Skip to content
Console
Try one

    Analytics

    Read Thalovant usage analytics, demand signals, geography, client posture, and hub activity.

    Analytics helps you understand how the workspace is being used over a selected time range.

    Use it to review demand signals, compare the current window with the previous one, find busy time buckets, inspect hub and client activity, and notice when missing telemetry limits a diagnosis. Use Analytics when the question is about usage volume, geography, trends, or which hub or client deserves the next look.

    Start with 24h for a normal review, 1h for a live issue, and 7d or 30d for trends. The Top 15 explorer ranks utterances, messages, intents, clients, hubs, and countries; select a row and the whole page filters to it, so every panel answers the same question. A workspace with no telemetry yet shows nothing until a connection exists and its client has sent a first heartbeat or message; add one, run it, then refresh. If Analytics isn’t available at all, that is a plan boundary rather than a fault: usage analytics need workspace hub capacity, and Billing says whether your plan has it. When you ask for help, capture the range, the active filters, any data-quality warning, the top signal, and the client or hub detail you were looking at.

    Workspace signal

    The headline for the range, the finding under it, and the next actions: open the top client, review clients, or clear filters.

    Current vs previous

    Compares the selected window with the same prior window.

    Telemetry completeness

    Shows missing metadata that can limit map, intent, client, or hub diagnosis.

    Demand signals

    Lists the top utterances, messages, intents, clients, hubs, and countries.

    Saved views

    Jumps to common views such as Live ops, Country focus, Stability, and Noisy clients.

    When usage happens

    Activity over time, and a day-by-hour heatmap of which hours carry the load.

    Posture segments

    Groups clients by connected, idle, reconnect-prone, noisy, quiet, or disconnected behavior.

    Where usage comes from

    Shows country-level activity when telemetry includes country metadata.

    New Private or Team workspaces start with an activation state instead of empty charts.

    1. Add a connection. Pair Thalovant Voice or another supported client.
    2. Keep the client running. Analytics starts after the first heartbeat, message, utterance, connect, or disconnect event arrives.
    3. Refresh Analytics. Start with 24h, then use shorter or longer ranges once signal is flowing.

    When the workspace has hubs or clients but no traffic in the selected range, Analytics shows a quiet telemetry brief. Use it to confirm whether clients are live or idle before opening detailed panels.

    The Analytics page before any telemetry: five counters at 0, a Workspace signal card reading Waiting for analytics signal, range buttons from 1h to 30d and an empty Attention queue.The Analytics page before any telemetry: five counters at 0, a Workspace signal card reading Waiting for analytics signal, range buttons from 1h to 30d and an empty Attention queue.
    A quiet telemetry brief. The counters stay at 0 and the attention queue stays empty until the first client event arrives.

    Thalovant records a small set of account-setup milestones so the product team can find where setup fails and how long first value takes. These milestones cover account creation, email verification, sign-in, the chosen onboarding path, accepted and Ready resources, and a successful preview or memory action.

    Activation milestone metadata is deliberately narrow. It can identify the milestone, setup intent, resource kind, operation kind, or product surface. It does not store prompt text, replies, memory content, email addresses, passwords, tokens, credentials, setup links, or secret material. Product reviews use aggregate counts and latency percentiles; the workspace Analytics page doesn’t expose a person-by-person activation ledger.

    Self-service account deletion erases product telemetry linked to the account before the account row is removed. Thalovant retains one unlinkable deletion audit marker so the deletion operation itself remains accountable without retaining the deleted account ID or profile snapshot.

    Free workspaces stay on the simpler voice path. Direct Analytics links route back to Dashboard, and the Free sidebar focuses on Dashboard, Public hubs, Connections, Billing, Profile, and Help.

    When that happens:

    1. Open Billing. Confirm whether the current plan includes owned hub capacity.
    2. Use Public hubs while deciding. Free keeps public hub discovery and one voice connection open.
    3. Return to Analytics after Billing allows it. Add or run a client, then start with the 24h range.

    Don’t treat a missing Analytics route as a broken chart. It’s a plan boundary until Billing says the workspace includes the needed capacity.

