Administer

Workspace settings & operations

The Administration area (/admin) unifies tenant configuration, navigation management, consumption monitoring, security keys, support escalations, and subscription billing.

Walkthrough: find your way around Administration

The rail groups every admin section. This is the one place to start from.

Choose Administration in the sidebar. There is one Administration link; everything administrative sits behind it.
Administration opens on the first section you may use. The rail on the left groups the sections: Organization, Access, Workspace, Billing and Operations.
Choose a section in the rail. The band above the page reads "Administration / Audit log", so you always know where you are.
A section that does not exist, or that your role cannot use, shows a way back inside Administration: choose "Back to Settings".
On a narrow screen the rail becomes the "Administration section" menu at the top of the page.

Your own settings

Every person has personal settings: profile, appearance, notification preferences, API keys and security, and Memory: what the AI keeps about them (Memory). They are separate from the workspace settings below.

Open Settings from the account row at the bottom of the sidebar. It opens on Profile, with the other sections in the list on the left: Appearance, Notifications, API keys and Security.
Change your Display name, then choose Save profile. Your email is your sign-in and is managed by your administrator.

Notification preferences

Saved delivery preferences are listed with a switch each.

Open Settings, then Notifications. Each kind of notification (Alerts, Shared with you, Mentions, High usage) has a switch and a delivery channel. Everything is on, in the app, until you change it.
Choose where a kind is delivered: In the app or Email. Your choice is saved at once and decides how you are notified; it applies to you only.
Turn a switch off to stop that kind of notification altogether.

General settings

Under Admin → Settings (/admin/settings):

Workspace name: The display name of the workspace shown across navigation and shared assets.
Default timezone: Full IANA timezone selection (e.g. America/New_York, UTC, Asia/Tokyo) used for report scheduling, cron evaluations, and timestamp formatting. Commentary the AI writes from computed numbers counts a month (or any period) as over once it has ended both here and in UTC.
Date format: Workspace-wide default display format (YYYY-MM-DD, MM/DD/YYYY, DD/MM/YYYY, DD.MM.YYYY).
Single sign-on (SSO): Configure SAML 2.0 / OIDC enterprise identity providers (Enterprise plan). See SSO & MFA.
Slack integration: Connect workspace Slack channels for automated report broadcasts and metric alert notifications.
Custom themes: Manage custom UI and chart color palettes. See Themes.
AI: every AI setting lives on its own page, Administration → AI; see AI below.
Watermark shared images: A company admin can require that every chart or dashboard image shared from the mobile app carries a visible watermark — who shared it, when, and from which workspace. Off by default; set it with PUT /api/workspaces/{id} (requireShareWatermark).

Frozen stories and presentations

Under Admin → Workspaces, Edit on a workspace has the switch Frozen stories and presentations, off by default. While it is on, editors may freeze a story’s data when they publish it (Frozen stories), and a published presentation’s values for a meeting. A copy of the frozen values is kept with each; readers still need access to the data to see them, and a presentation’s frozen values are shown only to readers whose access is the same as the person who froze it. Turning the switch off deletes every frozen value in the workspace and returns those stories and presentations to live data; records of what a presentation showed are kept. Set it with PUT /api/workspaces/{id} (allowStorySnapshots, its name unchanged).

AI

Under Administration → AI (/admin/settings/ai, company administrators), one page holds every AI setting of the workspace:

