Managing Your AI Employees

An AI employee isn’t a one-and-done tool. It builds up memory over time, runs more and more on its own, and quietly racks up usage. This page covers what you need to know to keep an employee healthy long-term: pausing, cleaning up memory, checking costs, tidying automatic runs, managing your skill library, adjusting its AI work computer specs, transferring ownership, and deleting.

Fast hire: create a new employee instantly, no conversation needed

Press + Hire at the top of the Home screen and a dialog opens asking how you want to grow your team:

👥 Hire a team: “Tell us what needs doing and we’ll put together the right team.” Describe your business or a task, and several employees are assembled into a team at once. See Working as a Team for the full flow.
🧑‍💼 Hire one employee: “Pick an experienced hire who can start today, or take on a new employee and train them yourself.” The fast hire flow covered on this page is this option.
Just create an empty team →: Doesn’t add anyone; it only creates an empty team so you can group employees you already have. It’s a smaller, secondary link at the bottom of the dialog, below the other two.

Choosing Hire one employee opens a confirmation dialog that can show two paths at once:

Experienced-hire cards (up to 3): employees from popular ready-made employee templates UpServe built, with a job already defined. Tap one and that employee is hired instantly (personal accounts only; this doesn’t appear when hiring into an organization).
Hire a new employee: tap “Confirm” (when no experienced-hire cards are shown) or “Hire a new employee” (when they are), and a hire with no name or role yet appears immediately, asking on its own: “Tell me what you’d like me to take care of.”

There’s no chatting with a manager first to decide what to assign. Starting an employee from scratch goes through this card and nothing else (installing an employee template from Templates, which brings over an employee someone already built, is always available separately). No experienced-hire cards showing is completely fine. They’re a side note, and without them you just see the plain hire-a-new-employee button.

Through the conversation that follows (Work setup), the new hire fills in its own name, role, and personality based on how you answer.

The + Hire button and the choices inside it exist only on web. In the mobile apps (iOS and Android), the + menu at the top right lists the same three options directly: Hire (fast hire, experienced or rookie), team hiring (only on accounts with team hiring turned on), and creating an empty team. There’s no extra dialog in between like on web (the exact wording for team hiring and empty-team creation varies slightly between iOS and Android). The confirm button reads Hire on mobile.

Hiring straight from a team’s screen is web only.

A hiring suggestion card may show up too

The hire confirmation dialog sometimes carries a hiring suggestion card. When one of your current employees is clearly overloaded (a backlog of pending work, a recent run of requests concentrated on that one employee, or a heavy load of recurring scheduled runs), UpServe surfaces that evidence on a card and asks “How about hiring someone to share this employee’s load?”

Accept it and the card’s “Hire someone to help” button hires right away, carrying “why this hire happened” into the new hire’s Work setup conversation.
Not interested? “Not now” dismisses it, and that same evidence won’t be suggested again.
No card showing is completely fine. It’s a side note, not a requirement: with no card, the plain “Confirm” button hires just the same. (When a suggestion is shown, that button is relabeled “Hire directly”, meaning you’re bringing someone new on regardless of the suggestion. If experienced-hire cards described above are also showing, this button reads “Hire a new employee” instead.)

This card works the same way on web, iOS, and Android.

What’s different during Work setup

When the instruction is specific enough, the new hire skips ahead: it starts on the first piece of actual work right away instead of asking about its name or other details first, and sorts out its profile (name, personality) after showing you that result. It only asks up front for what would genuinely block starting (like source material you haven’t sent yet); everything else gets a sensible default, shown in the result. If you don’t like it, a one-line correction is all it takes.

If you can’t yet name a specific job (you’re just starting a business and aren’t sure what to hand off, or you haven’t sorted your day into tasks), the new hire won’t throw an open-ended question back at you. Instead, it asks a couple of quick things about what you do, then proposes 2 or 3 concrete jobs it could take on. Pick one or reshape it, and that becomes its assignment. If you already know what you want to hand over, you never see this step at all.

While the new hire is editing its own profile (name, role description, personality) based on your answers, you won’t see the usual approval prompt that shows up before risky actions. Instead, a change card appears in the chat. It shows the before/after content side by side, and if you don’t like the change, the card’s “Revert” button restores the previous content right away.

Once the new hire summarizes its proposed name, mission, and description and asks you to confirm, and you agree, Work setup ends, and you can use the rest of this page’s management features (active/inactive, memory cleanup, logs, and so on) exactly like any other employee.

Work setup doesn’t end on its own just because time passes. Leave the conversation mid-way and the new hire just waits. If it sits untouched for a long while (72+ hours), the progress card at the top of the chat switches to “Stalled.” Either way, just talk to the new hire to pick it back up, or use the employee page → Settings subpage → “Resume Work setup” button.

How many employees can I hire?

