ThoughtSpot’s AI products are licensed through three mechanisms: AgentSpot, a separate AI work assistant with its own credit; Spotter, an AI analyst that is part of the Analytics plans and is also bundled with AgentSpot; and LLM access, which ThoughtSpot says it does not meter. The AgentSpot documentation describes AgentSpot as an AI-powered work assistant that connects business data, internal knowledge and external tools so that teams can build and use agents, workflows and apps on top of the organization’s data.[11] Because the products were introduced recently, the terms combine contract text in the Cloud Subscription Agreement and Program Guide with product documentation, so that is how this article cites them.
Editions
AgentSpot
AgentSpot is not one of the Analytics editions. The Cloud Subscription Agreement gives it its own term rules: the customer selects a monthly or annual term at purchase, monthly terms renew for successive one-month periods and annual terms for one-year periods, and notice of non-renewal must be given at least 10 days before the end of a monthly period or at least 30 days before the end of an annual period. Renewal pricing is subject to change on notice before that cut-off.[3] The Agreement also lists “free access to AgentSpot” among Evaluation Subscriptions, and an evaluation may be used for production purposes only in the case of AgentSpot.[3]
The AgentSpot documentation sets the scope of what is bought. “AgentSpot subscriptions cover ThoughtSpot Pro Edition strictly for AgentSpot setup and administration, with broader analytics requiring a separate license.”[1] Standalone AgentSpot customers receive ThoughtSpot Pro Edition solely to configure, administer and support AgentSpot; the access does not include general use of ThoughtSpot Analytics or ThoughtSpot Embedded.[1] Permitted uses are: setting up and managing user access and provisioning; configuring one or more data models for AgentSpot; configuring Spotter for use through AgentSpot only; and other administrative activities needed to enable and manage AgentSpot.[1]
Unless separately purchased, AgentSpot customers may not use the included Pro Edition to: create, author or publish any Answers or Liveboards; use Spotter or Spotter agents outside AgentSpot; schedule alerts; use the ThoughtSpot Embedded API or SDK; use more than one ThoughtSpot Org; or use ThoughtSpot for standalone business intelligence, reporting, analytics or embedded analytics use cases.[1] Catalog proof: AgentSpot customers may not author Answers, schedule alerts or use the Embedded API without separate rights.
Spotter
Spotter is available in two ways. Under the Analytics plans on the pricing page, the user-based Pro and Enterprise plans include “Spotter AI Agents (25 queries/month per user)”, while the usage-based plans list “Spotter AI Agents with Unlimited LLM tokens” and an “Unlimited Spotter” add-on is listed separately.[5] With AgentSpot, “Spotter is included with AgentSpot at no additional cost. Use of Spotter with other ThoughtSpot products requires a separate subscription.”[1] Administrators enable Spotter in the Admin portal and users or groups “must have Can use Spotter permissions to access Spotter”.[7] The Spotter version setting lists Spotter 3, Spotter Agent (Spotter 2), Spotter Classic (Spotter 1) and the earlier Sage / Ask Sage search, the last marked deprecated.[7] The administration documentation notes that, in the multi-tenant Pro edition, ThoughtSpot AI settings are configured by ThoughtSpot, with ThoughtSpot AI “included in consumption plans, per user” and available as an add-on for other plans.[12]
Analyst Studio and other add-ons
The pricing page lists Analyst Studio features (data prep, mashups and modeling; advanced analysis with SQL, R and Python) and an in-memory data cache (Analyst Studio or Spotcache) as add-ons, together with the Model Context Protocol (MCP) Server.[5] The Mode Subscription Addendum is a separate contract: it “applies to all purchases of ThoughtSpot Mode products”, replaces references to ThoughtSpot Cloud with the Mode Service, and prevails over the agreement it is attached to in a conflict.[9] The retrieved documents do not state how ThoughtSpot Mode products relate to the Analyst Studio features sold as add-ons, so which terms apply to an Analyst Studio purchase should be confirmed in the order form. A registration for the Studio version or another free trial of the Mode Service is governed by the evaluation provisions of the Cloud Subscription Agreement.[9] Mode support is by email and chat on weekdays from 5 a.m. to 6 p.m. Pacific Time, not under the Program Guide’s support terms.[9] The documentation does not state publicly whether Analyst Studio is licensed per user or per usage, beyond its credit consumption described below.
Metrics
Data to Outcome credit. “A credit is consumed anytime a user starts a new conversation with an agent, a Workflow is run on a user’s behalf or an agentic application is loaded.” DOs are measured weekly, and any single conversation, workflow or app load active beyond a continuous 24-hour period consumes one additional DO for each subsequent 24-hour period or portion of one.[1] The FAQ states billing is per conversation, that one continuing chat within 24 hours uses one credit, that three independent chats with the same agent use three credits, and that administrators cannot assign credit limits per user.[2] Catalog proof: A DO credit is consumed per conversation, workflow run or app load.
