The short answer
Only one of these is still available to new buyers. AWS closed Amazon Q Business to new customers on July 31, 2026, one day after it stopped accepting new Amazon Kendra customers. Existing Q Business customers remain fully supported with bug fixes and security updates, but AWS has said new feature requests will no longer be considered, and it points new buyers at Amazon Quick instead. Atlassian Rovo is actively developed and is included at no separate charge in paid Standard, Premium and Enterprise Cloud plans of Jira, Confluence and Jira Service Management, rationed by a pool of AI credits rather than sold per seat. So for most teams reading this in 2026 the real comparison is Rovo versus Amazon Quick, and the choice between them comes down to where your content lives: Rovo is strongest inside Atlassian, and the AWS products are strongest when your content is already in S3 and your team writes IAM policies for a living.
Last updated: August 2026.
These two products get compared constantly, and the comparison is usually written as though both were on the shelf. They are not. That single fact reorders the whole evaluation, so it is worth dealing with before any feature table.
Can you still buy Amazon Q Business?
No, not as a new customer. AWS closed Amazon Q Business to new customers on July 31, 2026. If your account was not already using it before that date, it is not something you can sign up for, and no amount of feature comparison changes that.
Existing customers are in a more comfortable position than the headline suggests. AWS has been clear that Q Business remains fully supported, with continued bug fixes and security updates, and it has published no shutdown date, no end-of-life and no forced migration deadline. What it has said is that new feature requests will no longer be considered. That is the sentence to plan around. A product that is supported but frozen keeps working and stops improving, while the category around it moves quickly.
The migration path AWS points at is Amazon Quick, using Bring Your Own Index so that an existing Q Business index can be connected to Quick without tearing anything down. It is not a clean lift and shift: Q Apps have to be rebuilt manually as Quick Flows, and Guardrails and User Store configurations have to be recreated. We covered the details of that transition separately in Amazon Q Business vs Amazon Quick Suite.
Amazon Q Business vs Atlassian Rovo compared
| Amazon Q Business | Atlassian Rovo | |
|---|---|---|
| Can you buy it today | No. Closed to new customers July 31, 2026. AWS directs new buyers to Amazon Quick | Yes. Included with paid Jira, Confluence and JSM Cloud plans |
| Status for existing users | Fully supported, bug fixes and security updates continue. New feature requests no longer considered. No published EOL | Actively developed across Search, Chat, Studio and agents |
| Price model | Per user. Lite $3 per user per month, Pro $20, plus hourly index units billed separately | No separate charge. Bundled into the plan, then rationed by a monthly pool of AI credits |
| What runs out | Nothing usage-based on the seat. You pay for index capacity whether or not anyone searches | Credits. The pool is org-wide, resets monthly and does not roll over, so heavy users spend the light users allowance |
| Successor cost | Amazon Quick starts at $20 per user per month plus a flat $250 per account per month infrastructure fee on the paid tiers | Not applicable. Rovo is the current product |
| Strongest where | Content already in AWS, S3 and enterprise repositories, with a team fluent in IAM | Content in Jira, Confluence and Jira Service Management |
| Weakest where | Frozen roadmap, and a migration to Quick that is not automatic | Outside Atlassian. Slack, Drive, Notion, Gmail and Salesforce are where most answers actually sit |
| Setup effort | An AWS project: IAM, Identity Center, connectors, index sizing | Close to zero if you are already on a paid Atlassian Cloud plan |
How much does Amazon Q Business cost compared to Atlassian Rovo?
They do not price the same way, so a per-seat comparison is misleading in both directions. Q Business charges per user, at $3 per user per month for Lite and $20 for Pro, and then separately for the index capacity that holds your content, billed by the hour whether anyone searches or not. Rovo charges nothing extra at all and instead caps how much you can use.
The sharpest number in this whole comparison belongs to the successor rather than either product. On the AWS 5,000-user example, moving from Q Business to Amazon Quick takes the bill from roughly $33,000 a month to roughly $100,250, because the cheapest paid Quick tier is $20 per user per month and adds a flat $250 per account per month infrastructure fee. Anyone treating the Quick migration as a like-for-like continuation should price it before committing.