Every plan lets you own and run employees without a headcount limit. You can hire more regardless of how many you already have, and new hires join ready to work without waiting for a free slot.

Work consumes SU based on actual usage, and SU balance, budgets, and automatic run AI usage limits still apply. Repeated hiring requests in a short time can trigger a request rate limit. If that happens, wait a little before trying again.

At a glance

What you want to do	Where to do it
Hire a new employee instantly, no conversation	Web: Home screen → ”+ Hire” → “Hire one employee” · Mobile: ”+” at top right → “Hire”
Hide an employee from the list and block all runs	Employee page → `Settings` subpage → “Employee Status” toggle → INACTIVE
Pause automatic runs temporarily	Employee page → `Automatic runs` subpage → toggle each scheduled run OFF
Stop the employee from waking up on its own	Same place, toggle `Self check-in` OFF
Undo something the employee learned wrong	Ask the employee in chat: “Forget that note about…” / refresh the Memory & instructions panel
See how many SU you used this month	Employee page → `Logs` subpage
Rate whether an automatic run’s result was useful	Right below that result message (👍/👎)
Clean up unused or outdated skills	Skills (`/skills`) page → “My Library” tab → cleanup banner/modal
See in one place what an employee runs without asking and what it can reach	Web: Employee page → `Permissions` subpage
Change its AI performance tier or AI work computer specs	Web: Employee page → `Permissions` subpage → “Computer & AI tier” · Mobile: the employee’s `Settings` screen
Keep a go-to employee at the top of the list	Web: right-click the employee in the avatar list on the left → “Pin to top”
Stop an employee from proposing its own tier changes	Web: Employee page → `Permissions` subpage → “Self-requests” → “Propose AI tier changes” toggle OFF · Mobile: the employee’s `Settings` screen → “Receive tier change proposals” toggle OFF
Adjust whether an employee can read-only or read/write a table	Files & data screen → “Permissions” button next to the store selector (for just one employee on the web: Employee page → `Permissions` subpage → “Data access”)
Transfer a personal employee to an organization	Employee table → drag the employee’s row onto an org section (or an org team), or the employee menu (⋯, right-click, or ⋮ at the top right of the chat) → More → “Transfer to organization”
Move an org-owned employee back to a member’s personal ownership	Employee list ”⋯” menu → “Transfer ownership (to member)” (organization owner only)
Remove an employee you no longer need	Employee page → `Settings` subpage → “Delete”

1. Active / Inactive: lock an employee without deleting it

If you want to put an employee on hold without removing it, flip the Employee Status toggle to INACTIVE.

On the employee page → Settings subpage, find the “Employee Status” row and switch it OFF (INACTIVE). Here’s what happens:

The employee is shown as inactive in your list (dimmed or marked separately from active employees).
All automatic runs (scheduled runs, self check-ins, webhooks) are blocked.
However, you can still chat with it directly. (If it was auto-paused for inactivity, sending a message reactivates it automatically.)

Switch it back ON (ACTIVE) at any time to restore everything immediately. Settings, memory, and skills are fully preserved.

This toggle is not shown for system employees such as the Manager.

Away from UpServe for a while? Your employees pause themselves

Even without touching anything, if you haven’t used UpServe at all for about a week, your employees are automatically switched to INACTIVE. You’ll get a heads-up notification one day before that happens, and another notification when the switch actually occurs.

The key thing to know: “using UpServe” counts across any of your employees. You don’t have to open a specific employee. As long as you chat with any employee or open the app, all of your employees stay active. This means an employee quietly doing its job in the background doesn’t get shut off just because you rarely open its own page. It only pauses if you’ve stepped away from the service entirely.

Waking it back up is simple. An employee auto-paused this way returns to ACTIVE the moment you send it a message, with no toggle to flip.

This automatic switch only applies when you’ve been away from the service entirely. If you turned the toggle off yourself, or the employee was paused due to low SU, sending a message won’t turn it back on. Fix the underlying cause first, then turn it on manually.

2. Pausing: turn off automatic runs for a while

UpServe employees can wake up without you in three ways:

Scheduled runs: fire at a set time or interval (e.g., every morning at 9)
Self check-in: during defined active hours, the employee wakes on its own, decides when to check in next, and does real work within its allowed scope when there’s something to do
Webhooks: external services ping the employee

If you want an employee to sit still for a few days, switch these automatic runs off. Each is a single toggle, and turning it back on restores the same setup.

Turn off scheduled runs

Open the employee page and switch to the Automatic runs subpage (/agents/<agent>/schedules).
Flip the toggle on the right side of each scheduled run card to OFF.
When you’re ready, toggle it back ON and it resumes from the next scheduled time.

You don’t have to delete scheduled runs to take a break; toggling is the easiest way.