AI in this workspace and Who may use AI: whether AI runs, in which AI features, for which roles, and the monthly AI credit budgets. See Control AI in a workspace.
AI assistant: Answer depth sets this workspace’s default effort, the one a conversation runs at until its person chooses one: Quick, Balanced, Deep or Max. Left at Company default, it is the company’s default effort (Set the company’s effort): Quick on Team and Balanced on Business and Enterprise unless a company admin chose another. Longest answer caps how long a reply may be, and Keep conversations sets how long chats are kept. Choose Save AI settings to apply.
Verified answers: questions you have checked, each answered by one governed metric or one dashboard card built on a data model, for the period it asks about. When someone asks one of them, in its words, in one of the other ways people ask it, or in other words that ask exactly the same for the same period and the same products, customers or places, the AI runs that metric or card again on fresh data, as the person asking (their row-level security applies), and answers in a few seconds, marked Verified answer; the numbers are never stored. To add one, type the question, the other ways people ask it (one per line), choose the metric, or the dashboard and its card, and the Period it asks about, then Add verified answer. A metric is evaluated over that period; a card runs as it is saved, so the period must be the one its own filters show. Leave the period at None for a question that names no period. When the question or one of its other ways names another period (a fiscal year, a weekend or since last year, say), or the metric or card cannot answer for it, the answer is not added and the message says what to change. Remove takes one away. One whose metric or dashboard was deleted is marked No longer runs, and the AI answers that question as any other; so does one whose metric or card has since been changed to cover another period.
AI tools: what the assistant may do for your people, such as running a query or building a chart. Switch a tool off to stop the assistant using it in this workspace; it applies within a minute, in conversations already under way too, questions answered at once included. Switching off Writing the plan turns plans off: no checklists, and a large build goes straight to its approval card (plans). On a server with the new tool set, the list also holds the new tools, each saying which of the tools above it takes over: a new tool is used while its own switch and one of those are on, and a part of it is refused only while every tool above that did that part is off (SQL of the assistant’s own while the query tool is off, say; a part two of them did stays on while either is). One that takes over none, such as Memory, is used while its own switch is on: switching Memory off stops the AI looking in or writing to its memory from a chat (switch the memory off).
AI brief: what every chat in the workspace should know about the business, read whole by every chat. Write it as a draft, try it in a chat, publish it; every published version stays in its history (the workspace AI brief).
Shared memory: what people propose or correct for everyone waits here until you accept it, as it is or edited, or dismiss it, and the notes that apply to everyone are listed, written and retired here (how a note reaches other people). People’s own notes are never shown.
What people said about AI answers: the thumbs-up and thumbs-down ratings; Period changes the window.
AI ledger: every AI call of the month, with the AI credits it used.

The assistant, its tools and the feedback card used to be under General.

Under Administration → Appearance, Design defaults for what the AI builds sets how the dashboards the AI builds look when a person set nothing of their own, and Use this theme for everything the AI builds applies the workspace’s theme to all of them (design preferences).

Walkthrough: workspace settings, AI assistant and AI tools

Set the workspace name and defaults on General, then, on AI, tune the AI assistant, switch AI tools on or off and read the feedback people left.

Open Admin, then General. Set the workspace name, and the default time zone and date format that every page uses.
Choose Save Settings at the foot of the page. A message confirms the change; nothing is applied until you save.
Open Admin, then AI. AI assistant: Answer depth sets how carefully the assistant works, Longest answer caps how long a reply may be, and Keep conversations sets how long chats are kept. Choose Save AI settings to apply.
AI tools lists what the assistant may do for your people, such as running a query or building a chart. Switch a tool off to stop the assistant using it in this workspace.
Below them, "What people said about AI answers" summarises the thumbs-up and thumbs-down ratings. Use Period to change the window.

Menu builder (`/admin/menus`)

The Menu Builder (/admin/menus) provides full customization of the sidebar navigation tree:

Hierarchy & Types: Create section headers, direct route links, dashboard shortcuts, paginated report links, and external URLs.
Reordering & Nesting: Move entries up/down and create nested submenus.
Role-Based Audience Grants: Control which roles can see specific menu items.
Per-Role Label Overrides: Customize navigation labels per role (e.g. renaming “Warehouse” to “Enterprise Data Lake” for viewer personas).
Viewer Preset: One-click configuration creating a streamlined view-only navigation tree (Home, Dashboards, Reports).
System Locks: Core system entries are protected against accidental deletion while remaining reorderable and renamable.

On the hosted service

On the hosted service the menu is shared by every company, so the menu tree is read-only for a company admin: entries can’t be added, renamed, moved, hidden or deleted, and no external link can be added. The page says so above the tree. What stays yours:

Labels for your company’s own roles: pick one of them under Audience, relabel entries, then choose Save role labels.
Which entries your company’s own roles see (Audience, Business plan). Ticking an entry changes the role, and changing a custom role needs Business; on Team, relabelling is what stays yours. Compare plans

The built-in roles, such as Admin, Editor and Viewer, are shared by every company too, so their labels can’t be changed there, and another company’s role can’t be read or relabelled. The api answers each refusal with 403.

