The best data intelligence platforms in 2026, compared

A customer-hosted, time-aware graph of the organisation with an evidence status on every fact, designed so that findings can become governed agents and workflows.

Best for: Organisations that want a controlled picture of people, groups and ownership to give AI agents context, and that can start from directory data. Not a warehouse, a lakehouse or a dashboard tool today.

Data sources and connectors: Built today: directory sources (Active Directory and LDAP, Microsoft Entra ID, Okta, Google Workspace). Cloud, code, Kubernetes, database, SaaS and document sources are designed for. The connector list beyond directory sources is open.[1][2]

AI-native analytics and agents: Built: a time-aware graph with an evidence status on every fact, and a local API that serves graph, as-of, coverage and evidence reads. Designed for: plain-language questions, trend detection, dashboards and wiring findings into agents.[1][2]

Data governance and access control: Built: token-scoped API with the tenant taken from the credential, row-level tenant isolation, a hash-chained audit log, and answers limited to what the person asking may see. Designed for: an action gateway with approvals.[1][2]

Deployment and data residency: Customer-hosted: a virtual machine with Docker, Kubernetes with Helm, or an air-gapped route. Connected, private and air-gapped modes. In air-gapped mode no outbound link is started.[1]

Pricing transparency: <pricing - founder to fill> <availability - founder to fill>[1]

APIs and extensibility: Built: a token-scoped local API for graph, people, groups, coverage and audit. Designed for: a context-package API and an MCP server.[1]

Built and designed, kept apart:

Watch-outs:

• Today it reads directory sources only. Cloud, code, SaaS and document sources are on the roadmap.
• Plain-language questions, dashboards and trend detection are design intent, and named dashboards and chart types are not published.
• Not a data warehouse, lakehouse or business intelligence replacement.
• Pricing and availability are not published. <pricing - founder to fill> <availability - founder to fill>
• A pre-model sanitisation gateway is in progress. Do not assume content is cleaned before it reaches a model.
• Compliance wording: this is built for compliance-by-design, not “compliant”. Swfte does not hold a SOC 2 report, an ISO 27001 certificate or a HIPAA BAA today. A SOC 2 Type I audit is in preparation; the trust page lists what is in place and what is in progress.

Sources: Swfte Intelligence Platform page, Swfte intelligence deployment page (read 2026-10-06).

Microsoft's business intelligence service for reports and dashboards, sold alongside Microsoft Fabric, its unified analytics platform with capacity-based pricing.

Best for: Organisations already on Microsoft Entra ID and Azure that want per-user report licences, plus Fabric capacity and an on-premises report server option where needed.

Data sources and connectors: Power BI connects to data through Power Query connectors, documented as a long alphabetical list; Microsoft does not state a total count on the page read. Fabric adds lakehouse, warehouse and OneLake storage as native engines.[1][2][3]

AI-native analytics and agents: Copilot in Power BI answers questions in chat and builds reports and DAX. Fabric data agents are configurable question-and-answer agents over lakehouses, warehouses and Power BI semantic models, with custom instructions. Both need a paid Fabric (F2 or higher) or Premium (P1 or higher) capacity.[1][2]

Data governance and access control: Row-level security uses DAX filter roles on semantic models; object-level security restricts columns. A Power BI activity log records user actions and is available to Fabric administrators. Users sign in through Microsoft Entra ID, and DirectQuery single sign-on is documented. RLS does not apply to workspace Admin, Member or Contributor roles.[1][2][3]

Deployment and data residency: Cloud service on Azure. Multi-Geo lets customers put a capacity in a region other than the tenant's home region. Power BI Report Server is an on-premises server, licensed through Fabric F64+ reserved capacity or SQL Server licences.[1][2]

Pricing transparency: Power BI Pro is $14.00 and Premium Per User is $24.00, each per user per month, paid yearly. Power BI Embedded and Fabric capacity (reservation or pay-as-you-go) show 'Variable' rather than a figure on the page.[1]