Turn off self check-in

Self check-in is the feature that wakes the employee inside defined active hours so it can ask “anything to do right now?” and handle it within its allowed scope if there is. The gap between check-ins isn’t fixed. It adapts to activity, stretching when things are quiet and tightening when the employee is busy. There’s one way to disable it:

Quick OFF: on the Automatic runs subpage, toggle the self check-in card OFF. (The editor always keeps at least one active-hours band, so emptying all bands to disable self check-in isn’t supported.)

While self check-in is off, the employee only runs when you message it directly or when a scheduled run or webhook fires.

Turn off webhooks

On the Webhooks subpage, switch the webhook OFF and external pings won’t wake the employee.

If you just want to stop the current burst, you can also chat “stop for today.” But that only applies to the run you’re in, and the next scheduled run will still wake it. Use the toggles above for a real pause.

On web, you can close a completed or stopped progress card with ×. A running card can be collapsed.

3. Cleaning up memory: undo what the employee learned wrong

Every employee keeps a personal notebook of what it has learned: your preferences, frequently-used data, lessons from past tasks. Sometimes a note ends up wrong or outdated.

See what’s in the Memory & instructions panel

In the employee chat view, click the memory icon in the right-side toolbar to open the Memory & instructions panel. You’ll see:

Identity, guidelines, and memory index: the employee’s core instructions and list of memories
User (user): your preferences and background
Feedback (feedback): lessons captured automatically after tasks
Context (context): ongoing situations, project state, and external-service details
Reference (reference): frequently-consulted source material
Other (other): memory files that don’t fit a specific type

Click any file to read or edit it. If the panel looks stale, hit the refresh button in the top-right of the panel.

Fixing wrong notes

The easiest way is to just tell the employee in chat. You don’t have to edit files by hand.

Situation	What to say
Drop a specific note	“Forget the note about me liking spicy food.”
A fact changed	“My office address changed. Please update memory.”
Refresh how the employee does something	“We don’t use that report format anymore. Clean up the related feedback notes.”

4. What you can change in the Settings and Permissions subpages

On the web, the employee page splits this work between the Settings subpage and the Permissions subpage. The first table below covers Settings; the second covers Permissions. The mobile apps (iOS and Android) still change the second table’s items from the employee’s Settings screen for now.

Settings subpage

Setting	What it controls
Avatar	Upload an image or generate one with AI
Name	How the employee appears in your list
Persona	Mission and description. Defines the employee’s role and personality
Analysis depth	Visible when the assigned model supports adjusting analysis depth. Deeper means slower replies and more usage. Choose default (same as medium, and automatic work such as scheduled runs uses the same depth), none (fastest, minimal reasoning), low (faster, cheaper), medium (balanced), high (more thorough, higher cost), or xhigh (deepest analysis, highest cost). For none/xhigh, the model automatically falls back to its lowest/highest supported level
Session Auto-Reset	When to automatically clear conversation history: never / daily at a set hour (UTC) / after idle timeout
Unread Message Limit	Once this many conversations go unread, runs the employee starts on its own (self check-ins, page watches, webhooks, schedules it set itself) are skipped until you open the chat. Schedules you asked for keep running (0 = unlimited)
Auto memory cleanup (every N conversation turns or scheduled runs)	Tidies memory automatically after N responses to people or teammate employees, or scheduled runs (default 6; 0 turns it off). Background self check-ins don’t count. Runs at most once every 6 hours, and once if 30 days pass without a cleanup. See the Advanced section of How AI Employees Wake Up for the exact rules
Automatic run AI usage limit (last 24 hours)	The cap on AI usage (SU) this employee’s automatic runs (scheduled runs, self check-ins, and so on) can spend in any 24-hour window. Once reached, no new automatic run starts; runs already in progress finish. Runs start again as earlier usage drops out of the 24-hour window. Messages you send directly are still handled, but child runs started while handling them can be held by the limit. Use the account default or set a separate per-employee limit (up to the platform default). Raising it past the platform default or making it unlimited (0) is not something you can do yourself. To change it, please contact the UpServe team
Tool indicators	Whether to show real-time tool execution indicators (including “Thinking…”) in the chat
Channels	Whether this employee’s results send app push notifications
Self check-in	Set the baseline interval and active-hours bands
Employee Status	ACTIVE / INACTIVE toggle (see section 1 above)
Skills	Add or remove installed skills
Delete	Permanently remove the employee (danger zone)

The Settings subpage has an “Open Permissions” card that takes you straight to the Permissions subpage. Per-connection auto-approval for external MCP connections is still changed in Settings.

Permissions subpage (web)

The employee page → Permissions subpage shows in one place what this employee can read and write, what it runs without asking, and which outside services it can reach. You can also open it from the Permissions row in the employee menu (formerly “Approval rules”). Changes here save right away, with no save button.

