live wire
▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub▸Snowflake agent evals can pin judge versions; unpinned metrics may roll forward and shift scoresSnowflake▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub
nl2sql.ai
newsPricing

Tableau’s capacity pricing covers AI agents—but authors still need seats

The new Viewer Blocks widen access without per-viewer licenses or overage charges, while Creator and Explorer licensing—and an opaque contact-sales price—remain part of the deployment math.

Named-user seats versus shared capacity for viewers and AI agents
Side by side: what changed
By The News Desk· Sep 30, 2026the quick take — two AI hosts go live when you do

Tableau has changed the licensing boundary for analytics consumed by large audiences and automated systems. Its new Tableau Cloud capacity model lets organizations buy shared Viewer Blocks for employees, partners, customers and AI agents instead of assigning every viewer an individual license. Tableau says the same blocks can cover MCP integrations and headless BI workflows, with no overage charges. Tableau’s announcement and pricing page describe the model.

Consumption moves to capacity; creation does not

The practical distinction is between producing analytics and consuming them. Tableau says customers first determine how many Creator and Explorer licenses they need, then use capacity blocks for broad Viewer access. Creators and Explorers therefore remain licensed per user under the announced Cloud model. Every deployment also requires at least one Creator license, according to Tableau’s pricing page.

That matters for agent deployments because an automated workflow does not erase the human authoring layer. Teams still need licensed people to create workbooks, curate data sources, administer content and build Tableau Knowledge graphs. Capacity changes how widely those assets can be consumed; it does not turn authoring into an unmetered service.

Agent access joins the Viewer pool

Tableau explicitly places AI agents and automated workflows—including MCP and headless BI uses—inside the capacity audience. The blocks are shared across a deployment, and Tableau says customers can add blocks as the audience grows. The model is offered for Tableau Cloud Standard, Enterprise and Cloud+, as well as the Tableau+ Bundle.

For procurement teams, the useful comparison is no longer simply “agent seat versus human seat.” It is the fixed Creator and Explorer base plus the block size required for viewers and machine consumers. Seasonal usage, customer-facing embedded analytics and background agent jobs can now draw from the same capacity construct.

The missing number is still the important one

Tableau has not posted a public price for Viewer Blocks. Its pricing page directs buyers to contact sales for capacity terms. It also lists starting per-user prices of $15 a month for Tableau Standard and $35 a month for Tableau Enterprise, billed annually, but those figures do not reveal the price, unit size or practical throughput of a Viewer Block.

Before treating the change as a predictable agent budget, buyers should request the block definition, measurement window, included machine activity and any throttling or fair-use rules in the contract. “No overage charges” removes one surprise, but it does not by itself show whether a block fits an interactive dashboard audience, a high-frequency agent workload, or both.

The launch is still a meaningful packaging shift: Tableau is acknowledging that analytics consumers now include software agents that do not map cleanly to named-user licensing. The cost case, however, remains impossible to calculate from the public tariff alone.

Filed by The News Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.