APIs and extensibility: A public Power BI REST API covers content management, admin operations and embed tokens. Embedding options and a custom-visuals SDK (.pbiviz packages) are documented. Open-source components are not stated on the pages read.[1][2][3]

Score 93.3 out of 100. Data sources and connectors: Rubric 3 allows a native lakehouse/warehouse engine plus a connector list: Fabric data agent page names lakehouses, warehouses and OneLake; Power Query page lists connectors. AI-native analytics and agents: Natural-language Copilot plus configurable Fabric data agents (agent building) plus Power BI semantic models as the governed model layer Copilot and agents read. Data governance and access control: Row-level security (fine-grained) and activity log (audit) and Entra ID sign-in/SSO all published. Deployment and data residency: Region choice (Multi-Geo) and an on-premises customer-run option (Power BI Report Server) are both published. Pricing transparency: Per-user tiers priced with a defined unit and billing term; embedded and Fabric capacity prices not shown on the page. APIs and extensibility: Public REST API and a custom-visuals SDK published; no open standard or open-source statement seen, so not 3.

Watch-outs:

• Copilot needs paid Fabric capacity F2 or higher, or Premium P1 or higher; Pro or PPU licences alone are not enough.
• Row-level security does not apply to workspace Admin, Member or Contributor roles.
• Fabric data agents currently return at most 25 rows and 25 columns and do not support unstructured files.
• Power BI Report Server is not included with F SKUs below F64 reserved instances.

Sources: Power BI pricing, Power BI activity log, Row-level security, Power BI data sources, Power BI security, Power Query connectors, Copilot for Power BI, Embedded analytics, Fabric Multi-Geo, Power BI Report Server, Fabric data agent, Power BI REST API, Power BI custom visuals (read 2026-10-06).

A lakehouse data and AI platform on AWS, Azure and Google Cloud, with Unity Catalog governance, Genie for natural-language questions and Agent Bricks for agents.

Best for: Data teams that want one governed platform for SQL, notebooks, natural-language data questions and agent development, and that can run compute inside their own cloud account.

Data sources and connectors: Lakeflow Connect lists managed connectors in six categories (database CDC, SaaS, file sources, query-based, streaming, community) and names about 16 sources, such as MySQL, PostgreSQL, SQL Server, Salesforce, Workday, SharePoint. The page gives no total count. The platform is described as a lakehouse.[1][2]

AI-native analytics and agents: Genie lets users ask data questions in natural language. Genie Agents are set up by data teams with trusted data sources, metrics and business rules. Agent Bricks is described as the agent developer platform, with MCP support. Metric views are described as a governed semantic layer in Unity Catalog that Genie Agents can query.[1][2][3]

Data governance and access control: Unity Catalog manages access with privileges, attribute-based policies and row and column filters, keeps audit logs of data access and system activity, and tracks lineage. The security page lists single sign-on with providers such as Microsoft Entra ID and Okta.[1][2]

Deployment and data residency: Runs on AWS, Azure and Google Cloud, and customers choose a region. A customer-managed VPC puts workspace compute in the customer's own AWS account, while Databricks runs a separate control plane. Some designated services use Databricks Geos for data residency.[1][2][3]

Pricing transparency: Pay-as-you-go billing in Databricks Units (DBUs) at per-second granularity, with no up-front costs. The pricing pages fetched do not show dollar rates; they point to a pricing calculator and to sales for committed-use contracts.[1][2]

APIs and extensibility: A REST API covers almost all resources. SDKs exist for Python, JavaScript, Java, Go and R, plus a CLI. Agent Bricks supports MCP servers, and the SQL pricing page describes open formats and APIs.[1][2][3]

Watch-outs:

• Prices vary by cloud, region and billing terms; the fetched pages give no dollar rates.
• Azure Databricks pricing is set by Microsoft and published on Azure.com.
• Connector count is not published; connectors are in various release states.

