Question 1

We use Looker Conversational Analytics. Does Orion replace it?

Accepted Answer

No, and it is not meant to. Looker Conversational Analytics is a strong way to ask questions of your LookML models conversationally, inside Looker, for the people who work there. Orion is complementary: it reads the same LookML you have already built and puts it to work across your whole stack. It watches your metrics, investigates why they moved, and delivers a written answer to Slack or email, including to people who never open Looker. A common setup keeps Looker and Conversational Analytics for in-Looker exploration and adds Orion for proactive, cross-warehouse investigation that comes to them.

Question 2

We've invested in LookML. Does Orion reuse it the way Looker does?

Accepted Answer

Yes. Orion reads your LookML through a dedicated Looker user, with optional read access to the LookML repository. It uses your dimensions, measures, and joins as the logic behind every answer, the same definitions your team already trusts. Looker Conversational Analytics grounds its answers in LookML, from inside Looker or an app built on Google's Conversational Analytics API. That is always through your Looker connection. Orion reuses that investment across the warehouses you connect. A cited Knowledge Base adds the context that is not in the model. Same semantic layer, put to work in more places.

Question 3

Looker CA already answers questions over our data. Isn't that enough?

Accepted Answer

Looker Conversational Analytics is genuinely good at what it does: Gemini-powered, multi-step analysis grounded in your governed LookML. Three limits show up in Google's own documentation. First, the optional Gemini key-driver analysis in its agentic workflows explains only the change that triggered the alert. Second, the Explore or agent you point it at bounds the conversation. Google documents at most five Explores per agent. Third, no path from a chat to a saved dashboard is documented. Results open in an Explore as a manual step. Orion clears all three. It auto-detects significant changes and investigates the root cause across the warehouses you connect. It delivers the written explanation to Slack or email, and builds the dashboard, report, or deck when you ask for one.

Question 4

Is Looker Conversational Analytics only available inside Looker?

Accepted Answer

No. The Conversational Analytics API covers Looker, BigQuery, Looker Studio, and Google's operational databases. Agents your team builds publish to Gemini Enterprise and custom apps through the API, with agent-to-agent (A2A) support. Looker's agentic workflows push alerts with Gemini key-driver analysis to Slack, email, and mobile. Orion also works inside Gemini Enterprise, Claude, and Cursor through an open connector standard, MCP. The differences that remain are depth and direction. The key-driver analysis is optional and explains the change that triggered the workflow. No path from a chat to a dashboard is documented. Orion investigates beyond the triggering metric, across every warehouse you connect. It hands you the finished output: a written narrative, a dashboard, or a deck, in Slack or email.

Question 5

Can Looker Conversational Analytics build a dashboard from a chat?

Accepted Answer

Not directly. You can ask it questions, get charts and answers back, and open the results in an Explore. From there, building the dashboard is a manual job in Looker. There is no documented path from a conversation straight to a saved dashboard. Orion does this directly. Ask for a dashboard, report, or slide deck in plain language. Orion builds it and delivers it to Slack or email, grounded in the same LookML definitions. If your business users' follow-up to every good answer is 'can I get this as a dashboard,' that difference is the workflow.

Question 6

Can Looker Conversational Analytics reason across our other warehouses?

Accepted Answer

Through Looker's connections it reaches the warehouses Looker models, including BigQuery, Snowflake, Redshift, and Databricks. The Conversational Analytics API covers Looker, BigQuery, Looker Studio, and Google's operational databases. But the Explore or agent you point it at bounds a single conversation, at most five Explores per agent. Inside Looker, one conversation reaches nothing beyond those Explores. Orion is built to reason across the warehouses you connect directly, in a single investigation. A question that touches Snowflake and BigQuery together is one analysis. If your data and questions live beyond what Looker models, that is the difference.

Question 7

Does Orion need a Looker instance or Gemini in Looker?

Accepted Answer

Not for the people asking. Orion connects to the warehouses you already run and reads your LookML through one dedicated Looker user, with optional read access to the LookML repository. Everyone else gets value from that semantic layer without a Looker account of their own, and without an admin enabling Gemini in Looker. Looker Conversational Analytics runs inside Looker with Gemini in Looker enabled, or in an app your team builds on Google's Conversational Analytics API. Orion is a way to let the LookML you have already invested in pay off for people and questions that live outside the Looker instance. And if part of your stack has no LookML at all, Orion builds a semantic layer for it from your warehouse metadata and Knowledge Base.