Rovo looks free next to that, and for teams already paying for Atlassian it very nearly is. The catch is the credit pool, which Atlassian describes but does not quantify on its pricing page. Widely reported figures put the allowance at 25 credits per seat per month on Standard, 70 on Premium and 150 on Enterprise, with a chat or agent request costing around 10 credits. At Standard, a 50-person company therefore shares roughly 125 requests a month across the entire organization. We work through that arithmetic, and label carefully what Atlassian publishes versus what only third parties report, on our Atlassian Rovo pricing breakdown. The comparable vendor breakdowns are Amazon Q Business pricing and Gemini Enterprise pricing.
Which one searches more of your company?
This is the question that decides most real evaluations, and neither product answers it well, because both are extensions of a platform rather than neutral search layers.
Rovo is excellent at Atlassian and progressively less useful the further you get from it. If the decision you are looking for was made in a Confluence page and tracked in a Jira ticket, Rovo will find it and explain it. If the decision was actually made in a Slack thread, promised to the customer over Gmail, and only summarized in Jira afterwards, Rovo sees the summary. It returns a confident partial answer, which is a worse failure than returning nothing, because the person reading it has no way to know what was missing.
Q Business inverts the bias. It is strong on content that already lives in AWS and in traditional enterprise repositories, and it assumes a team comfortable with IAM, Identity Center and index sizing. That is a real advantage for a company whose documents sit in S3 and whose security model is already expressed in AWS policy. It is a real cost for a company whose knowledge lives in SaaS apps and whose IT team does not write IAM policy for a living.
Both vendors are also betting that the near future is agents rather than search boxes, Rovo through Studio and agents, AWS through Q Apps and now Quick Flows. That is probably right in the long run, and it is worth noting that a whole market of ready-made agents you can browse by category has grown up around that bet, so building one from scratch is rarely the first move. It does not change the near-term problem, though. An agent is only as good as what it can read, and both of these read one platform well.
Permissions: the difference that actually matters
Any tool that reads across a company inherits the messiest part of that company, which is who is allowed to see what. The two products handle it very differently, and there is one AWS detail worth knowing before you plan a Quick migration.
In AWS wording, without IAM Identity Center, all Amazon Quick users automatically receive access to connected Q Business indexes. Read that twice if you are migrating. It means the identity configuration is not a detail to tidy up later, it is the control that decides whether your search tool quietly becomes a company-wide document leak. Teams that ran Q Business without Identity Center and move to Quick need to fix that first, not after go-live.
Rovo inherits Atlassian permissions, which is a genuine strength inside Atlassian because that permission model is one system rather than seven. The limitation is scope rather than correctness: it can only enforce permissions on content it can see, and most of your content is not in Atlassian.
The general test to apply to any vendor here, including us, is when permissions are evaluated. If access is copied into an index on a sync schedule, then the sync interval is your revocation window, and someone removed from a project this morning may still be able to surface its documents this afternoon. If access is checked at the moment of the query, that window does not exist. Ask which one a vendor does and get the answer in writing.
Which should you choose?
If you are an existing Q Business customer, do not panic and do not rush. Nothing is breaking, there is no deadline, and the product is supported. Do take the freeze seriously in your planning: keep contracts short, price the Quick migration properly including the $250 per account monthly fee, and fix Identity Center before you move rather than after.
If you are a new buyer, Q Business is not an option, so the honest shortlist is Amazon Quick, Rovo if you are an Atlassian shop, and cross-app search tools. Choose Rovo if your company genuinely runs on Atlassian and your questions are about tickets and pages. It is included in something you already pay for, and no third-party tool competes with already-paid-for on its home ground.
Choose a cross-app tool when your answers are spread across Slack, Google Drive, Notion, Gmail and Salesforce as well as Jira, or when you want predictable per-seat cost instead of a metered credit pool that punishes the teams who adopt it most enthusiastically. That is the case InSearch is built for: one query across every connected app, a written answer with inline citations back to the source documents, and item-level permissions re-checked at query time for whoever is asking. If you want to see the wider field first, the shortlist with a price column is on enterprise search tools, and Atlassian Rovo alternatives goes deeper on the Rovo comparison specifically. Teams whose requirement is simply one answer across everything should start at unified search.
One closing note on method, because it applies well beyond these two products. The most decision-relevant fact in this comparison was not a feature, a benchmark or a price. It was a purchasing status buried in a documentation page. Before you spend a week on a feature matrix, spend ten minutes confirming that both products are actually for sale, that neither is in maintenance mode, and that the vendor is still accepting feature requests. That check would have saved several teams a quarter this year.