Sources: Databricks pricing overview, Databricks SQL pricing, Unity Catalog docs, Genie docs, Metric views docs, Agent Bricks docs, Lakeflow Connect docs, Developer tools docs, Supported regions docs, Security overview docs, Customer-managed VPC docs (read 2026-10-06).

Enterprise platform that puts analytics, machine learning and AI agents in one governed system, with orchestration and oversight.

Best for: Large organisations that want data preparation, machine learning, AI agents and governance in one platform and can accept sales-led pricing and self-managed or customer-cloud deployment.

Data sources and connectors: Dataiku's documentation lists connections to file storage, cloud warehouses (Snowflake, Databricks, BigQuery, Redshift, Synapse), many SQL and NoSQL databases and application sources such as Salesforce and ServiceNow. A Dataiku overview page says 40+ connectors; the documentation list looks longer, so no single count is confirmed.[1][2]

AI-native analytics and agents: Documentation describes building, evaluating and deploying AI agents (simple visual, structured visual, code and external agents) and an Agent Hub and Agent Chat for conversation. Semantic Models define entities, attributes, relationships and business metrics that agents use to query data.[1][2]

Data governance and access control: SSO via OIDC, SAML and Kerberos is documented, as are group-based project permissions, an audit trail of user actions and column-level data lineage. The pages read did not describe row-level or policy-based access to data inside datasets.[1][2][3][4]

Deployment and data residency: Documentation describes Dataiku Cloud (hosted by Dataiku), Dataiku Cloud Stacks (managed in your own AWS, GCP or Azure tenant) and custom self-managed installs on your own Linux server, on-premises or in the cloud. No region choice was seen on the pages read.[1]

Pricing transparency: No prices were found. A 14-day free trial and a free edition are offered, and paid and hosted plans are handled through the sales team.[1]

APIs and extensibility: Documentation covers a Python API client, an HTTP REST API and an R API. No open-source or open-standards statement for the APIs was seen.[1]

Watch-outs:

• Paid plans carry no published price; the vendor directs buyers to its sales team.
• The free trial excludes Govern and advanced LLM Mesh and is limited to 2 users and 4 CPUs.
• Pages read did not document row-level data access control.

Sources: Dataiku product, Dataiku get started, Dataiku at a glance, DSS supported connections, DSS installation, DSS security, DSS permissions, DSS audit trail, DSS data lineage, DSS public API, DSS agents, DSS semantic models (read 2026-10-06).

A work AI platform that searches and answers across connected company apps, with chat, an agent builder and developer APIs.

Best for: Companies wanting permission-aware search, chat and agents across many workplace apps, including those that need the platform hosted in their own AWS or GCP account.

Data sources and connectors: Glean publishes 275+ connectors, covering native connectors, MCP-based connectors and Push API connectors. Custom sources can be added through an Indexing SDK.[1][2]

AI-native analytics and agents: Chat answers questions across connected apps with citations. An agent platform offers a no-code builder, code, MCP and human-in-the-loop steps. This is knowledge search and chat rather than SQL analytics, and no semantic layer was seen.[1][2]

Data governance and access control: Admin roles (Setup Admin, Admin, Super Admin), SSO configuration in the Admin console, and admin audit logs with a default 30-day retention and CSV export. Results follow each user's source-system permissions. Audit logs cover admin actions, not end-user activity.[1][2][3][4]

Deployment and data residency: Glean Hosted runs a single-tenant instance on GCP in multiple regions worldwide. Customer Hosted deploys Glean as a managed service inside the customer's own GCP or AWS environment, with data-residency guarantees.[1]

Pricing transparency: not verified

APIs and extensibility: Search and Chat APIs with client libraries in Python, TypeScript, Go and Java, a Web SDK, an Indexing SDK for custom data sources, and MCP and OpenAPI-based actions.[1][2]

Watch-outs:

• Admin audit logs exclude end-user activity and default to 30 days' retention.
• Customer Hosted deployments do not support manual deployment, patching or architecture changes.

