FAQ

Why don’t I see a create button for automations, workflows, notes, or secrets?

There isn’t one — that’s by design. These resources are created by asking Ronja in an exploration (“Schedule a daily revenue summary every morning at 9 AM”). The pages you see in the app are for viewing and managing what Ronja built. See The resource types.

Why can’t my teammate run analyses?

They most likely have the User Read-Only role — the default for new invitees — and your organization has turned off Read-only AI access. Read-only members can ask Ronja questions, run analyses, attach files and have Ronja make documents and files into their own My Files in their own chats, or in a chat shared with them with Full access, while that setting is on, which it is by default; with it off they can only view shared content. Either way they can’t create or change anything shared. An Admin can turn the setting back on, or upgrade their role, under Users & Access. See the roles capability matrix.

Where did my exploration go?

Brand-new explorations don’t appear in the sidebar until they have a title, which Ronja writes automatically from your first message. The Chats and Shared sections initially show six recent conversations. Click Show more to expand a section, or Show less to shorten it again. Show less keeps your current conversation in the list if it was included in the expanded section. The sidebar loads up to 50 recent explorations; use the magnifier on the Chats section header to search beyond that list.

Who can see my private feature?

You and your organization’s admins. A private feature is invisible to your colleagues until it is shared with the organization: you hand it over and an admin approves, or an admin shares it for you, and you get a notification saying so (unless the feature was in the Trash). Admins see everything in the organization — every member’s private features included — and can open members’ explorations for support. See Scopes and sharing.

My Private Feature is the exception: it is your personal one, where a Saved Agent you create without choosing a feature lands, along with every copy you take with Make my own and every SharePoint key you create, and it can’t be handed over at all. To share something in it, create a feature and move that thing into it.

I promoted a feature to Organization scope — why can’t anyone see it?

Organization scope makes a feature available, not automatically visible. An admin still has to attach it to access groups using the access matrix on Shared features — or an access group needs All features turned on. See the scopes reference.

Why can’t I pick the AI model or effort in chat?

The model/effort picker is open to everyone by default, but an admin can turn Choose model & effort off for your role or for you on the Control Center’s Rules tab. Super Admins always have it and User Read-Only members never do. It’s also hidden for everyone while your organization is out of credits, and comes back when credits are added. The reduced-models step does not hide it. See Govern AI spend.

Why did AI pause for the day?

This is a safeguard, not a billing problem. Your organization’s daily AI safety limit kicked in: it caps how much AI the whole organization uses in one day, so a runaway agent or a loop can’t cause an unexpected spike. A grey banner above the chat box says “AI is paused for now.” and when AI resumes, in your own time — at the latest when the day resets at 00:00 UTC, or sooner if an admin raises the limit. Nothing is wrong with your account or billing, and nothing you typed is lost: a message you sent stays in the chat as Not sent with Retry. Workflows that don’t use AI keep running. Admins can see today’s use and change the organization’s own limit on the Control Center’s Rules tab. See Daily AI safety limit.

Why did new chats and runs stop for my organization?

Your organization has used its monthly budget: the limit on how much the whole organization uses in a billing period, which some organizations Ronja bills by invoice have. A note in the sidebar says “Your organization has used its monthly budget.”, and a message you send is refused with “…New chats and runs resume when the next billing period starts on {date}, or as soon as an admin raises the budget in Control Center › Rules.” What you typed stays in the chat box. It stops everyone, admins included: new chats, Saved Agent runs, automations and workflow runs. Work already running finishes, and runs waiting on someone carry on once the budget allows. Any Admin can raise the budget on the Control Center’s Rules tab, which lifts the stop within about 20 seconds; otherwise it lifts when the next billing period starts. Admins are emailed at 80% and again when it is used up.

Why was my prompt refused with a message about credits?

Five different gates can refuse a prompt, and four of them are spending caps: your billing-period allocation is used up (“You have reached your credit allocation for this billing period…”), your weekly AI budget is used up (“You’ve used your weekly AI budget. It resets on Monday at 00:00 UTC…” — it resets then on its own, or an admin can raise it, or give you extra credits for the rest of the week), your allocation was set to Blocked by an admin, or the organization’s wallet is empty (“Your organization has used all its available credits. Top up your wallet to resume runs.”). The same checks apply when you answer or skip a form, or approve a step, in a chat, so once your weekly budget is used up those are refused too; Reject still works and ends the step without Ronja continuing. The message tells you which — contact your admin, or for an empty wallet, your Super Admin can buy credits. See Credits and AI spend. The fifth gate never mentions credits: the organization’s daily AI safety limit pauses AI for everyone with a grey banner beginning “AI is paused for now.” — see Why did AI pause for the day? An organization on a monthly budget has no wallet: its message says it has used its monthly budget instead — see Why did new chats and runs stop for my organization?

Does the reduced-model switch apply to weekly budgets?

