track time lies in their intelligence and flexibility. Lawyers are typically working simultaneously in many documents, web browsing windows, emails, meetings and more. Prior to LLMs, time tracking systems were not powerful enough to properly understand and interpret all this activity.

Modern AI time tracking software captures information about the tasks a user is performing. Typically this includes data from their local computer (e.g., what windows are being opened), as well as from their mobile device and all the different apps they work in (e.g. Microsoft 365, Chrome, Zoom, etc). The system then uses AI to sort through thousands of logs, interpret what work was being performed, assign it to a client and project, and generate a time entry.

Unlike historical ML models, LLMs that are given proper context and prompting can be very effective in condensing down many tasks into a simple, well-constructed time entry. This allows modern systems to produce time entries that are much closer to human-level quality and are accurately assigned to clients and projects.

As LLMs have gotten faster and more intelligent, the quality and accuracy of time entries they produce has also improved. This trend will likely continue as better and better models are released.

Benefits of AI Time Tracking

There are a number of benefits for firms and attorneys who use AI timekeeping software.

Increased billable hours.

Since software automatically captures all the tasks lawyers are performing, these systems can ensure that lawyers don’t forget to record any billable hours. This typically results in an increase in total hours of 5-20%+, depending on the firm and practice.

Less time spent on admin.

Automatic timekeeping software can significantly reduce the amount of time spent on timekeeping and billing admin. At PointOne, we often see daily timekeeping admin go from 20-30 minutes down to less than 5 minutes.

Better billing compliance.

Since AI generates time entries based on the firm’s billing guidelines and outside counsel guidelines, those entries are less likely to be contested by clients. This results in the firm being able to collect on bills at a higher rate, and get paid faster.

Improved quality of life.

A large proportion of attorneys find time tracking to be a daily pain point. And given the heavy demands of client work, timesheets are often deprioritized and end up eating into nights and weekends. AI timekeeping massively speeds up and simplifies this process, giving attorneys back time with family and friends.

Higher quality data.

In theory, time entry data could be used for a variety of purposes within law firms, from pricing and budgeting, to experience management, and more. Unfortunately, time entries are drafted in natural language and are usually “block billed” and not well tagged or categorized — making them challenging to use. AI timekeeping systems can capture higher quality data up front, which enables better use by firms.

Integrations with Other Tools

App Integrations

Given that AI time tracking systems need to automatically figure out what work a lawyer is performing, there are a number of integrations that are required.

Some common app integrations include:

• Email providers (Microsoft and Google)

• Calendar providers (Microsoft and Google)

• Microsoft Word, Excel and Powerpoint documents

• Adobe PDF

• Phone providers (Cisco, Zoom, Teams, etc.)

• Web browsers (Chrome, Edge, Firefox, Safari)

• Integrations with bespoke legal software such as Westlaw, Lexis, document management systems, and more

Practice Management Integrations

Once that data is collected, the software uses AI to interpret what work was performed and generates time entries. Those time entries are then usually synced with an underlying legal practice management system, such as:

• Aderant

• Elite

• Surepoint

• Centerbase

• Actionstep

• Clio

• Filevine

• MyCase

Integrations on both sides — data collection and syncing with the practice management system — are critical for a tool to be accurate, effective and easy to use.

Popular AI Time Tracking Tools

There are a number of different solutions on the market that use AI to help lawyers capture time. Six of the most popular tools are:

• PointOne

• Intapp Time

• Aderant iTimekeep

• Laurel

• TIQ Time

• WiseTime

Each tool has a different sector focus, with some serving big law firms, some focused on small firms, and some serving firms of all sizes. Some of the tools also focus on industries outside of legal, such as accounting and consulting.

These tools also vary in their level of automation. PointOne allows for one-click timekeeping, with AI systems doing all of the work to interpret data, aggregate tasks, and produce a human-like time entry. Laurel also uses AI to automatically generate narratives.

Other tools like Intapp, iTimekeep and TIQ Time use AI to identify tasks and attempt to assign them to clients, but still require the user to turn those into time entries and write narratives describing their work.

Final Thoughts

Although the concept of AI timekeeping has been around for several years, it is only recently that LLMs have made it possible to generate time entries automatically, with quality and accuracy that’s comparable to a human.

For firms willing to explore this new way of entering time, there are significant opportunities to improve profitability, efficiency, and quality of life for their attorneys.

While law firms can be slow to adopt new technologies, over the next several years we expect to see AI time entry systems become the dominant method for attorneys to track their time.