Sources: Glean connectors, Glean developer platform, Glean agents, Glean security, Glean deployment models, Glean admin audit logs, Glean administrator roles, Glean docs home (read 2026-10-06).

Collaborative analytics workspace with notebooks, an AI agent, conversational Threads and a shared context engine built on semantic models.

Best for: Data teams on a cloud warehouse that want notebooks plus self-serve AI questions on governed definitions, and are happy with per-editor pricing and OIDC single sign-on.

Data sources and connectors: The data connections documentation lists 16 database types in three support tiers, including Snowflake, BigQuery, Databricks, ClickHouse, Postgres and Redshift. The homepage also mentions dbt.[1][2]

AI-native analytics and agents: Threads lets users ask natural-language questions answered from semantic models and warehouse tables. Built-in agents include a Notebook Agent and a Modeling Agent, and the Modeling Workbench defines governed dimensions and measures. The AI docs describe built-in agents only and do not describe building custom agents.[1][2][3]

Data governance and access control: Enterprise plan: OIDC SSO (SAML not supported), audit logs, SCIM directory sync and role-based permissions. Signed embedding supports row-level security. SSO and audit logs are not on lower plans.[1][2][3][4]

Deployment and data residency: Multi-tenant cloud including separate HIPAA and EU stacks, plus single-tenant VPCs set up by Hex with secure peering. Pricing page lists single-tenant as an Enterprise add-on and EU multi-tenant by contract.[1][2]

Pricing transparency: Community is free. Professional is $36 per editor per month and Team is $75 per editor per month; Enterprise is custom. The page does not state whether monthly or annual billing applies. AI features use credits, with per-seat monthly grants on paid plans.[1]

APIs and extensibility: A public API runs and manages projects, users and groups, with personal and workspace tokens. A hextoolkit Python package wraps it, and an MCP server is available on Team and Enterprise plans. The API is limited to 60 requests per minute per user.[1][2]

Watch-outs:

• SSO supports OIDC only; SAML is not supported.
• SSO and audit logs are Enterprise-only; the MCP server needs the Team or Enterprise plan.
• Public API rate limit is 60 requests per minute per user, with a maximum of 25 concurrent kernels.

Sources: Hex homepage, Hex pricing, Hex enterprise, Hex data connections, Hex SSO, Hex workspace security, Hex embedding, Hex AI overview, Hex modeling workbench, Hex MCP server, Hex API overview (read 2026-10-06).

Google Cloud's business intelligence platform built around the LookML semantic model, sold as Looker (original) and Looker (Google Cloud core).

Best for: Teams that want governed metrics defined in code (LookML), embedded or API-driven analytics, and a choice between Google-hosted or customer-hosted deployment.

Data sources and connectors: Looker connects to SQL databases through a published dialect table: 60 rows (46 marked Supported) in Looker (original), counting version variants separately, which is roughly 45 distinct database products. Looker (Google Cloud core) supports fewer dialects.[1][2]

AI-native analytics and agents: Conversational Analytics in Looker is a Gemini-powered chat-with-data feature grounded in the LookML semantic layer. Users can create data agents with custom context that query up to five Explores. A Looker-managed MCP server lets agents such as Gemini CLI, Claude and Cursor reach a Looker instance.[1][2]

Data governance and access control: access_filter in LookML applies user-specific row restrictions via user attributes. Looker supports SAML 2.0 authentication. Looker (Google Cloud core) writes Cloud Audit Logs (Admin Activity, Data Access when enabled, System Event); Looker (original) has System Activity, described as usage monitoring rather than compliance auditing.[1][2][3][4]

Deployment and data residency: Looker (original) supports customer-hosted and other-cloud deployments. Looker (Google Cloud core) is hosted by Google in the Google Cloud location you pick at setup, with private networking options, and is not available customer-hosted or multicloud.[1][2]

