The short answer
Above a hundred people, the deciding factor is usually not which product searches best. It is which product does not reprice when you cross a threshold. Dropbox Dash moves from $15 to $35 per user per month at 101 users. Microsoft Copilot Business stops at 300 users, after which you move to a $30 add-on that requires a base plan underneath it, making the real seat $69 on E3. Gemini Enterprise caps self-serve accounts at 25 seats. Amazon Quick runs the other way, spreading a fixed $250 per account per month across more seats as you grow. Glean publishes no price at all. Get the threshold, not just the rate, onto the order form.
Last updated: August 2026. Every figure re-read from the vendors' own pricing pages and documentation between August 21 and August 31, 2026.
Software usually gets cheaper per seat as you buy more of it. That is the whole logic of volume pricing: the vendor trades margin for commitment. Enterprise search mostly does not work that way, and buyers who assume it does get an unpleasant surprise in year two.
The pattern across this category is that pricing is tiered rather than discounted, and the tier boundaries are set at headcounts that successful rollouts pass. A pilot for forty people is cheap everywhere. The same tool for a hundred and forty is where the vendors diverge, sometimes by a factor of two on the same product. This is a comparison written for that second purchase, because it is the one with real money attached and almost nothing published about it.
What happens to the price when you cross the line?
Here is every threshold we can document from a vendor's own materials, and what crossing it does. Where a vendor publishes nothing, the row says so rather than repeating a figure from a procurement blog as though it were a list price.
| Product | The threshold | What happens when you cross it |
|---|---|---|
| Dropbox Dash | 100 users | Teams at $15 per user per month becomes Business at $35, billed yearly only. An $18,000 bill becomes $42,420 |
| Microsoft Copilot Business | 300 users | You move to the $30 Microsoft 365 Copilot add-on, which is not standalone. With E3 at $39 the honest seat is $69, with E5 at $60 it is $90 |
| Gemini Enterprise | 25 seats self-serve, 1,000 invoiced | The price does not change, the purchasing path does. You place an offline order through a Google Cloud representative. The Business edition is separately documented as "1-500 users" |
| Amazon Quick Professional | No seat threshold | Runs the other way. A fixed $250 per account per month spreads across more seats, so the effective rate falls from $45 at ten users to $21 at 250 |
| Amazon Q Business | Not available at any size | Closed to new customers on July 31, 2026. Existing applications keep bug fixes and security updates, but AWS states new feature requests will no longer be considered |
| Glean | Seat minimums, not published | No list price exists. Procurement platforms report averages between roughly $98,890 and $142,019 a year, from negotiated deals of unknown size. Secondary sourcing only |
| Dust | 3 connectors on Business | The cap is on integrations rather than people, and prices are quoted in euros. Credits do not roll over |
| Dashworks | 10 seat minimum on Business | Team is $10 per seat yearly with no minimum, Business $12 yearly with a floor of ten. The cheapest published rate in this comparison |
| Onyx | Permission sync is Enterprise only | Business is $20 per user per month annually, and it is MIT licensed if you self-host. Permission-syncing connectors sit behind the Enterprise Edition |
| InSearch | No headcount threshold | $20 Team, $32 Business, $49 Scale per user per month billed annually. The rate does not change at 100 or 300 people |
We sell one of the products in that table, so read it with that in mind. What we have tried to do is put the threshold in the same row as the price, because every one of these vendors publishes both and almost no comparison shows them together.
How much should 250 seats of enterprise search cost?
Between roughly $36,000 and $207,000 a year at list, depending entirely on which product and which tier. The spread is not about quality, it is about whether the product bundles a base plan and whether it re-tiered on the way up. Here is the arithmetic at a single headcount, so the tiers are comparable.
| 250 seats, list price | Rate | Annual |
|---|---|---|
| Dashworks Business | $12 yearly | $36,000 |
| Onyx Business | $20 | $60,000 |
| InSearch Team | $20 | $60,000 |
| Amazon Quick Professional | $20 plus $250 per account per month | $63,000 |
| InSearch Business | $32 | $96,000 |
| Dropbox Dash Business | $35 | $105,000 |
| Amazon Quick Enterprise | $40 plus $250 per account per month | $123,000 |
| InSearch Scale | $49 | $147,000 |
| Microsoft 365 Copilot on E3 | $30 add-on plus $39 base | $207,000 |
Two honest observations about our own position in that list. Dashworks at $12 a seat is cheaper than anything we sell, and Onyx matches our Team tier while also being MIT licensed if you have the engineers to run it. Neither of those is a reason to buy from us. The Microsoft line is also not quite a fair fight, because most companies buying Copilot already pay for E3 or E5 and would not book the base plan against the search project. It is in the table because the seat cost is real money even when it is already in someone else's budget line.
Why does enterprise search get more expensive per seat as you grow?