    1. Choose a range. Start with 24h; use 1h for live incidents, 7d or 30d for trends.
    2. Read the brief. It tells you whether activity, quality, or client posture needs attention first.
    3. Check data quality. If metadata is missing, avoid treating the map or top list as complete.
    4. Use Demand signals. Switch between utterances, messages, intents, clients, hubs, and countries.
    5. Filter into detail. Select a hub, client, country, message, utterance, intent, health segment, or time bucket.
    PanelWhat it answersWhat to do next
    Attention queueWhat matters most in the selected range.Use the suggested action or clear filters.
    Saved viewsWhich common lens to apply.Choose Live ops, Country focus, Stability, or Noisy clients.
    Current vs previousWhether usage rose or fell from the prior matching window.Widen the range if there is no baseline.
    Telemetry completenessWhich missing fields limit the read.Capture the warning before drawing conclusions.
    When usage happensUtterances, messages, connects, and disconnects over time.Select a peak bucket when a spike needs review.
    Which hours carry the loadA day-by-hour heatmap of the selected range.Select the peak to filter the page to that hour.
    Signal compositionHow the activity splits across speech, messages, connects, and disconnects.Select a segment to focus the explorer.
    Where usage comes fromThe busiest countries by requests, when country signal exists.Select a country to filter the page.
    Usage by hubWhich hubs carry the most volume.Open the hub detail or filter the hub.
    Connected surfaceHub volume, active clients, idle clients, and last seen time.Open the hub row when a hub needs inspection.
    Posture segmentsConnected, idle, disconnected, flapping, noisy, and quiet clients.Select a segment to filter the client ledger.
    Clients worth checkingClients with reconnect churn, stale presence, unusual volume, or missing country signal.Open the client detail before changing setup.
    Client usage ledgerClient-level utterances, messages, country, state, and age.Open a client detail or filter that client.

    Saved views reset the page to a useful starting point:

    ViewUse it whenWhat it changes
    Live opsSomething is happening now.Uses 1h and starts with message signals.
    Country focusYou need geography.Uses 7d and opens country signals.
    StabilityYou are checking reconnects or disconnects.Uses 24h and message signals.
    Noisy clientsOne client may be too active.Uses 24h, filters noisy clients, and opens client signals.

    Active filters appear as chips under Filtered by. Select one chip to remove that filter, or use Clear all to return to the full selected range.

    The Top 15 explorer shows five ranked rows at a time and keeps the rank stable as you page through results.

    Use the tabs to switch the question you’re asking:

    • Utterances: what people are saying most often;
    • Messages: which bus messages or message shapes are busiest;
    • Intents: which intents appear most often;
    • Clients: which clients are carrying the most activity;
    • Hubs: which hubs are carrying the most activity;
    • Countries: where activity is concentrated when country data is available.

    Selecting a top row applies the matching filter. That lets the rest of the page answer the same question without making you rebuild the search.

    Use the search box inside the explorer when the Top 15 list is long enough that you already know the signal name. Search filters the visible top signals; it doesn’t change the selected time range.

    Explorer actionResult
    Switch tabsChanges the ranked signal type.
    Search top signalsNarrows the current tab by label.
    Select an utterance, message, or intentApplies that text filter.
    Select a client or hubFilters to that client or hub.
    Select a countryFilters to that country.
    Use Previous or NextMoves through the ranked results five at a time.

    Client detail shows utterances, messages, connects, disconnects, posture, geography, and last seen time. From there you can filter the selected client or its hub.

    Hub detail shows active and idle clients, recent usage, and the clients attached to that hub. Use it when the top list names a hub but the next check belongs to a specific client.

    Open details before changing setup. Analytics is strongest when it helps you choose the next owner page: Clients for connection state, Access rules for access, Hubs for hub setup, Topology for live presence, or Billing for plan limits.

    Analytics can still be useful when some enrichment is missing.

    What you seeWhat it means
    Country enrichment pendingActivity is still tracked by time, hub, client, and message, but map detail is incomplete.
    No country signal yetTelemetry has not included country metadata for the selected range.
    No activity in this rangeTry a wider range, then confirm Dashboard and Topology are seeing fresh activity.
    Telemetry completeness warningCapture the missing field before opening support.

    For an Analytics question, include:

    • selected range;
    • active filters;
    • the exact locked, unavailable, or data-quality message;
    • the Top 15 explorer tab and row you selected;
    • client detail or hub detail if one was open;
    • whether Dashboard and Topology show the same story.

    Analytics is clear when:

    • the selected range matches the question;
    • active filters are visible and intentional;
    • the top signal or busiest time bucket is understood;
    • data-quality limits are captured before drawing conclusions;
    • you know whether the next page is Dashboard, Hubs, Clients, Topology, or Support.