Watch the Walkthrough
Rize co-founder Macgill Davis walks through the whole product in about fifteen minutes, from how tracking works to how to configure tagging so it stays accurate.
The sections below follow the same order, with the exact settings paths and links to the deeper guides.
Start on the Calendar
The calendar is where tracking is visible and where you will spend most of your review time. The left column is your live activity stream, captured by the macOS or Windows app once tracking is enabled.
Rize records the metadata of your active window only: app name, window title, and URL. No screenshots, no keystrokes, no window contents. That is what keeps passwords, client records, and health information out of the system entirely. See how automatic tracking works for the full list, and privacy and security for the controls.
Raw activity is never shared at the team level. It is an input to the algorithm, not a report.
Review Time Entry Suggestions
Rize turns your activity into time entry suggestions: a start time, an end time, an AI-written description, and tags. They are the same fields you would fill in by hand, drafted for you.
Every suggestion is editable before it counts. You can:
- Edit the description or the tags
- Merge two adjacent entries that should be one block of work
- Split a long entry that covered several things
- Regenerate the suggestion from the underlying activity
- Accept it as is, or reject it
Rejected time is never shared. Accepted time goes to your team.
If your workspace leaves automatic approval on, a pending entry is approved for you roughly eight hours after it ends, so nothing sits in limbo when you are heads down. Workspace admins can turn that off in team settings and require an explicit approval for every entry.
The workflow that works: let Rize track in the background, then review either a few times a day or once at the end of it. Reviewing time entries covers the review panel in detail.
Know the Four Tag Types
Rize has four tags, and they answer different questions.
- Task answers which unit of work. Tasks carry assignees and due dates, which is why they map cleanly onto ClickUp, Linear, or Asana.
- Project answers which body of work. Tasks belong to projects.
- Client answers who the work is billable to. Projects belong to clients.
- Label answers what kind of work it is. Every entry gets one, and labels never carry a due date.
Tasks, projects, and clients form a hierarchy that describes who the work belongs to. Labels sit outside it and describe the nature of the work instead.
All four are scoped per team, so switching teams switches the whole tag set.
Turn Off the Levels You Do Not Use
You can disable task, project, or client tagging so Rize only asks for what you actually bill or report on. This is the single fastest accuracy win, because the AI stops guessing at a level you do not care about.
- Accounting and professional services firms often keep clients and labels, and disable projects and tasks
- Product teams often keep projects and labels only
- Teams synced to a task tracker often keep tasks and labels, and disable projects and clients
Labels stay on regardless. See setting up a team for where these live.
Configure Labels With Prompts
Labels are tagged entirely by AI prompt, which is why they are the most accurate tag out of the box. Each label has a title, a color, and a description, and Rize passes the full set to the model when it tags an entry.
Go to Settings > Teams > Settings to see the labels for a team. The defaults cover common work types. Rewriting the descriptions to match your own vocabulary is worth doing on day one, and it is the reason a law firm, an agency, and a product team end up with different label sets.
You can also create and edit labels through the Rize agent or the MCP server rather than the UI.
Configure Clients, Projects, and Tasks With Keywords and Prompts
Clients, projects, and tasks are tagged from two inputs: keywords and an AI prompt. Open any client, project, or task and you will find both fields.
- Keywords are literal strings that appear in your activity: a project code, a ClickUp task ID, a Google Workspace domain, a repository name. Keywords mostly drive how the timeline gets split into entries.
- Prompts are plain-language instructions describing when this tag applies. Prompts mostly drive which tag gets picked.
Use both. A keyword catches the obvious matches, and the prompt handles everything the keyword misses. Rize also updates these fields as it learns from your corrections.
When the model is confident it applies the tag automatically. When it is less sure you get a recommendation with a confidence score, and one click applies it.
Tune Entry Size and Tagging Instructions
The Settings tab on the calendar holds the two controls that change how entries get built.
- Time entry size is a slider. Drag left for shorter entries, right for longer ones. Adjust this first if entries feel too coarse or too fragmented.
- Custom tagging instructions is a full AI prompt covering how your time should be tagged. It accepts org skills, and Rize updates it as it learns your patterns.
Both can be set for yourself under Prompts, and admins can set them at the team and workspace level so a whole team starts from the same instructions. Personal guidance is the place to describe what you work on and how you want Rize to interpret it.
Read Your Timesheet
The timesheet page groups your entries by project and is the fastest way to fix categorization in bulk. Drag an entry onto a different project to recategorize it, and accept all pending entries in one action.
Reporting for Admins
The Time Entries tab is the team view. Members see their own data. Team and org admins see everyone's, with a team switcher and a date range in the top right.
The page gives you a breakdown by member, label, and client, plus a pivot table you can regroup. Group by member to see how each person's work splits across labels. Group by label or client to see it the other way. Expand any row for the underlying entries, jump to a member profile to compare them against the team or against their own previous period, or open a chat with the agent scoped to what you are looking at.
The Timesheets page is the one to use for capacity and utilization. Filter by client, project, label, or task, set the range, and read who worked on what.
AI Reports, Routines, and the Agent
Three agentic surfaces sit on top of your tracked time.
- AI reports are saved questions that run on a schedule and deliver a written analysis by email, Slack, or in-app. Start from a template, export finished runs to PDF, or chat with a run to dig into it. See AI reports.
- Routines run from your own perspective rather than the team's: a start-of-day plan, an end-of-day recap, a tagging accuracy check.
- The agent is the chat interface over everything, with shared skills you can publish across a team so an improvement reaches everyone at once. See using the Rize AI agent.
Track Client Profitability
Upload an hourly or retainer contract per client and set an hourly cost rate per team member, and Rize reports margin live instead of at month end. Tracking profitability covers the setup.
Connect Rize to Your Own Tools
Anything you can do in the Rize interface you can do through the API or the MCP server, which is what connects Rize to Claude, ChatGPT, or an agent you build yourself. See using Rize with AI assistants.
Your First Week
- Install the app, grant permissions, and let it track for a day
- Disable the tag levels you do not bill or report on
- Rewrite your team's label descriptions in your own vocabulary
- Add keywords and a prompt to your top few clients or projects
- Review your entries for a minute or two each day, correcting rather than only approving
- Adjust the entry size slider if entries are consistently too long or too short
Corrections in week one are what the AI learns from. Everything after that is reporting.