Cauliflower
Cauliflower is an EU-hosted research analytics platform that turns SPSS, Excel, CSV and verbatim data into one governed research model, then builds dashboards, AI text analysis and natively editable PowerPoint reports on top of it.
What is Cauliflower?
Cauliflower is a research analytics platform built in Hamburg by Cauliflower GmbH & Co. KG, and sold under a category the vendor coined for itself: an Insights Operating System. The argument behind that label is worth stating, because it explains the whole product. Research data, Cauliflower contends, is not complex because of its size but because of its methodology — item batteries, questionnaire routing, waves, weightings, subsamples, base sizes, significance, implicit brand hierarchies. Generic BI tools and general-purpose language models can store, aggregate or paraphrase such a file, but they do not recognise what any given variable methodologically means.
So the platform starts by mapping a finished dataset into a shared research model. You upload SPSS, Excel or CSV files, connect sources through an API, or feed in verbatims and public reviews from Google Maps, Trustpilot, the App Store and Google Play. A guided setup assigns columns, roles, weightings and recodings. From that single structure five outputs are generated, all resting on the same definitions: dashboards with tracking and benchmarks; text analysis that codes open answers into aspects, detects sentiment and summarises themes against a codebook you own; driver analysis using Shapley values and Kruskal-Wallis tests; an AI agent that answers questions in context; and a natively editable PowerPoint export in your corporate design.
Four places use AI, and the vendor names them: the Surveybot, which asks respondents follow-up questions while a survey is still open; text analysis; Smart Captions, which draft a chart takeaway you can edit, keep or discard; and the research agent, which answers only from your own model and cites the base size, filter and chart behind each response.
Governance is the fourth pillar: five roles from Owner to Guest, free viewer licences, expiring password-protected public links, audit logs and single sign-on on the top tier. Everything runs in the EU — Frankfurt, Gävle and Dublin — with a data processing agreement signed as standard. Access is arranged after a demo; there is no self-service sign-up, and pricing starts at 49 EUR per named seat per month.
What it does
- Import SPSS, Excel, CSV, verbatims or API feeds and map them into one shared research model
- Code thousands of open-ended answers into aspects, sentiment and summarised themes
- Build reusable dashboards that respect waves, weights, base sizes and significance
- Rank what actually drives a KPI using Shapley values and Kruskal-Wallis tests
- Ask an in-context AI agent questions and get answers citing n, filter and source chart
- Export native, editable PowerPoint in your own corporate master template
- Share findings through password-protected, expiring links with role-based access
When to use Cauliflower / When not to
A quick filter to help you decide if Cauliflower is the right fit.
When to use Cauliflower
- Corporate insights teams running tracking studies across several waves, markets and brands
- Market research institutes and agencies producing recurring, presentation-ready client reports
- CX and consumer insights teams buried under thousands of open-ended answers and public reviews
- UX and product research teams that need findings tied back to base sizes and significance
- European organisations bound to EU data residency and a signed data processing agreement
When not to use Cauliflower
- Anyone hoping to sign up and test the tool alone: there is no self-service trial and no free plan
- Occasional or one-off analysis, where 49 EUR per named seat per month is hard to justify
- Teams wanting a full REST API on the entry plan: it only unlocks at the 199 EUR Business tier
- Buyers needing single sign-on, a private tenant, on-premise or audit logs without an Enterprise quote
- Organisations shopping for a survey tool, a panel or fieldwork: Cauliflower analyses data, it does not collect it
How to use Cauliflower
A typical end-to-end flow, from setup to results.