Group	What it covers
Data access	How much the employee can read and write table data (data role), and the files it can read or edit
Run automation	The Tool Approval Policy (`Run without approval` / `Approval Required` per tool), which computer use actions are allowed (click, type, screenshot, etc.), the “Also allow public GitHub repositories” switch, and the Active auto-approvals list chosen on approval cards, along with approval rule suggestions
External access	Connection details, and the scope you approved for the employee to call outside services and web apps
Computer & AI tier	The AI performance tier (`Standard` or `Ultra` (💪); UpServe decides which AI model backs each tier, and `Ultra` spends far more SU) and the AI work computer specs (see section 8 below)
Self-requests	Whether the employee may ask with an approval card when it needs more. Propose AI tier changes (e.g., “I’d suggest moving up to Ultra”) lives here too. It defaults to on; turning it off means this employee never proposes a tier change on its own (this is separate from you changing its tier manually). A request applies only after a person approves it
Recent changes	Automatic changes, approval cards, and direct changes, newest first. Computer spec, AI tier, and tool approval changes aren’t recorded yet

For employees in an organization space, only the organization owner or a delegated operator can change the AI performance tier and AI work computer specs. Other members see them locked with “Ask the organization owner to change this.”
The AI tier switch in the chat input’s ”+” menu is still there.
Analysis depth and a pinned specific model (when shown) stay in the Settings subpage.

5. Reviewing runs and cost

On the employee page → Logs subpage (/agents/<agent>/logs) you can see recent usage and SU consumed.

Use the 7D / 30D / 90D tabs to select the time range.

What’s shown:

Total SU used in the selected period
Total requests: number of model calls
By run condition: split across your messages / scheduled runs / webhooks / self check-ins, etc. (mentions and other run types fall back to a gray bar)
By cost type: AI model usage vs. Tools & computer usage (AI work computer, etc.). AI model usage covers not just visible replies but also model calls for internal planning, judgment, and memory cleanup
By model: which model was used how much
Daily trend: a small bar chart of recent activity
Scheduled run history: when each automatic run happened and how it ended

If your SU are dropping faster than expected, check the “By run condition” breakdown first to see if self check-ins or scheduled runs are doing more than you intended. A self check-in with too short a baseline interval will keep spending SU even when you’re not around.

6. Tidying scheduled runs and webhooks

Toggling a scheduled run or webhook OFF stops it from running. To keep the screen clean, delete the ones you definitely don’t need anymore.

Delete a scheduled run: on the Automatic runs subpage, hit the Delete button on the card.
Disable webhooks: on the Webhooks subpage, switch OFF.
App push notifications: on the employee page → Settings subpage → Channels section, toggle the app push notification switch off.

For self check-in, prefer toggling OFF or clearing the active-hours bands instead of deleting. That way the same profile is ready to go when you turn it back on.

7. Skill Library Cleanup

Install enough skills and things start to pile up: skills you never used, skills that keep failing, two skills that do essentially the same thing. This feature lives on the Skills tab’s “My Library” screen, not in an individual employee or a separate account settings page. Open Skills (/skills) → My Library; a banner at the top shows the number of cleanup candidates, and clicking it opens a modal listing what UpServe scans for periodically, organized as individual cards.

What categories does it detect?

There are six.

Category	What it means
Unused (UNUSED)	Installed more than 30 days ago but never run
Failing (FAILING)	Has a high failure rate over the last 30 days (only judged once there’s enough run history)
Outdated (OUTDATED)	The original skill has been updated but your copy is still on an old version
Deprecated Origin (DEPRECATED_ORIGIN)	The original skill has been marked deprecated and is no longer actively maintained
Duplicate Function (DUPLICATE_FUNCTION)	Overlaps significantly with another skill you already have
Revoked (REVOKED)	Your installed version was confirmed malicious and pulled. Update to a safe version or remove it

How to use it

Open the “My Library” tab on the Skills (/skills) page.
At the top there is a periodic-scan toggle. When it is ON, UpServe runs the scan automatically on a regular schedule. When it is OFF, automatic scanning is skipped. When there are no current cleanup candidates, the empty screen shows a “Run Now” button you can use to trigger a manual scan.
When cards appear, select the ones you want to act on using the checkboxes, then click “Apply Selected” or “Dismiss Selected”.
- Apply: carries out the recommended action for each item (remove the installation, pull in an update, etc.).
- Dismiss: skips those items for this round. They may reappear in the next scan cycle.
To clear everything at once, use “Dismiss All”.

To learn how the autonomous skill system improves skill quality behind the scenes, see Skills & the Autonomous Skill System.

8. Adjusting its AI work computer specs

Every time an employee runs, it uses a dedicated AI work computer. This isn’t your computer; it’s one UpServe prepares for the employee. You can pick that computer’s CPU/memory spec from six tiers.