Question 8

Can our whole team work with Orion at the same time, or is it one person per chat?

Accepted Answer

Rooms lets a team share one conversation with Orion. The data lead and the marketer work the same thread, and attribution tracks who asked what. Each participant acts at their own access level. Viewers can follow without being able to query. Chat in a BI tool answers one person at a time. A Room makes the conversation itself the shared workspace, with Orion in it.

Question 9

How does Orion connect, and what does security look like?

Accepted Answer

Orion connects read-only, through a service account you scope to the data it may analyze. For Looker that is a dedicated analyst-level user, and on BigQuery it can run queries as the asker through per-user OAuth. Credentials are encrypted at rest, and two independent layers of roles, tenant and group, scope who sees which projects, data sources, and Knowledge Base pages. Every analysis is captured in a notebook: the instructions, the logic, and the code, so you can see how each answer was produced.

Question 10

Which should we choose, Orion or Looker Conversational Analytics?

Accepted Answer

Keep Looker either way. This is not a rip-and-replace decision. Looker Conversational Analytics is the right fit if your questions live inside what LookML models. It also fits when your users are in the Google ecosystem and want conversational Q&A over governed Explores. Orion is the better fit if you want the same LookML working proactively across your whole stack. That means changes detected and investigated beyond a single triggering metric. It also means dashboards and reports built from a request, and answers delivered to Slack or email, including to people who never open Looker. Many teams run both: Conversational Analytics for in-Looker questions, Orion for the investigation and delivery layer on top.

Question 11

Can we embed Orion in our own product?

Accepted Answer

Yes. Embedded Orion is a white-label analyst for your customers, presented under your brand inside your product. Each customer account is isolated: they see only their own data, business context, and permissions, and nothing passes between accounts. Your customers sign in to your product, and Orion inherits who they are and what they can see. There is no second login. You integrate once. Every new account is provisioned with its own branding, data connections, and access controls, so the second customer costs a fraction of the first.

Question 12

Can Orion and Looker work together, or should we switch?

Accepted Answer

Yes, and coexistence is the natural starting point. You keep Looker and Conversational Analytics as the governed, in-Looker analytics your team relies on. You add Orion as the analyst. Orion reuses the same LookML, watches metrics across the whole stack, investigates what changed, and delivers the answer to Slack or email. What tends to happen from there: as Orion answers more of the day-to-day questions, fewer of them start with someone opening a dashboard. Because Orion already understands your LookML, how much you lean on each tool stays your decision, on your timeline, with no forced migration.

	Orion	Looker CA
Proactive investigation	Auto-detects a significant change, investigates the root cause, and delivers a written narrative to Slack or email	Agentic workflows, with optional key-driver analysis of the triggering change^↗
Everyone in one live conversation	Everyone posts into one thread and watches the same answer stream, each at their own access level. The Room stays the single record, and nobody can copy it into a private thread	Per-user chat. Google documents Explore conversations as typically limited to one user, and not shareable^↗
Turns a chat into a dashboard	Builds dashboards, reports, and slide decks from a plain-language request. Each dashboard has filters, and you can click a number to see why it moved	No chat-to-dashboard path documented. Results open in an Explore^↗
Reaches your whole stack	Reasons across BigQuery, Snowflake, Databricks, Redshift, and the other warehouses you connect, in one investigation	One Explore per conversation, up to five per agent. Or Google sources through its API^↗
Reaches people who are not Looker users	Connects through one dedicated Looker user, with optional read access to the LookML repository. The people asking need no Looker account of their own	Each person asking must be granted Looker permissions of their own^↗
Delivered where you work	Sends the written answer to Slack or email. Ask questions and start analyses from any Slack channel	Workflow alerts reach email, Slack, and the Looker mobile app^↗
First-party Google integration	Works inside Gemini Enterprise, Claude, and Cursor. Delivers to Slack or email	One vendor across Looker, Gemini Enterprise, and Google Chat, governed by your Google contract^↗
Governed exploration inside Looker	Works outside Looker, on your LookML. Explores stay in Looker	Native to the Looker experience your analysts already govern^↗

Orion vs Looker Conversational Analytics

TL;DR

What is the difference between Orion and Looker Conversational Analytics?

How do Orion and Looker Conversational Analytics compare?

What does your LookML buy you outside Looker?

When should you choose Looker Conversational Analytics instead?

Orion vs Looker Conversational Analytics: what do buyers ask most?

More Orion comparisons