Because most vendors in this category sell two products under one name rather than one product at two volumes. The lower tier is a self-serve product for small companies: credit card, trial, no implementation. The upper tier is a supported enterprise product with admin controls, directory features and a named contact. Pricing them as separate products is defensible. It just produces a staircase instead of a discount curve, and the step is placed where the vendor believes a company stops being small.
The practical consequence is that your headcount forecast is a pricing input, and most teams do not treat it as one. If your plan for next year takes you from 85 to 115 people, you are not buying the $15 product, you are buying the $35 one with a year of the cheaper rate attached. Model both sides before the first invoice, not at renewal.
Which enterprise search tools have seat limits or caps?
Four of the products here document a hard limit rather than a price change. Gemini Enterprise caps self-serve and resold accounts at 25 seats and invoiced accounts at 1,000, with the Business edition separately documented as covering 1 to 500 users and the Frontline edition unavailable until you already hold 150 Standard or Plus seats. Microsoft Copilot Business supports up to 300 users. Dropbox Dash for Teams stops at 100. Dashworks Business sets a floor rather than a ceiling, at ten seats.
A fifth belongs on that list as of September 2026. Anthropic documents the Claude Team plan as being for teams of 2 to 150, and that is the plan bundling Claude Cowork at $20 a seat billed annually. Above the ceiling you move to Enterprise, where the seat is $20 plus usage billed at API rates, so the model changes shape rather than just the rate. The full breakdown is in Claude Cowork pricing, plans and admin controls.
A sixth landed on September 2, 2026, and it is the strictest wording of the lot. OpenAI caps a ChatGPT Business subscription at 200 paid Standard and Premium seats, counts unassigned seats against that total, and states that its support team is unable to change the maximum. At the ceiling you cannot buy a seat or even approve a join request that needs one. The crossing math is in the ChatGPT Business 200 seat cap and what Enterprise costs at 201 seats, and the published figures sit on our ChatGPT Enterprise pricing page.
Limits behave differently from prices in one important way: you cannot negotiate around a console that will not sell you a twenty-sixth seat. A price is a conversation. A cap is a procurement dependency with a lead time attached, and it lands on the week you promised the rollout would finish. When you scope a pilot, ask the vendor what happens at the boundary and get the answer in writing.
What should we check before signing a multi-year deal?
Five things, in the order they tend to cause problems.
- The threshold and what sits on the other side of it. Get both rates on the order form, not just the one you are paying today.
- Whether the sticker is the whole bill. Amazon Q Business bills index capacity hourly whether or not anyone searches, roughly $193 a month for one enterprise unit before a single seat. Amazon Quick adds $250 per account per month on its paid tiers. Microsoft 365 Copilot needs a base plan underneath it. A per-seat number that omits these is not a quote.
- Eligibility, before price. Some products will not sell to you regardless of budget. Dropbox restricts Dash team deployments to teams located in the United States, United Kingdom or Australia, requires the account language to be English, and lists accounts that require HIPAA compliance as not supported.
- Whether limits are pre-announced. Dropbox states that plans may in future include a limit on how many items each member can use, with at least 30 days of notice, and counts emails, calendar invites and meeting recordings as items. That is unusually candid, and it is still an open variable on a three-year term.
- Who actually renews it. Search tools are bought by IT and renewed by finance, and the two rarely see the same numbers. Whoever owns the renewal needs the threshold math in a form they will read, which usually means turning the cost model into a few clear slides rather than forwarding a spreadsheet nobody opens.
So what is the best enterprise search software for a team over 100 users?
There is no single answer, but the shortlist narrows fast once you apply the thresholds. If you are between 100 and 300 people and already standardized on Microsoft 365, Copilot Business is genuinely hard to beat on price until the 300 user cap, because the base plan is already paid for. If you want the widest connector list and a governance console alongside search, and you are under a hundred people in a supported country, Dropbox Dash at $15 is the best value in this comparison and we will say so plainly.
If you are past a hundred people, running a mixed stack rather than a single vendor's suite, and what you actually need is one search box that answers questions across Drive, Slack, Notion, Confluence, Gmail, Jira and Salesforce with citations back to the source documents, then a focused tool priced on headcount alone is the simpler purchase. That is what we built, with item-level permissions inherited from each source and enforced at query time, no training on your content, and a published per-seat rate that does not re-tier.
The detail behind each vendor sits on its own page. Dropbox Dash pricing covers the 100 seat step and the deployment restrictions in full, Microsoft Copilot pricing untangles the add-on and the 300 user cap, Gemini Enterprise pricing covers the editions and pooled quotas, Amazon Q pricing separates the three AWS products sharing that name, and Glean pricing explains what can and cannot be said about a vendor that publishes nothing. If you are still building the shortlist rather than pricing it, enterprise search tools is the wider survey and unified search explains the product category itself.