On the web, open the employee page → Permissions subpage and look for “AI work computer specs” in the “Computer & AI tier” group. The mobile apps still show it in the employee’s Settings screen. For employees in an organization space, only the organization owner or a delegated operator can change it.

Tiers	Screen	Best for
NANO / SMALL / MEDIUM	No screen	Lighter, cheaper work (running code, processing documents, etc.)
LARGE / XL / XXL	Desktop screen	Work that needs a screen, like browser automation or computer use, and heavier workloads

The default is “Auto.” Leave it there and the employee gets the default spec UpServe picks.
If you pick one manually, higher tiers cost more SU per minute. Each option in the list shows its per-minute SU rate.
You’re not charged while the computer is powered down (asleep). SU are only spent while it’s actually running.
🔴 Picking a “no screen” tier (NANO, SMALL, MEDIUM) takes away that employee’s web browser and screen control. That computer has no screen at all. The screen where you pick the spec shows the same warning, and moving back to LARGE or above restores them. If the employee’s job involves browsing the web or driving a screen, keep it at LARGE or above.
Employees that must have a screen (such as one whose published web app visitors operate directly) can’t go below LARGE.
XL and XXL are only available if you have a paid plan or belong to an organization on the managed service.
Changes apply starting from the next activation (restart), not to the computer currently running.

The employee can also propose a spec change on its own

If the employee runs into memory trouble during a task (evidence it needs more), or notices it’s had far more capacity than it’s using for a while, it may propose a spec change in chat. Even then, the employee never changes its own spec. You’ll see an approval request showing the current/new spec and the expected cost change, and it only takes effect once you tap approve yourself.

When a downgrade proposal targets a “no screen” tier, the approval card also lists the tools the employee would lose (web browser and screen control), so you’re not approving on the SU saving alone.

Data access grade: how much of a table each employee can touch

For every table stored in your account, whether it’s personal or shared with an organization, you can set how far each employee can go with it. There are three grades:

Grade	What it means
No access	This employee can’t view or use this table at all
Read	Can view the contents, but can’t add or edit rows, or attach files to it
Read & write	Can view, and also add/edit rows and attach files

A new hire starts at “Read & write” in both personal and organization spaces. That way the employee can put its results straight into tables, and its grade won’t get in the way when you say “keep these results in a table” in team chat. The exception is the built-in system employees UpServe provides, which start at “Read.”

Organization employees that were at “Read” are raised to “Read & write” once. Employees someone deliberately lowered to “Read” are left as they are.
If an employee is still at “Read,” it is raised to “Read & write” when the organization owner (or an UpServe operator working through delegated access the owner granted) approves a table creation or extension card linked to a workflow, or one raised in team chat. The new grade applies from that employee’s next task.

This grade is about what the employee can do with a table. Whether a person can change a table’s structure is separate: any member can add a new table or column, while removing or retyping a column and turning a table off or back on are for the organization owner. See Organizing Tables in Files & Data.

To change a grade, go to the Files & data screen → the “Permissions” button next to the store selector. It lists every employee that belongs to that store, and you can change each one’s grade right there.

For a personal store, only its owner can change grades.
For a store shared with an organization, only that organization’s owner (or an operator working through delegated access) can change grades. Everyone else sees the same screen as read-only.

On the same screen, you can also toggle “Allow self-request” per employee. When it’s on, that employee can ask on its own, mid-task, for write access to a table, and a chat approval card shows up when it does. The grade only changes once you approve; declining leaves it unchanged. In an organization space, only the organization owner or an operator working through delegated access can approve this card. If another member taps approve, it’s refused and the card stays open. Turn the checkbox off and that employee won’t make these requests at all. An employee can only ask for a higher grade, and one already at “Read & write” won’t ask. Lowering a grade is something you do directly on this screen.

“Read” or above lets an employee see every column of that table, including old values. There’s no way to hide specific columns; to keep something entirely out of view, set the grade to “No access.”

9. Transferring a personal employee to an organization by drag-and-drop

The fastest way to move a personally-owned employee into an organization. In the employee table, drag a personal employee’s row onto an org section (or onto the row of another standalone employee already sitting there), and it’s transferred to that organization on the spot.

If dragging is awkward, use the menu instead. In the inbox Employees & teams list, click the employee’s ⋯ button or right-click it (in a chat, use ⋮ at the top right), then pick the organization under More → “Transfer to organization”. The same confirmation dialog opens. If you belong to several organizations, each gets its own row. System employees, managers, team leads, and employees that already belong to an organization don’t have this item.

Drop it directly onto a specific team inside the organization, and two things happen at once: the employee is transferred to the organization first, then immediately joins that team. You only see one confirmation dialog; the two steps run automatically in sequence.

This works in the web employee table and in the sidebar list on narrow screens (mobile web). It’s not available in the mobile apps yet.
The only direction supported is personal → organization. You can’t use this to pull an org-owned employee back to personal ownership, or to move it between two organizations. To send an org-owned employee back to a member, see 10. Moving an org-owned employee back to personal ownership below.

