Plans 💳 : Everyone

User Permissions 👥: Everyone

Having an automatic digital memory of your working day is valuable for a ton of reasons, but even more so is the ability to quickly transform that data into billable or logged hours. Though you can log time in Timely manually or with a timer, the real power of our tool lies with automated time tracking powered by a feature we call Memory AI.

Memory AI essentially takes all the activities collected by the Memory app, assigns them to the right project and drafts a time entry for you. It does this by using your time logging patterns to learn how to group your tracked activities. Every time you log a time entry using tracked memories in your Memory Inbox, Memory AI learns associations between your projects and activities. The more hours you log using tracked memories, the more accurate Memory AI's draft entries will become! Before you know it, the Hours tab will be filled with automatically constructed entries. 🤖

By default, Memory AI is enabled for each new user as they are invited or sign up for Timely.

You can enable to disable this feature by clicking the Memory icon, shaped like a brain, at the top of your timeline in the "Hours" section. Simply click the toggle button to switch this feature on or off.

Draft time entries will shortly start appearing alongside logged and planned time entries in the Hours tab. You can recognize them by their diagonal stripes. The draft entry color relates to the project Memory AI thinks your grouped activities belong to:

From Week view, you'll see a card at the top of each day letting you know how many drafts you have to review:

If an AI draft contains memories which don't belong to the project indicated ー or if it’s missing any memories ー you can drag and drop memories between your Memory Inbox and your entry to correct it. Once correct, hit "Log entry". You can then add notes and tags, and click "Accept" to save the entry:

For those new to Timely or who have just enabled the AI features, you might not receive drafts right away or on a regular basis. If this is the case, ensure you are logging your hours from your Memory Inbox. When drafts do arrive, remember to adjust, accept or reject them as appropriate. By doing these two things, you are helping your AI to learn and serve you better.

Note: Memory AI won’t make suggestions based on memories that have already been logged to projects. If you find you haven’t received any drafts after a while but have been logging consistently, try waiting for Memory AI to make some suggestions for you. The end goal is for you to not worry about logging your hours, so hand the wheel to Memory AI and keep an eye out for drafts you can review and accept! ✅

You can train the AI at any point just by reviewing memories on your timeline and making sure they're associated with the correct projects. You can create new projects, search through all of your active projects, ignore memories or skip the memories you don't want your AI to include.

If your logging patterns or the way you use Timely ever change, you also have the option to reset your AI model from scratch. After clicking the Memory icon you can click the "Reset" button, then confirm "Reset Training" to remove any learned behaviors and restart your AI's training afresh.

Note: Hitting reset will fully delete the previous model, so any patterns Memory AI established will be gone forever. 💨

Why am I not receiving AI drafts?

Before receiving AI drafts, first make sure you have Memory installed (Mac and Windows) and that you're receiving memories. If you're having trouble, here's a helpful article to get you started.

Also, keep in mind that Memory AI needs to be trained before it can start drafting your time entries accurately. As you log entries, Memory AI will learn your logging patterns and project classifications. With this knowledge, it can start to create drafts for you. Then all you need to do is review your drafts, make any edits and approve! Easy as 1-2-3!

Where can I learn more about Memory AI?

Our developers are both really smart and really generous with their time, so they’ve offered up some personal insights while answering FAQs from our users, as well as a tour under the hood explaining how our machine learning processes work.