Pricing transparency: Looker (Google Cloud core) has Standard, Enterprise and Embed editions. The edition documentation read does not publish prices and points to contact sales. The pricing page itself returned almost no content when fetched.[1]

APIs and extensibility: Looker has an API with official SDKs in Ruby, Python and TypeScript/JavaScript, with source code on GitHub under the looker-open-source organisation, and a legacy codegen project for other languages.[1]

Watch-outs:

• Looker (Google Cloud core) supports fewer database dialects than Looker (original).
• Looker (Google Cloud core) does not support LDAP or email/password sign-in; it uses Google OAuth and SAML/OpenID Connect.
• access_filter applies per Explore; an Explore without it is not restricted.
• Looker (Google Cloud core) is unavailable for customer-hosted or multicloud deployments.

Sources: Looker dialects, Looker core vs original, Conversational Analytics overview, Looker MCP server, access_filter reference, Looker SAML, Looker core audit logging, System Activity, Looker core overview, Looker core editions, Looker API SDKs (read 2026-10-06).

Foundry is Palantir's data platform built around an Ontology; AIP adds AI tools (Logic, chatbot studio, evals) on top of it.

Best for: Organisations with strict access-control needs that want data modelled as business objects and actions, with AI tooling on that model, and that accept a sales-led engagement.

Data sources and connectors: The Data Connection docs list many source types, including cloud storage, Snowflake, BigQuery, Databricks, SAP, Oracle, Salesforce and file systems, with agent-based connections for on-premises sources. No count is published, and the page was longer than the portion read.[1][2]

AI-native analytics and agents: AIP includes AIP Logic, AIP Chatbot Studio (formerly Agent Studio) for conversational agents, and AIP Evals, all built on the Ontology. The Ontology maps data into object types, properties and links with security and governance. Natural-language analytics over data was not confirmed on the pages read.[1][2]

Data governance and access control: Single sign-on and multi-factor authentication; resource roles (Owner, Editor, Viewer, Discoverer); markings and mandatory controls; row and column filtering through restricted views and object and property security policies; security audit logging. SAML 2.0 is supported for identity providers.[1][2]

Deployment and data residency: not verified

Pricing transparency: not verified

APIs and extensibility: A REST API with OAuth 2.0 and JSON, an Ontology SDK (OSDK) in TypeScript, Python and Java, and a Platform SDK.[1]

Watch-outs:

• Several Palantir product and docs URLs returned no usable content, so deployment and pricing are not verified.
• The Data Connection source page was longer than one fetch; the last part was not read.

Sources: Foundry security overview, Foundry administration overview, Foundry API overview, AIP overview, Ontology overview, Data Connection sources, Architecture Center overview (read 2026-10-06).

Embedded analytics platform with an AI assistant, natural-language queries and a Compose SDK for putting analytics inside other products.

Best for: Software companies embedding customer-facing analytics, especially those needing row-level security, an SDK and an on-premises or dedicated-cloud option; pricing is not published.

Data sources and connectors: Documentation lists 12 native live connectors (e.g. Snowflake, BigQuery, Databricks), 19 native ElastiCube connectors, partner-supported connectors and vendor JDBC sources. No total count is published.[1]

AI-native analytics and agents: Sisense Intelligence offers an interactive Assistant, natural-language query, narratives and semantic enrichment of data models. The homepage mentions governed semantics and shared formulas. Agent building was not confirmed on the pages read.[1][2]

Data governance and access control: Documentation covers row-level data security rules, SSO and role-based permissions, and audit logs that record logins including SSO. The pricing page lists column-level security and an SSO Router under Enterprise.[1][2][3][4]

Deployment and data residency: Enterprise plan lists SaaS, dedicated cloud on AWS, Azure or GCP, and on-premises. Self-Serve runs in Sisense's cloud. No region choice was seen on the pages read.[1][2]