Before the transfer happens, a confirmation dialog explains:

This can’t be undone. Getting it back to personal ownership requires a separate transfer by the organization’s owner.
Every member of the organization will use this employee, but only conversation and task history from after the transfer is visible to them. Only you, who transferred it, can still see the earlier history.
Ownership moves to the organization, so it’s no longer your personal asset.

The dialog also shows:

Deleted during the transfer: the archived conversation files the employee referenced, its recall search index, and attachment copies in its workspace, so the employee can’t bring pre-transfer conversations into the organization. The original conversation records remain.
Moves along as is: workflows, canvases, and scheduled runs.
Memory list: memory moves with the employee. Check for anything the organization shouldn’t see, pick files to delete if needed, and tick “I’ve reviewed the memory” to continue.
Files to bring along: choose which files this employee made or received should move to the organization. Only the files you pick open in the organization (and no longer in your personal space); the rest stay in your personal space. Nothing is selected at first.
Tables: shows whether the tables this employee used move to the organization. When they can, you choose which ones; only the table structure moves (rows added for testing don’t), and tables that already exist in the organization under the same name aren’t overwritten. Tables that won’t move are listed beforehand.
You can’t transfer while a task or approval is pending or the employee is working. Finish or cancel it first.
Any “next steps” proposals this employee posted in a team conversation that haven’t been answered yet are closed during the transfer.

When you drop it directly onto a team, the transfer can occasionally succeed while joining that team fails. In that case the transfer has already gone through. You’ll see a message, and dragging the employee onto the same team again retries just the team join.

This only appears on accounts that use an organization (for a company or department). To move an entire team (the team and everyone on it) to an organization at once, see the “Teams on an organization account” section of Building Teams. This section covers transferring one individual employee at a time.

10. Moving an org-owned employee back to personal ownership

You can move an employee that belongs to an organization back to a specific member’s personal ownership. This is available only to the owner of the organization the employee belongs to.

Open the ”⋯” menu on the employee in your employee list.
Click “Transfer ownership (to member).”
Pick the receiving member from the list and confirm.

Before you transfer, keep in mind:

This cannot be undone. Moving it back into the organization requires a separate transfer.
Other members of the organization will no longer see this employee, and the receiving member won’t see its conversation and task history from the organization either. History starts from the transfer.
The archived conversation files the employee referenced, its recall search index, and attachment copies in its workspace are deleted during the transfer. You need to review the memory list first, and any pending task or approval must be finished beforehand.
Connection details and service connections attached to the employee move with it to the selected member’s personal ownership.
If it belonged to an org team, it’s removed from that team.
If any external services were connected through the organization’s account, you’ll need to reconnect them under a personal account after the transfer for the employee to keep using them.
The transfer goes through no matter how many employees the recipient already has, and the transferred employee arrives ready to work. There is no headcount limit.

11. Deleting an employee

On the employee page → Settings subpage → danger zone at the bottom, click “Delete”.

What goes away

All automatic runs (scheduled runs, self check-ins, webhooks, run conditions) stop immediately.
The AI work computer (code execution environment) is cleaned up.
If the employee belonged to a team, its team links are removed.
The employee disappears from your employee list.

What stays

Conversation history rows remain in the database (soft delete). There’s no UI to view or recover them yourself; they’re kept to preserve linked records (e.g., who authored which skill). The employee’s memory data, however, is permanently and immediately removed on delete and cannot be recovered.
Attachments and avatars stay, so other conversations and shared records aren’t broken.
Skills and canvases this employee created that you’ve shared remain available.

When delete is blocked

While the employee is running, delete is rejected and the screen tells you why. If the delete dialog closes and you land back on the list, the delete definitely went through. There’s no case where it fails but looks like it succeeded. Wait until the run finishes and try again.
System employees such as the Manager can’t be deleted by regular users.

After deletion, the employee no longer appears in the detail view, logs, or cost statistics pages. If you call delete again on an already-deleted employee, it returns successfully without error.

12. Rating automatic-run results on the spot

When an employee produces something on its own while you’re not watching (from a scheduled run or a webhook, for example), or right after you approve a risky action mid-conversation, a “Was this result helpful?” rating row appears right below that message (same on web, iOS, and Android).

Tap Useful and it’s saved right away.
Tap Needs fix and it’s saved right away too, opening an optional comment box below where you can note what went wrong. You can send it empty.
After rating, tap “Undo” to go back to unrated.

13. Pinning a go-to employee to the top of the list

On web, the left side of the screen shows your employees as a vertical list of avatars. If someone you talk to often keeps getting buried further down, pin them to the top.

In the list on the left, right-click the avatar of the employee you want to pin.
In the menu that opens, click “Pin to top.”

