Decide · The money conversation
Noise suppression license costs at seat scale
The per-seat price is the number vendors put on the page. The number you have to defend is that price multiplied by a seat count you chose, over a term you agreed to, minus the seats nobody ended up using. This page does all four of those steps in public so you can walk into the budget conversation with arithmetic instead of a vibe.
Annual spend, computed
Twelve months of licensing at each seat count. Run python3 tools/license-math.py in this site's repository and you will get these exact rows — that is the point of shipping the script rather than the spreadsheet.
| Seats | Core, annual billing | Core, month-to-month | Advanced, annual billing | Advanced, month-to-month |
|---|---|---|---|---|
| 25 | $2,400 | $4,800 | $4,500 | $9,000 |
| 50 | $4,800 | $9,600 | $9,000 | $18,000 |
| 100 | $9,600 | $19,200 | $18,000 | $36,000 |
| 250 | $24,000 | $48,000 | $45,000 | $90,000 |
| 500 | $48,000 | $96,000 | $90,000 | $180,000 |
| 1,000 | $96,000 | $192,000 | $180,000 | $360,000 |
Twelve months of licensing, by seat count
Core
Advanced
Per seat per year, which is the unit finance will actually ask for: Core annual $96, Core month-to-month $192, Advanced annual $180, Advanced month-to-month $360.
One seat, for one year
The two mistakes that double the bill
1. Not flipping annual billing
Annual billing is 50% off the month-to-month rate on both Core and Advanced. At 100 seats that is $9,600 a year versus $19,200 — the same software, the same deployment, the same outcome, for double the money.
It is the most expensive checkbox in this entire project, and it gets missed for a mundane reason: pilots start month-to-month because that is sensible, and nobody goes back and changes it when the pilot converts. Put “switch to annual” on the conversion checklist, not in your memory.
2. Buying a seat for every employee
Headcount is the wrong denominator. The right one is people whose job involves talking to other people through a computer, which in most organizations is a large minority rather than everyone.
The difference between 250 seats and 500 seats on Core annual is $24,000 a year. That is real budget, spent on filtering the background noise of people who attend two meetings a month with their microphone off.
Scoping seats honestly
Work down this list with the people who own each group, not from an org chart in a vacuum:
- Tier 1 — buy without debating. Customer-facing voice roles: support, inside sales, account management, recruiting, anyone whose calendar is calls. These people generate the complaints and the revenue exposure.
- Tier 2 — buy if the budget allows. Managers and ICs in hybrid or fully remote roles with heavy internal meeting loads. Real benefit, less measurable.
- Tier 3 — usually don't. Deskless staff, people in controlled office environments who rarely take calls, and anyone whose meeting participation is passive. The platform-native levers cover them adequately and cost nothing.
- Then add 10%. Not waste — headroom. New hires, role changes, and the person in tier 3 who turns out to work next to a construction site. A rollout that can't absorb an exception generates a purchase order for every exception.
The waste math nobody runs
Assume a 250-seat purchase on Core annual: $24,000/year. Now assume what actually happens without provisioning automation — people leave, roles change, and nobody reclaims the seat. An idle-seat rate of 15% by month twelve is not pessimistic for a tool that nobody's offboarding checklist mentions.
That is roughly $3,600 a year paying for licenses attached to people who no longer work there. It is also, conveniently, the cheapest possible argument for doing the identity work: SCIM deprovisioning turns that leak into a non-event, which is one of several reasons the SSO & SCIM page argues for doing identity before installers. Below the tier that includes SCIM, a calendared quarterly reconciliation is your compensating control — and it has to be calendared, because the failure mode is forgetting, not deciding.
The leak, drawn to scale
250 seats on Core annual: $24,000 a year. Idle by month twelve: 15%, three laptops in every twenty — roughly $3,600 a year.
If you run a contact desk, price the other line
The Call Center product is a separate line with separate economics: CC Core at $15/agent/month billed annually — $180 per agent per year, or $9,000/year for a 50-agent desk. Pricing a support organization against the Meeting AI tiers understates what you need and overstates what you'll get; if agents are your population, quote the agent product.
The agent line, multiplied out
$15per agent, per month, billed annually$180per agent, per year$9,000a 50-agent desk, per year
The thing that costs nothing
Before any of the above: configure the platform levers you have already paid for. Zoom's administratively-set suppression level and Teams' voice-isolation policy are included in licenses you hold today, and for a single-platform fleet they may end this project at a cost of one afternoon. No affiliate relationship exists between this site and Microsoft or Zoom, and we would still put this paragraph above the pricing table if we could. The mechanics: Zoom admin audio settings and Teams noise-suppression policy.
Four questions that shrink the number
- Did the trial actually beat our configured baseline? Not “was it good” — was it better than Teams and Zoom properly policy-configured? If nobody configured the baseline first, the comparison is meaningless and you are probably buying a control group.
- Are we paying the Advanced delta for features we will govern? At 100 seats that delta is $8,400/year. Name the feature, name the owner, or buy Core.
- What is our renewal date, and who owns the 30-day review? A name and a calendar entry, set the day you sign.
- What happens at 80% adoption? Decide in advance whether that triggers a comms push or a seat reduction. Deciding afterwards means doing neither.
Price it after the pilot, not before
Krisp's 7-day trial needs no card and no procurement conversation (verified 20 September 2026), which makes it the only honest input to every number on this page. Run it on the people in tier 1, measure against your configured platform baseline, and let the seat count come out of the evidence.
Run the trial, then do the math
Referral link, declared: when an organization licenses Krisp after clicking this button, Krisp pays RolloutDesk a fee, and the quote you receive is the one you would get going direct. We are aware that a page arguing you should buy fewer seats is a strange thing for an affiliate to publish. It is also the only version of this page worth reading.
Related: the company-wide playbook, what each tier gives an admin, and the pilot plan that produces the evidence these numbers depend on.