The short version. Guru pricing is no longer published. The page now offers a package built around your scale, knowledge complexity and AI maturity, with a call booking rather than a rate card. The quote is driven by seats, connector count, assistant usage and how much implementation support is bundled. Price the vendors who still publish first so you have a baseline, and be clear whether you are paying for verification, which is what Guru is genuinely best at.
Guru pricing is now a conversation
Worth stating plainly, because a great many comparison articles still list per-user tiers that no longer appear on the vendor’s site.
We checked on 17 October 2026. The pricing page describes a package built around your scale, knowledge complexity and AI maturity, with expertise to make it work from day one. The call to action is booking a call with their team.
There are no currency figures on it. The product has repositioned as an enterprise AI knowledge solution, and the commercial model moved with the positioning.
Document360 made the same change during 2026 and Glean has no pricing page at all, so this is a category trend rather than one vendor’s decision.
What the package now includes
The change is not only about the number, and this is the part that matters for comparison.
The current packaging bundles the platform with solution engineering and implementation support. You are buying the knowledge layer and a team to make it work, rather than software you configure yourself.
For a large organisation with a complex estate that is genuinely valuable, and it explains why a rate card stopped being meaningful. For an internal IT team with eighty runbooks and two maintainers it is a considerable amount of service attached to a product that would otherwise be configured in a week.
So when you compare Guru pricing against a published alternative, you are not comparing like with like, and it is worth saying so explicitly in any internal recommendation.
What drives the quote
Four variables, and having your numbers ready shortens the first call noticeably.
Seats. How many people can read answers, which for internal knowledge is usually the whole technical team and often beyond it.
Connector count. How many systems the knowledge layer reads from, since each one is work on their side.
Assistant usage. Answering on top of the knowledge base is typically metered separately from the content itself, and usage-based elements are the ones most likely to move after signature.
Implementation scope. How much of the solution engineering you take, which is now part of the product rather than an optional extra.
Build a baseline from vendors who publish
Five publish openly, and pricing them all takes about twenty minutes.
Helpjuice publishes by included users rather than per seat, at 249 per month for 30 users, 449 for 100 and 799 for unlimited. Its page carries no currency symbol anywhere.
Slab is free to 10 users, then Startup at $6.67 and Business at $12.50 per user per month, both billed annually.
Tettra publishes $8 per user per month on Scaling, with a ten user minimum and yearly billing, and answers inside Slack.
Notion is free, then $10 Plus and $20 Business per member per month.
Confluence is a configurator. At 25 seats Standard is $7.15 per user per month and at 300 seats it is $5.78, so any figure without a seat count describes a different company.
What you are actually paying for
One capability justifies the premium, and it is worth testing whether you use it.
Verification is the strongest thing Guru does. Content carries an owner and a review cycle, unverified material is visibly unverified, and that surfaces staleness, which is a failure that otherwise produces no alert at all.
Before accepting a quote, open the product and check how much of your content is currently verified, when the last review cycle completed and how many owners responded to it.
A team with most content verified and owners responding within a week is genuinely relying on the feature. A team that finds review cycles were switched off after the first quarter is paying for a capability it abandoned, which changes the comparison substantially and is more common than anyone admits in a renewal meeting.
Replacing verification with a process
If you move to a cheaper published option, this is the work that replaces what you give up, and it costs nothing.
Put an owner’s name on every document, visible to readers rather than buried in revision history, so anybody who spots an error knows who to tell.
Put a last-reviewed date beside it, so readers can judge for themselves whether to trust the page.
Split the library into quarterly and annual review, with the things people ask about most in the quarterly set, and create one recurring calendar entry per owner. An intention to review is not a mechanism; an interruption is.
And stamp the page even when nothing changed, because a visible review with no edits is still useful information to a reader.
The narrower option
If what you want is answers from documents rather than a governed knowledge platform, there is a smaller shape.
An answering layer reads documents where they already sit and returns a sentence with a link to the source, with no implementation project.
Matram is one of these, and it is owned by the same people who publish this site, which you should weigh against this section. It publishes flat pricing at $29, $69 and $199 per month with unlimited seats, which in a category where several vendors have stopped publishing is worth noting in itself.
Its limits are specific. There is no free tier, only a 30 day trial without a card. It answers rather than actioning tickets, so it will state what the runbook says and will not run the procedure. And it is not an ITSM or service desk platform, so there is no ticket queue, routing or case management in it.
It also has no verification workflow, which is the direct trade here. You would gain answering and lose the mechanism that makes staleness visible, so the ownership process above has to come from you.
Questions for the first call
Six, and asking the first early tells you a lot about the weeks ahead.
What is the range at our seat count and connector list. How much of the package is implementation rather than software, and can we take less of it. How is the assistant metered. What is the minimum term. What happens to the price when we add a connector. And what does the product do when the content does not contain an answer, which separates these tools and which almost nobody asks.
A vendor that gives a range in the first call is not necessarily cheaper, and it is reliably faster to evaluate.
Why can I not find Guru pricing any more?
Because it is no longer published. As of October 2026 the page describes a package built around your scale, knowledge complexity and AI maturity, with a call booking rather than a rate card, and there are no currency figures on it. Document360 made the same change during the year and Glean has no pricing page at all, so any per-user tiers you find elsewhere describe a previous commercial model.
What drives the quote now?
Seats, connector count, how heavily the assistant is used on top of the knowledge base, and how much implementation support you take. That last one matters because the current packaging bundles solution engineering with the platform, so you are buying software and a team to deploy it rather than software alone. Ask explicitly whether you can take less of the service component.
How do I budget without a rate card?
Price the vendors who still publish and use them as a floor. Helpjuice, Slab, Tettra, Notion and Confluence all list their rates openly, and pricing all five at your headcount takes about twenty minutes. That gives finance a defensible number and turns the vendor conversation into a comparison rather than an open-ended discovery session run on their terms.
Is the verification feature worth the premium?
It depends entirely on whether you use it, and that is checkable. Open the product and look at what proportion of content is currently verified, when the last review cycle completed and how many owners responded. Heavy use makes it the strongest argument for staying. Review cycles switched off after the first quarter means you are paying for an abandoned capability, which is common and changes the comparison substantially.
What replaces verification if we move?
A process, and it costs nothing. Put a visible owner name and last-reviewed date on every page, split the library into quarterly and annual review sets, and put a recurring calendar entry in each owner’s diary. Stamp pages even when nothing changed. That gets most of the benefit, and teams who skip it find the library stale a year later with nobody aware, because staleness never announces itself.
Should we stay or move?
Stay if verification is load-bearing and you can demonstrate you use it, because nothing in the published alternatives replaces it cleanly and the process substitute requires discipline. Move if what you actually use is a searchable library with chat answering, since several published options cover that for considerably less. Either way, price the published vendors first so the decision is made against a number rather than a feeling.
