AIBM RESEARCH

Software pricing studies, checkable down to the row

Original studies on how business software is priced and how it behaves when you use it. Every study states its population, period and denominator, ships as a fixed edition with its data, and keeps its earlier editions and corrections in view. Quote the numbers; check them first if you like.

Press & data: contact, reuse, correctionsHow current is each study?Open dataset

Market observations

Counts over a defined population of tools, vendors or AI answers. Each card shows the edition it describes.

Market observationEdition 2026-10Corrected 2026-10-08
Panel: software terms, month by month
2 of 78 panel tools with a confirmed price change in the last reviewed round (2026-09 to 2026-10)
78 fixed panel tools · measured 1 August 2026 to 7 October 2026 · next re-check monthly: capture on the 1st-2nd, review of the differences, then a new edition
Market observationEdition 2026-10-07
Report 01: SaaS Pricing Transparency
398 of 513 of the researched tools carry at least one documented cost caveat
513 researched tools · measured 1 July 2026 to 7 October 2026 · next re-check monthly: a new edition is computed at most once per calendar month
Market observationEdition 2026-10-01Corrected 2026-10-07
Report 02: who do AI engines cite for software pricing?
73 of 227 citations in AI answers about software prices pointed at the vendor's own site
40 answers, 20 tools · measured 1 October 2026 · next re-check monthly: new run on the 1st
Market observationEdition 2026-09-04Corrected 2026-10-07
Report 03: which plan do you need for a vendor's MCP server?
42 of 74 of the vendors with a first-party MCP server state no plan, tier or price for it in their own docs
425 vendors checked · measured 17 August 2026 to 4 September 2026 · next re-check 16 November 2026: full sweep of every partner
Market observationEdition 2026-07_2026-10Corrected 2026-09-15
Report 04: how often does a SaaS pricing page change its price?
37 of 753 monthly page comparisons showed a price change (all change types together)
753 monthly page comparisons · measured 2 July 2026 to 2 October 2026 · next re-check monthly: capture on the 2nd, published after review
Market observationEdition 2026-10-01
Report 05: do AI answers get software prices right?
11 of 25 assessable AI answers stated at least one price we could not find on the vendor's pricing page that morning
25 assessable answers · measured 1 October 2026 · next re-check monthly: AI run on the 1st, compared with the same day's capture

Hands-on tests

The same task, the same instructions, checks written down before the first run. One run per tool is a case study, not a ranking; repeat runs become new editions.

Hands-on testEdition 2026-10-01
Hands-on: Clonable on WordPress
15 seconds at most for a price edit on the English page to appear on the German clone
1 site, 1 language · measured 1 October 2026 · next re-check re-test after a relevant product change; repeat runs are planned as a protocol series
Hands-on testEdition 2026-09-30
Hands-on: Active Intelligence 3.0 on a new ActiveCampaign account
1 of 2 campaign drafts contained a sentence our source page does not support
11 tasks, 1 account · measured 30 September 2026 · next re-check re-test after a relevant product change; repeat runs are planned as a protocol series
Hands-on testEdition 2026-09-28
Hands-on: one pitch deck, four AI presentation tools
1 of 4 tools changed the funding ask everywhere from one prompt (check C09)
4 tools, 1 run each · measured 28 September 2026 · next re-check re-test after a relevant product change; repeat runs are planned as a protocol series
Hands-on testEdition 2026-09-26
Hands-on: one booking app, three AI builders
2 of 3 first versions let ten simultaneous bookings into an eight-spot class (check T05)
3 builders, 1 run each · measured 26 September 2026 · next re-check re-test after a relevant product change; repeat runs are planned as a protocol series

How current is each study?

A rebuild of a page is not a new measurement. This table shows when each study was last verified and when the next full re-check is due.

StudyTypeCurrent editionLast verifiedNext full re-checkCorrections
Panel: software terms, month by monthMarket observation2026-107 October 2026monthly: capture on the 1st-2nd, review of the differences, then a new edition2
Report 01: SaaS Pricing TransparencyMarket observation2026-10-077 October 2026monthly: a new edition is computed at most once per calendar month0
Report 02: who do AI engines cite for software pricing?Market observation2026-10-011 October 2026monthly: new run on the 1st1
Report 03: which plan do you need for a vendor's MCP server?Market observation2026-09-044 September 202616 November 2026: full sweep of every partner1
Report 04: how often does a SaaS pricing page change its price?Market observation2026-07_2026-107 October 2026monthly: capture on the 2nd, published after review1
Report 05: do AI answers get software prices right?Market observation2026-10-011 October 2026monthly: AI run on the 1st, compared with the same day's capture0
Hands-on: Clonable on WordPressHands-on test2026-10-011 October 2026re-test after a relevant product change; repeat runs are planned as a protocol series0
Hands-on: Active Intelligence 3.0 on a new ActiveCampaign accountHands-on test2026-09-3030 September 2026re-test after a relevant product change; repeat runs are planned as a protocol series0
Hands-on: one pitch deck, four AI presentation toolsHands-on test2026-09-2828 September 2026re-test after a relevant product change; repeat runs are planned as a protocol series0
Hands-on: one booking app, three AI buildersHands-on test2026-09-2626 September 2026re-test after a relevant product change; repeat runs are planned as a protocol series0
The machinery behind the studies: the Atlas · Price Watch · open dataset · MCP server · how we review

AIBuilder Marketplace earns affiliate commissions on some tools it covers. Every study names its commercial relationships; partners and non-partners are counted by the same rules, and per the manifesto we may earn from tools our research criticises — it never changes the numbers.