On a private server, which is one company, a company admin keeps the whole Menu Builder, the built-in roles’ labels included.

Each person’s own sidebar choices (hide, pin, order) are theirs on either.

API keys (`/settings/api-keys`)

API keys are personal, so they live under Settings → API keys rather than in Administration. The old /admin/api-keys address redirects there. Manage programmatic access to the DataSquares REST API:

Scoped minting: Generate API keys scoped to Read, Read & Write, or Read, Write & Admin. A key never does more than your own role allows.
No AI: None of the three lets a key use AI features, which are available in the DataSquares app only. A key that calls one is refused; see Scopes for what a key gets instead.
One-time display: Full tokens are displayed once upon generation. Only key prefixes remain visible in the management table.
Lifecycle management: Deactivate, reactivate, or permanently revoke API keys. View last-used timestamps and request activity.

Walkthrough: create a personal API key

Open Settings, then API keys. Your keys are listed here. A key lets a script or another tool act as you, with the scopes you give it.
Choose Create API Key. Name the key for what will use it, and choose its Scopes: Read, Read & Write, or Read, Write & Admin. Give it no more than it needs.
Choose Create. The key is shown once. Copy it now and keep it somewhere safe, because DataSquares cannot show it again. Then choose Done.
The key is now in the list, with its Key Prefix, Last Used and Status. Select a key to deactivate or delete it. A deleted key stops working at once.

Integrations (`/admin/settings/integrations`)

The Slack bot answers governed-metric questions in a channel with a slash command. Its answers leave DataSquares, so, as for an activation, a metric that reads a sensitivity-labelled field is not answered, whatever the role of the admin who connected it, and a metric that can’t be read says only why it was refused (row-level security, a labelled field, its definition) or that it could not be read.

Every command reads as the admin who connected Slack, and that admin is checked again on each command: if they are no longer an admin of the workspace’s company, the command is refused with a pointer to Admin → Settings → Integrations, the workspace’s admins are told, and an admin connects Slack again to resume it.

Walkthrough: connect Slack

Open Admin, then Integrations. The Slack bot answers governed-metric questions in a channel with a /datasquares slash command. Create a Slack app with a slash command that points at the address shown, then bring its details here.
Enter the Slack team ID and the signing secret from the Slack app's Basic Information page (the secret is masked), then choose Connect Slack. Answers run with the asking person's own data permissions. Nothing is saved in this walkthrough.

Usage & capacity (`/admin/usage`)

The Usage tab shows how the workspace uses what it pays for:

Pipelines (with a Pipelines size): execution hours used this month and what is left — the size’s monthly hours plus any hour packs, each pack with what is left of it and when it ends — and the runs it can have at once (guaranteed, and up to the burst when slots are free), with what is running and waiting now.
Warehouse (with a Warehouse size): storage now against its limit (the size plus any storage add-ons) and query hours this month against the size’s plus any query-hour packs, each pack with what is left of it and when it ends.
Past the allowance: this month’s use past the allowances so far, which goes on the next invoice (see Seats & billing).
Capacity changes: each hour pack, add-on or agreement added or removed, with who and when.
More hours, More query hours, More refresh hours, More storage and Bigger size, for billing administrators who may change the plan: each opens the Pipelines & Warehouse sheet on Billing at that part, where hour packs, sizes, storage and the rates past the allowance are bought. Anyone else who can open the page sees the numbers.
AI credits: the AI credits used and the number of AI requests over the last 30 days, the credits used on each of the last seven days, and which features used them — the whole company’s for company administrators, this workspace’s for anyone else the page admits — and, for billing administrators, the included credits used this period, the balance (bought credits below zero read as such), the order credits are used in, the company’s below zero and what the next invoice will carry, and the credit activity. Credits are shown in whole numbers; a use smaller than half a credit reads “under 1”. Platform administrators also see each tenant’s monthly totals.

