Analytics and Business Intelligence Platforms (Gartner)

Two analyst firms looked at this market from different documents and reached the same conclusion, which is rarer than it sounds.

Forrester's last evaluation of it ended on the argument that the bottleneck was never the user interface. It was semantics. Gartner's research now attaches a forecast to the same problem: agentic analytics projects that rely on a protocol without a governed semantic layer underneath it are expected to fail at scale.

The vendor lists are worth reading. The agreement underneath them is worth more.

The June 2026 Magic Quadrant for Analytics and Business Intelligence Platforms, and the vendor it demoted

The quadrant was published on 29 June 2026 and credited to Anirudh Ganeshan, Christopher Long and Edgar Macari. Six vendors are positioned as Leaders.

Microsoft is a Leader for the nineteenth consecutive year, with Power BI and Fabric positioned around semantic models and AI-ready data. Qlik is a Leader for the sixteenth consecutive year. Google is a Leader for the third consecutive year with Looker. Salesforce holds a Leader position with Tableau. ThoughtSpot is a Leader, cited for its Spotter agent and for connecting to external semantic layers including dbt, LookML, Snowflake Semantic Views and Databricks Unity Catalog.

The sixth is the one that moved. Amazon Web Services entered the Leaders quadrant with Amazon Quick, promoted from Challenger. Gartner's stated strengths are the integration with the rest of the AWS data and analytics stack, a pricing model that is comparatively predictable, and agentic workflow management that reaches into external systems through OpenAPI and MCP. The cautions are equally specific: a smaller user community than the incumbent Leaders, and administration, monitoring, auditing and cost management that run through AWS services sitting outside the product itself.

The demotion is on the other side of the chart. Oracle, a Leader in the previous edition and the vendor Forrester assessed as having the strongest current offering in its 2023 evaluation of this market, moved to Visionary. Gartner's reason, as Oracle reports it, is limited visibility for Oracle Analytics Cloud and infrequent adoption as a standalone choice outside the Oracle estate.

Why the semantic layer decides whether any of this works

Gartner's Market Guide for Agentic Analytics, published on 9 February 2026, contains the number that matters here. By 2028, sixty percent of agentic analytics projects that rely solely on MCP are expected to fail because of the absence of a consistent semantic layer.

That is a forecast about plumbing, not about models. MCP standardizes how an agent asks a question. It says nothing about what the answer means. An agent that asks for revenue will receive whatever the underlying layer decides revenue is, and it will act on that answer without the hesitation a human analyst brings.

Alongside that prediction sits a second one: by 2028, sixty percent of users are expected to run ad hoc and exploratory analysis through general-purpose language models, while production-grade reporting stays in platforms like the ones this quadrant scores. The exploratory surface and the reporting surface are separating, and only one of them is being governed.

A retired name on the other side of this market

The Forrester evaluation of this market is covered separately on this site at Augmented Business Intelligence Platforms, and its subject is a name that no longer exists.

Forrester evaluated this market as Augmented BI Platforms in 2021 and as Augmented Business Intelligence Platforms in 2023, then dropped the adjective for The Forrester Wave: Business Intelligence Platforms, Q2 2025. The word left because every vendor could do the thing it described. Augmentation went from a differentiator to a baseline, which is the specific way successful technology becomes invisible.

Two facts from that lineage are worth carrying into the Gartner reading. In the 2023 Wave, Oracle was a Leader with what Forrester assessed as the strongest current offering in the market, and the only vendor to take maximum scores in both core enterprise BI and augmented analytics at once. Three years later, Gartner has moved Oracle down and promoted AWS past it.

And the 2025 Wave made generative AI functionality a scored criterion in its own right. Microsoft placed as a Leader with the highest score of any vendor in that criterion. The capability that replaced augmentation as the thing worth measuring is now, in the Gartner quadrant, simply part of the field.

The market is splitting again, from the other direction

Something is happening to this category that did not happen the last time.

When Forrester retired the augmented adjective, the market contracted into one name. Now Gartner is publishing a Market Guide for Agentic Analytics as a subject distinct from the quadrant for analytics and business intelligence platforms, which means the market is widening into two.

The two documents score different things. The quadrant scores vendors as companies: their vision, their execution, their ability to be bought and run. The Market Guide describes how a market is forming, and names representative vendors rather than ranking them. Cube, Dremio and Unsupervised appear there as representative vendors, not as Leaders, because a Market Guide does not have Leaders.

A buyer who reads only the quadrant will select a platform. A buyer who reads only the Market Guide will learn that the thing they are selecting is changing shape. Neither document is sufficient alone, and the one that is easier to find is the one that will be quoted in the sales deck.

Four questions for the analytics buyer

Who owns the definition of your metrics? Not who owns the dashboard. Who decides what revenue, active customer and margin mean, where that decision is recorded, and what happens when two departments disagree. Every question below depends on this one.

Can your platform answer the same question two ways and return the same number? This is the semantic layer test in one step. Take a metric your business cares about, ask for it in three different phrasings, and compare. If the numbers diverge, you have found the problem before paying for it.

Do you know what your agents are going to do with a wrong answer? A human who receives an implausible figure usually notices. An agent acts on it, at volume, without asking. The governance question is not whether the agent is accurate. It is what happens when it is not.

Which of the two markets are you actually buying in? The quadrant scores the reporting platform your finance team will live in. The Market Guide describes the exploratory, agentic surface that is forming around it. They have different vendors, different risks and different timelines, and a single procurement will not cover both well.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner's research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

The Forrester evaluation of this same market is covered on this site at Augmented Business Intelligence Platforms (Forrester), where the subject is a name Forrester retired. The market was scored there as Augmented BI Platforms in 2021 and Augmented Business Intelligence Platforms in 2023, before the adjective was dropped in the Business Intelligence Platforms Wave of Q2 2025.

The agentic layer this market is growing into already has coverage of its own: Agentic Analytics, where the question of what an autonomous agent does with an answer nobody checked is the entire subject.