No. Your organization’s spend steps, for example an email at 70%, a warning to members at 85% and reduced models at 95%, measure the whole organization’s credits for the billing period (monthly). They never apply to a person’s weekly budget. A weekly budget has no percentage steps: at 75% only that person sees a warning above their chat, and at 100% their chats stop until Monday 00:00 UTC. They can check how much is left at any time in the account menu. Neither a weekly budget nor a billing-period allocation ever switches anyone to a different model. The Control Center’s Rules tab has a table of what each limit does.

Why don’t I see the Build option on the start screen?

Build mode is on for everyone by default, and an admin can turn it off per role or per person on the Control Center’s Rules tab. User Read-Only members never have it. If it’s off for you, the Build segment simply isn’t offered — ask an admin to turn it on. Deep analysis doesn’t need it. See Standard, Deep analysis, and Build.

Under an answer, what’s the difference between the chips and the dark buttons?

Where the chip takes you. The chips under Continue here ask the next question in the conversation you’re in — clicking one sends it as your next message, word for word as the chip reads, and nothing else happens. The dark buttons with a branch icon (“Explore this in a separate thread.”) fork a tangent: a private branch on that question, leaving this thread untouched. Ronja picks the shape deliberately — the obvious next step of an analysis belongs in the thread that produced it; a genuine side-angle is worth a branch.

Can I change an exploration’s mode after it starts?

No — the mode is fixed once the first message is sent. The one exception: Ronja herself can request to enter Build mode mid-conversation, which you approve in chat (and it’s one-way for the session). To use a different mode, start a new exploration or fork a tangent in the mode you want.

I deleted something by accident — can I get it back?

Almost certainly. Deleted items go to the Trash for 30 days, where an admin can restore them. See Restore from trash.

Why can’t I edit a note or the Knowledge page directly?

Notes are written and edited through chat — ask Ronja to change them, and she writes the note and opens it beside the chat. Organization Knowledge is the same idea, deliberately: “tell, don’t edit”. An admin clicks Curate in chat and tells Ronja what is wrong; she rewrites or retires the entry. She no longer writes new Knowledge entries — new org-wide rules go in the organization policy. See Curate Knowledge.

Your policies are the exception: Your policy and the Organization policy are short documents you type into and save yourself. See How Ronja learns.

Why did my table edit not go live?

Edits to shared tables become a personal draft first. The live table keeps serving everyone until an Admin reviews and commits your draft — the banner shows “Draft pending review”. See Versions, drafts, and approvals.

Why isn’t Ronja answering in our Slack channel?

First, the channel has to be bridged — no channel is connected automatically. Add Ronja with /invite @Ronja or Add bridge on the Bridges page (if Admin-only connections is on, a Ronja admin has to do it). Once it’s bridged, check three things: the channel’s trigger setting (by default Ronja only answers when @-mentioned), whether your Slack account is linked (Ronja auto-links by email on your first message — “You don’t have a Ronja account yet” means you need an account first), and the bridge’s status on the Bridges page (a disconnected or reauthorization-needed bridge stops replying). See Use Ronja in Slack and Teams and Troubleshooting.

Why does my uploaded file show a “truncated” badge?

Text files can be up to 5 MB, but only roughly the first 50 KB is passed to Ronja to fit her context budget. For big files, upload data as CSV/Excel instead — data files become tables and aren’t truncated.

What’s the difference between the Usage page and the Billing page?

Usage shows where credits went; Billing is the Super-Admin page where credits are bought. Full answer: What’s the difference?

Why can’t I approve my own request?

That’s a deliberate guardrail. The Inbox withholds Approve on a proposal you filed yourself in the browser, and your organization may also disallow self-approval as a policy (set on the Organization rules page) — in which case a different admin reviews anything you authored, whatever credential you filed it with or try to approve it with. The policy is the rule that always holds; the hidden button is a convenience, and it can’t recognise a proposal you filed with an API token. See Propose and review changes.

Why did my automation stop running?

The most common causes: it was paused (the on/off toggle — deleting its access group also pauses it), its run cap was reached (webhook, table and mailbox triggers all have daily caps), or the organization’s credits ran out (an empty wallet blocks automation runs too), or its monthly budget is used up: each fire then shows in its run history as Paused — monthly budget used, which is not a failure, sends no failure alert and is not run again later. Run now starts it again once the budget allows: an admin has raised it, or the next billing period has started. An automation that runs a Saved Agent, or a workflow that asks AI something, also doesn’t get its work done while the organization’s daily AI safety limit has paused AI: each fire during the pause shows in its run history as Paused — daily AI safety limit and is not run again later, and what an email or webhook carried is not kept. Run now starts it again without it once AI resumes; for an email or webhook that mattered, have it sent again, or ask Ronja to handle it. A Mailbox automation has two causes of its own: disconnecting its mailbox switches it off, and a mailbox that merely stops working — one that needs reconnecting — stops it firing while it still reads Active. Open the automation’s run history to see what happened. See Troubleshooting.

How many favorites can I pin?

Ten per type in each Space, counted in the Space the chat or app is in, even when you star it from another Space. Adding an eleventh there shows “Max 10 favorites in your Space — unfavorite one there to add another”, naming that Space. When it is in a Space you can’t open, the message says “Max 10 favorites outside the Spaces you can open” instead.