Here’s what changes once they’re pinned:

They move into a pinned group at the top of their section (your personal section or each organization’s). It sits right below the manager, and an employee who belongs to a team is pulled out of the team box and shown here instead. A short divider marks where the pinned group ends.
A small pin badge appears at the bottom left of their avatar.
If you pin several employees, their order inside the group follows the same sorting as the rest of your list.

To undo it, right-click the same avatar again and click “Unpin.” The employee goes back to their team box or original spot.

The manager always sits at the top of its section, and right-clicking it doesn’t open this menu, so it can’t be pinned.
For organization employees, the menu also shows “All org members.” Pinning an organization’s employee isn’t a personal preference: it applies to every member of that organization.

(Upcoming) Confirming the plan before building

The following isn’t deployed yet as of this update. Once it ships, this section will reflect what actually happens.

When you hand off work that produces something visible (a document, an image, code), you’ll be able to set the employee so it doesn’t dive straight into building. Instead, it lays out a short plan first and asks whether it’s okay to proceed. A new toggle will appear on the employee page → Settings subpage, and it’s especially recommended for employees that face external clients or handle outsourced work.

When it’s on, the employee presents a plan of roughly three to six steps and waits for your go-ahead before starting the actual work. If your request changes partway through, instead of quietly reworking what’s already built, it shows you the updated plan and checks in again. Simple questions and small, quick fixes skip this step entirely. If it asks about the plan and the question times out with no answer, it goes ahead with the plan it showed for work that stays within your request and can be undone, and tells you it did so. When you come back, just tell it if you want a different direction. Work that needs your approval, such as sending email, publishing, or payments, still waits for that approval, and it also holds off on deletions, new recurring schedules, and extra paid generation.

This setting works by instructing the employee to show its plan first. It isn’t a hard block. In rare cases, the employee may not follow that instruction and start working right away.

Advanced

The rest of this page isn’t required reading. Use it only if you want to look at memory structure directly, tune self check-in in detail, or check the fine print on AI work computer specs and ownership transfer.

Memory & instructions file layout

The Memory & instructions panel lets you inspect identity, guidelines, preferences, work notes, current context, and reference material. Edit only the memory entries you need to change.

These memories are private to the employee. Ask in chat to move anything teammates need into the shared team notes. Refreshing the panel wakes the work computer; leaving it open alone does not incur a charge.

How self check-in works

Self check-in works like this:

Active-hours bands: a unit of weekdays + start time + end time + interval. You can stack multiple bands.
Interval options: 30 min / 1 h / 2 h / 3 h / 4 h / 6 h / 8 h / 12 h / 24 h. This is a baseline, not a fixed cadence. It stretches when things are quiet and tightens when there’s activity.
Timezone: defaults to your timezone; editable in the card.
At least one band is always required: the editor enforces a minimum of one active-hours band, and removing all bands to disable self check-in isn’t supported. To fully disable it, use the toggle on the Automatic runs subpage instead.

During a self check-in the employee wakes with a lightweight context (smaller than a normal chat turn) and exits immediately if there’s nothing to do, so empty check-ins stay cheap. When there is nothing new to look at, a check-in can be skipped without calling the model at all, but once a check-in actually runs, even concluding “nothing to do” takes a model call that spends SU. You can check actual usage on the Logs subpage under “By run condition.”

The six computer-spec tiers in detail

Tier	vCPU	Memory	Screen
NANO	1	1GiB	None
SMALL	2	2GiB	None
MEDIUM	2	4GiB	None
LARGE	4	4GiB	Desktop
XL	4	8GiB	Desktop
XXL	8	8GiB	Desktop

Disk space is fixed regardless of tier; tier only changes vCPU and memory.
When you pick a tier directly (in Permissions on the web, in Settings on mobile), it’s saved immediately as soon as it passes server-side checks (meets the tool-required minimum, meets paid-plan eligibility), with no separate approval step.
When the employee proposes a tier change on its own, it always requires your approval. A downgrade proposal is rejected if the employee had a memory-shortage incident within the last 7 days (recent evidence of that risk is treated as more reliable than the employee’s current read on things). Either way, the change takes effect starting from the next activation.

What the tier-proposal toggle actually blocks

Turning off the tier-proposal toggle (“Propose AI tier changes” in Permissions on the web, “Receive tier change proposals” in Settings on mobile) closes three places at once: the tier-change tool itself is no longer listed or described to that employee, it can’t be executed even if something slips through, and the decision step blocks it again in case the toggle was switched off while an approval card was already showing (so an approval can’t sneak through during that window). Once it’s off, rejection history and cooldown timers for that employee also stop accumulating. Saving the employee as an employee template and sharing it carries this setting along.