Spotter AI Agent query allowance. The 25 queries per month per user on user-based Pro and Enterprise is a published allowance, but the retrieved pricing page does not define what counts as a query for it or what applies above 25.[5] On consumption plans the counting rule is documented separately: each time a user uses natural language to generate or modify a data answer with Spotter, it counts as a query, and click drill-downs are not additional Spotter queries.[6] A Spotter-built Liveboard counts one billable query per ThoughtSpot query on it.[6]
LLM token. The pricing page states: “ThoughtSpot does not meter or charge for LLM tokens. Platform use is governed by your subscription. If you use your own LLM provider, that provider’s fees may apply.”[5] AgentSpot is the exception to the idea of unlimited tokens, because it has a Reasonable Use cap.
Counting / floors
Reasonable Use. “Unlimited use of AgentSpot includes foreseeable uses such as long conversations and complex workflows”, but heavy programmatic query commands or performance stress-testing are not included, and ThoughtSpot “reserves the right to limit use above Reasonable Use”.[1] Reasonable Use is monthly usage that does not exceed 1 billion input tokens and 50 million output tokens from language models in the account. If use beyond that is found, ThoughtSpot gives written notice describing the usage patterns and a time in which the customer may purchase additional Data to Output Credits, failing which it may suspend or terminate use until the issue is resolved.[1] The documentation names the credit unit both “Data to Outcome” and “Data to Output” in different places, which a reader should treat as one unit. Catalog proof: AgentSpot Reasonable Use is 1B input and 50M output tokens a month.
Abnormal use. The contract adds that the customer may not use AgentSpot “to exceed the abnormal use restrictions as described in the Documentation or otherwise for large-scale automated batch processing or agent actions, mass data extraction, or scraping or re-syndication of outputs to services outside your organization”.[3]
Connectors need their own licences. Adding a connector to AgentSpot requires the MCP server URL for the external application, and “Your company must have a license for the external application” together with access to it.[11] A customer’s use of a third-party service through a connector is therefore licensed by the third party, not by ThoughtSpot.
Analyst Studio rows and refreshes. For row capacity, ThoughtSpot sums unique rows per Connection including rows made available through an Analyst Studio dataset.[10] For consumption plans, a scheduled Analyst Studio dataset refresh counts as five queries, a manual refresh as one and an ad-hoc SQL IDE query as one.[6]
Support. AgentSpot support is available 24 hours a day, 7 days a week, with the same response targets as ThoughtSpot Pro: P0 within 12 hours, P1 within 24 hours and P2 within 48 hours.[4]
Virtualization & partitioning
No virtualization or processor rule applies. The structural limit in AgentSpot is that the included Pro Edition may not be used with more than one ThoughtSpot Org.[1]
Cloud / BYOL
AgentSpot is hosted by ThoughtSpot in locations listed in its documentation (GCP us-central1 and europe-west1; Azure eastus and westeurope).[1] For Spotter, the LLM comes from ThoughtSpot by default. Administrators may choose a ThoughtSpot-provided LLM or bring their own LLM key, and ThoughtSpot says it recommends its managed auto-select option and that custom keys are configured by contacting support.[7][8] If a customer connects its own LLM, “you are responsible for its performance, cost, and maintenance”, so token costs move to the customer’s own provider contract.[8] The Program Guide adds that AgentSpot is an AI-Dependent Service whose reliance on Improving Features cannot be opted out of while still using the service.[4] The AgentSpot FAQ states that data is sent to the LLM without a masking layer and that enterprise-plan data is not used for LLM training.[2]
Programs
- AgentSpot rights of use for ThoughtSpot Pro. Pro Edition for setup and administration only.[1]
- AgentSpot Reasonable Use. Monthly token cap with notice before enforcement.[1]
- AgentSpot monthly or annual term. Auto-renewing with 10 or 30 days’ notice.[3]
- Unlimited LLM tokens. No metering of tokens on ThoughtSpot-provided models.[5]
Out of scope
This article does not cover AgentSpot prices, which ThoughtSpot does not publish in the retrieved pages, the number of credits in an AgentSpot subscription, or the commercial terms of any specific AI model provider. It does not cover Sage as a product, because the retrieved documents show it only as a deprecated Spotter setting, and it does not cover the data-protection and AI regulatory aspects of using these products, which are outside the scope of a licensing wiki. See generative AI licensing for the general treatment.