Pricing transparency: The pricing page names Enterprise and Self-Serve plans but shows no prices or billing unit. Enterprise uses 'Talk to us'; Self-Serve offers a free trial.[1]

APIs and extensibility: A REST API v1.0 reference and the Compose SDK (data and UI packages for React, Angular and Vue) are documented. The developer docs link to GitHub; no explicit open-source licence statement was read.[1][2][3]

Watch-outs:

• No prices are published for either plan.
• Audit logs are retained 30 days by default, configurable up to 9999 days.
• Row-level rules restrict whole rows; the data access page does not describe column restrictions.

Sources: Sisense homepage, Sisense pricing, Sisense data sources, Sisense data access security, Sisense audit logs, Sisense security, Sisense Intelligence, Compose SDK, Sisense REST API (read 2026-10-06).

A cloud data platform on AWS, Azure and Google Cloud, with Cortex Agents, Cortex Analyst and semantic views for AI over governed data.

Best for: Teams that keep data in a managed cloud warehouse and want natural-language analytics and agents under existing role, masking and row-policy controls.

Data sources and connectors: Snowflake is a native warehouse engine with tables, Iceberg tables and hybrid tables. Openflow, built on Apache NiFi, lists more than 25 named connectors in its docs navigation (for example Salesforce, SharePoint, MySQL, Kafka) and refers to hundreds of processors. No exact connector count is given.[1][2]

AI-native analytics and agents: Cortex Agents is described as a managed platform for building and running agents, using tools such as Cortex Analyst (SQL over structured data), Cortex Search, custom tools and remote MCP servers. Business users ask questions in natural language in a CoWork application. Semantic views store metrics and business definitions as governed schema objects.[1][2][3]

Data governance and access control: SSO through SAML 2.0 and OIDC; role-based access control; dynamic data masking and row access policies; Access History records who read or wrote which columns, with column lineage. Access History needs Enterprise Edition or higher.[1][2][3][4]

Deployment and data residency: Customers choose the region when requesting an account, across AWS, Azure and Google Cloud; each account sits in a single region. The key-concepts page says the platform cannot be installed on private cloud or locally. Openflow has a bring-your-own-cloud option that runs in the customer's VPC, but that covers the integration service only.[1][2][3]

Pricing transparency: Four editions are named: Standard, Enterprise, Business Critical and Virtual Private Snowflake. Pricing is consumption-based, with on-demand or pre-paid capacity. The pricing page shows no dollar figures and points to a consumption table PDF, which could not be read.[1]

APIs and extensibility: Documented REST APIs (OpenAPI-compliant), a Python API, language drivers, Snowpark, a CLI and a Terraform provider. Iceberg tables are supported, Openflow is built on Apache NiFi, and Cortex Agents connect to remote MCP servers.[1][2][3][4]

Watch-outs:

• Access History (audit and column lineage) requires Enterprise Edition or higher.
• Each account is hosted in a single region; using several regions needs an account in each.
• The docs page at the Snowflake Intelligence URL describes the app as Snowflake CoWork; naming may have changed.

Sources: Snowflake pricing options, Snowflake key concepts, Snowflake regions, Snowflake Intelligence / CoWork docs, Cortex Agents docs, Semantic views docs, Openflow docs, Federated authentication docs, Column-level security docs, Row access policy docs, Access History docs, Snowflake REST API docs, Snowflake ecosystem docs (read 2026-10-06).

Salesforce's analytics product line: Tableau Cloud (vendor-hosted), Tableau Server (customer-run) and the newer Tableau Next with agentic analytics.

Best for: Teams that want per-user licensed visual analytics with a choice of vendor-hosted regions or self-run Tableau Server, and Salesforce-linked AI features.