Walkthrough: see how the workspace is used

The Usage page shows capacity first, then the AI credits used.

Open Admin, then Usage. Pipelines and Warehouse come first: their hours, storage and the runs they can have at once, with Buy more for billing administrators. Further down, the AI credits used this period and in the last 30 days.
Daily Usage charts the credits used on each of the last seven days, and Feature Breakdown shows which features used them. A new workspace shows "No usage data yet" until people start asking questions.
Further down, the plan allowance and Credit activity list what was granted and what was spent. Use Rollup month, where offered, to look at another month.

Experience baseline (`/admin/experience`)

The Experience tab shows workspace admins aggregate counts from typed product events over a date window (the last 14 days by default). It shows no raw events and nothing per person.

Readiness — whether the window holds enough days with events (14) to count as a baseline. A shorter window is labelled provisional, however good its numbers look.
Journey rates — Home to a dashboard, command palette to an action, a signal to an action, draft recovery, onboarding steps, viewer to share and to edit, and a first relevant signal within a minute. Each rate shows its numerator and denominator; an empty denominator reads “No data”, never 0%.
Mobile app outcomes — what people did in the DataSquares mobile app, grouped by the job a phone does: Are we OK? (a change opened from Home, the morning brief), Something fired (an alert opened from a notification, Home or the Inbox), Before the meeting (a saved view applied, a snapshot or link shared), Answer my boss (an explanation, an analysis, Ask AI from a dashboard or metric) and Did the data land?. Each outcome is the share of people who opened the app in the window, with people and times as evidence. A job the app does not count yet says so.
What these numbers do not say — the server’s own limitations for the window.

Walkthrough: read the experience baseline

Journey rates and the evidence behind them.

Open Admin, then Experience. This page counts typed product events for the workspace: no raw events and no per-person tracking. A banner says when there is not yet enough traffic to call the numbers a baseline.
Journey rates show how often people complete each path, such as opening a dashboard from Home. The Evidence column shows the counts behind every rate.
Change From and To to move the window, then choose Refresh to recompute the rates for it.

AI escalations (`/admin/escalations`)

The AI assistant in AI chat opens an escalation when someone asks it for a person, when it has tried and can’t resolve the request, or when the request needs authorization it doesn’t have. It tells the person that the people of your company who handle such requests see it in their queue and can answer them in the conversation, and the escalation lands in the AI Escalations queue:

Each one shows who asked (their name and email), the priority the assistant gave it, the reason it wrote, when it was opened, who claimed or resolved it, and the answers already given. The conversation itself stays private to the person who asked.
Answer writes in their conversation: your answer appears there under your name, as from a person of your company and never as the AI’s, and they’re told in the app (their Mentions and replies notification setting). The AI doesn’t read it. Answering an escalation nobody has claimed makes it yours.
Claim an escalation to make it yours, then resolve it with a note. Awaiting lists the ones nobody has claimed.
The AI escalations permission decides it all: viewing it shows the queue, and its update right lets you claim, answer and resolve. Company administrators hold it, and so does a role given it.

Billing (`/admin/billing`)

Billing is one page (the Billing page): a bar only when something needs doing, a summary of the Plan, the Next payment and the AI credits, the latest invoices, and Details rows that each open a sheet over the page:

Change plan: the plan, its billing period and the editors and viewers, reviewed with the price and the date it starts before you pay on Razorpay or schedule it.
Renewal, Pay automatically and Billed to: when the plan renews, paying invoices automatically, and who invoices are made out to, with the GSTIN in India.
AI credits: the company’s credits with each workspace’s own, buying credits, below zero, auto top-up, viewers’ own credits and the credit activity (see AI credits).
Pipelines & Warehouse: the add-ons, hour packs, extra use and the Platform bundle.

All invoices is a page of its own, with the invoices to pay, paid and void, their PDFs, and the account’s activity.

Users, roles & invites — member permissions and custom roles.
SSO & MFA — SAML and OIDC single sign-on, SCIM, and two-factor authentication.
Compute & workers — managing VM execution fleets and worker queues.
License — offline license verification and feature entitlements.
Audit log — tamper-evident administrative audit logging.