What the data access grade actually controls

The grade governs UpServe’s table tools (querying, adding/editing rows, attaching files). Below “Read,” the query tool itself is refused; below “Read & write,” the tools that change values are refused.
A missing or unrecognized stored value is always treated as “Read” (unknown defaults to the lower grade, never up to “Read & write” automatically).
Both raising and lowering the grade from the Permissions screen take effect immediately, with no separate approval step (since only the store owner or the organization owner can make that change directly).
A card approval only raises employees that are at “Read.” An employee you deliberately set to “No access” stays there even if you approve the card. If the person who approves the card isn’t the organization owner or an operator working through delegated access the owner granted (or, in a personal space, its owner), the grade doesn’t change.
Table creation and extension cards linked to a workflow, and those raised in team chat, can raise the grade. Check the permission change shown on the card before approving. An extension card that only makes table fields optional does not raise the grade.
An employee linking a table to a workflow on its own does not raise its grade. Only a card approved by a person does.
An employee that worked at “Read & write” in your personal space keeps that grade when transferred to an organization. The organization owner can lower it from the Permissions screen.

What moves (and doesn’t) in a personal → organization transfer

Moves with it: connection details for connected external services, webapp links, its email address, Slack channel/thread connections, and any versions it has pinned in the Skill Library.
Team membership breaks at the same time. Unless you drop it directly onto a team inside the organization, it lands as a standalone employee in that organization right after the transfer.
Doesn’t move: External tools (MCP) connections registered on your personal account (moving these would drag along your other personal employees that share the same server), auto-assignment of external services connected with the organization’s account (an organization owner has to assign these separately), attribution for past SU usage (the ledger is never rewritten retroactively), and authorship credit for any skill or template this employee made.
The destination must be an organization you’re already a member of. This action can’t create a new organization for you. System employees, public-version employees, and employees that already belong to an organization can’t be transferred this way.
No special permission is needed: you’re moving your own personal employee into an organization you already belong to, so the regular Member role is enough. Sending an org-owned employee back to a specific member, on the other hand, is a separate action only that organization’s owner can do.

What’s preserved during a reverse ownership transfer

Organization membership itself doesn’t change. The receiving member stays in that organization after the transfer.
Past usage history (SU-spend records) is never retroactively re-attributed.
Conversation history, canvases, and other content are preserved; other members of the organization just lose visibility into them.

Data retention after delete (what we can confirm)

Employee record: marked status = DELETED with a deleted_at timestamp. Hidden from lists and search right away.
AI work computer (code execution environment): torn down immediately.
Automatic runs: scheduled runs, run conditions, and webhook callbacks are all disabled in one go.
Conversation, canvas, and task rows: preserved. For permanent removal of your content rows, contact support.

The above is what the code (AgentService.delete) actually does today. If that changes, this page is updated alongside it.

The fast-hire Work setup window

A new hire created through fast hire keeps its self-editing profile tools until you confirm (approve the summary it shows you). 🔴 There is no time cap (UPS-1072, removed 2026-08-25). Previously the window auto-closed 72 hours after hiring, but now an unfinished Work setup can sit for days without the tools being revoked or automatic runs turning on.

The hire-time-plus-72-hours value (rookie_until) is still stamped by the server, but it’s display-only now. Once that time passes, the progress card switches to “Stalled” to flag the neglect, and nothing else is gated on it. Only your confirmation revokes the tools; from then on the profile can only be changed the way any other employee’s can, through the Settings subpage.

When the rating row appears

There are two cases where the rating row shows up.

Unattended-run output: deliverables from scheduled runs, webhooks, work reports, or tasks the employee proposed and carried out on its own (initiative). It doesn’t appear for self check-ins or employee-to-employee mentions.
Approved actions mid-conversation: if a risky action asks for your approval during a live chat and you approve it, the resulting output also gets a rating row, because approval only judges the action beforehand, so it can’t speak to how the result actually turned out. If you decline approval, nothing runs, so there’s nothing to rate.

Ratings are only used in aggregate. The employee doesn’t see its own score or react to it in a later conversation.

(Upcoming) How the plan-confirmation policy works

Technical detail for the “Confirming the plan before building” section above. Not yet deployed to production as of this update.

This setting isn’t a system-enforced approval gate. It’s a single instruction line added to what the employee sees. Whether it actually presents a plan is ultimately up to the employee’s own judgment.
It only applies to normal conversation turns responding to a message you sent directly. It’s not added during unattended runs (scheduled runs, webhooks, or self check-ins).
If the plan question expires unanswered, the instruction has the employee treat your original request as the go-ahead for reversible work, follow the plan it showed, and tell you so. Actions that need an approval card (sending, publishing, payments, and so on) are still held by the approval policy regardless of this instruction.
The toggle is per-employee and defaults to off.

Managing Employees on the Home Screen: see every employee at a glance and act on what needs attention
Building Teams: group employees into a team so they can collaborate