Data sources and connectors: Tableau documents a Connect pane listing file types and server connectors, and points to a connector list on its website. No connector count was stated on the help pages read.[1]

AI-native analytics and agents: Tableau Pulse is a metrics service with metric definitions (a metrics layer) and AI-generated insights; Tableau Agent in Pulse takes natural-language questions across metrics. Salesforce's pricing page lists Tableau Next with agentic analytics, Tableau Agent and Semantics.[1][2]

Data governance and access control: Tableau Cloud sign-in supports SAML, OpenID Connect and Google or Salesforce accounts, with MFA required. Row-level security options include user filters, data policies on virtual connections (needs Data Management) and entitlement tables. Admin Insights (Tableau Cloud) includes a sign-in, publishing and content-access events data source described as an audit source.[1][2][3]

Deployment and data residency: Tableau Cloud lets you choose the region where your site and data are stored (regions across Asia, Europe and North America). Tableau Server installs on Windows or Linux on customer-managed infrastructure.[1][2]

Pricing transparency: Salesforce's analytics pricing page lists Tableau Cloud starting at $15 per user per month, billed annually, and Tableau Next starting at $40 per user per month, billed annually. Annual contracts are required. Premier support is 30% of net licence fees.[1]

APIs and extensibility: The Tableau REST API manages Tableau Server, Tableau Cloud and Prep Conductor resources, with Tableau Server Client (Python) as a client library. The Embedding API v3 is a JavaScript library with a tableau-viz web component. Open-source status was not seen.[1][2]

Watch-outs:

• Salesforce's pricing page states annual contracts are required for all editions.
• Admin Insights is stated as a Tableau Cloud-only project.
• Data policies on virtual connections require Data Management.

Sources: Salesforce analytics pricing, Tableau Cloud security, Tableau Cloud authentication, Tableau row-level security, Tableau Admin Insights, Tableau Server overview, Tableau Pulse, Tableau REST API, Tableau Embedding API, Tableau connect overview (read 2026-10-06).

A search-driven analytics platform with Spotter, an AI analyst, that queries data left in external warehouses and can be embedded in other products.

Best for: Teams that keep data in a cloud warehouse and want natural-language analytics with row-level security, embedded in an app or used on a per-user plan.

Data sources and connectors: ThoughtSpot Cloud supports 34 external database connections, including Snowflake, BigQuery, Redshift, Databricks and Oracle, using live queries rather than importing data. Connections do not support joins across connections.[1]

AI-native analytics and agents: Spotter is described as a personal AI analyst answering questions in everyday language. Spotter Agent can plan multi-step analysis and connect tools such as Slack, Confluence or Jira. Business context comes from the underlying data model, with coaching on company terms.[1][2]

Data governance and access control: Single sign-on with SAML and OIDC providers; roles, groups and privileges; object, column and row-level security rules; customers can view login and activity records.[1][2]

Deployment and data residency: Customers pick the region where their data resides, and warehouse connectivity can use VPN, private links or a proxy. No region list appears on the trust page, and other deployment options were not listed.[1]

Pricing transparency: Analytics plans: Essentials from $25 per user per month billed annually; Pro from $0.10 per credit, usage-based; Enterprise custom, contact sales. Embedded: Developer free for one year; Enterprise Embedded flexible pricing, contact sales.[1]

APIs and extensibility: A Visual Embed SDK and REST APIs for embedding search, Liveboards and visualisations, with rebranding and custom actions. The pricing page lists API and SDK access for the Developer plan.[1][2]

Watch-outs:

• Spotter is included in new per-user and consumption plans but is an add-on for Embedded and legacy plans.
• Connections do not support joins across connections.
• Essentials is capped at 5-50 users and 25M rows; Pro at 1,000 users and 250M rows.

Sources: ThoughtSpot pricing, ThoughtSpot trust security, ThoughtSpot trust overview, ThoughtSpot connections docs, ThoughtSpot security docs, ThoughtSpot Spotter docs, ThoughtSpot developer docs (read 2026-10-06).