- Request a demo through the contact form or book a 30-minute slot: there is no self-service sign-up
- Bring a real dataset — a live tracker, messy verbatims or a stack of public reviews
- Upload SPSS, Excel or CSV files, or connect a data source through the API
- Work through the guided setup to assign columns, roles, weightings and recodings
- Let Cauliflower map the file into a research model of questions, waves, segments and open answers
- Check bases, weighting, filters and significance before passing any number on
- Analyse: filter, segment, run significance tests, code open ends and rank drivers
- Build dashboards per stakeholder from the same definitions and calculations
- Use Smart Captions and the research agent to draft and interrogate the takeaways
- Export a native editable PowerPoint, or share a password-protected dashboard link
Pros & Cons
Pros
- Survey methodology is handled natively: waves, weights, base sizes, low-base warnings and significance are first-class, not afterthoughts
- Every AI answer cites its base size, filter and source chart, so a number can be traced before anyone quotes it
- An explicit, unconditional commitment that research data never trains Cauliflower's models or the foundation models behind them
- EU hosting documented down to the region — Frankfurt, Gävle and Dublin — with a data processing agreement signed as standard
- PowerPoint export is native and editable rather than a pasted image, removing the step that costs research teams the most time
- An unusually complete pricing page: four tiers, nine credit top-up bands, and plain definitions of a seat and a data point
- AI credits never expire and are shared across plans and workspaces, and viewers cost no seat from the Professional tier up
Cons
- No free plan and no self-service trial: every route into the product goes through a sales demo
- The full REST API is locked behind the 199 EUR Business tier, and no public API documentation exists
- Single sign-on, private tenant, on-premise, your own DPA and audit logs are Enterprise-only, at an unpublished price
- Chat-with-the-Data is a paid add-on at 19 EUR per seat per month and is unavailable on the entry plan
- The platform subprocessor list is shared only on request, as are the DPA, technical measures and security one-pager, under NDA
- No security certification of the vendor's own: the ISO 27001 cited on the security page belongs to the data centres, not to Cauliflower
- The terms of use are published only in German, even at the English address, and neither they nor the privacy policy carries an effective date
Pricing & Plans
There is no permanent free plan and no self-service free trial: access is arranged after a demo. The lowest published price point is EUR 49 per named seat per month, for the Starter OS tier, reduced by 20% on annual billing. Seats are personal rather than concurrent, and viewers who only open a shared dashboard consume no seat from the Professional tier upwards.
- the full analysis platform
- PPTX
- Excel
- CSV and PDF export
- EU hosting and GDPR
- email support
- 500 one-time AI credits per seat and 250
- 000 data points per month
- everything in Starter OS plus your own PPTX master template
- included viewer licences
- the optional Chat-with-the-Data add-on at EUR 19
- and email support with an SLA
- with 2
- 000 credits per seat and 1 million data points per month
- everything in Professional plus multiple PPTX masters
- the full REST API and integrations
- white-label options and a dedicated Customer Success Manager
- with 5
- 000 credits per seat and 5 million data points per month
- everything in Business plus SSO (SAML
- Okta
- Azure AD)
- private tenant or on-premise deployment
- your own DPA
- audit logs and a dedicated technical account manager with an SLA
- with unlimited credits and data points
- AI credit top-ups — nine bands from 1
- 000 credits at EUR 60 to 1
- 000
- 000 credits at EUR 40
- 000
- with volume discounts up to 33% and a further 20% on annual billing
- credits never expire and are shared across plans
Data, GDPR & hosting
A consolidated view of how Cauliflower handles your data.
GDPR overview
GDPR compliance is claimed explicitly and repeatedly, and the detail behind the claim is unusually concrete. The controller is named — Cauliflower GmbH & Co. KG, Fischertwiete 2, Chilehaus A, 20095 Hamburg — with privacy@cauliflower.ai as the data protection contact and the Hamburg Commissioner for Data Protection and Freedom of Information as the competent supervisory authority. Articles 15 to 21 are listed individually, alongside portability under Article 20 and the right to lodge a complaint under Article 77. Legal bases are cited — Art. 6(1)(f) for server logs and hosting, Art. 6(1)(b) for demo requests — and processors are named under Art. 28. A data processing agreement is signed with every customer. Two caveats: the English privacy policy is a courtesy translation whose German original prevails, and neither it nor the terms of use carries an effective date.
Who owns the data?
The German terms of use are explicit: the research results and the data you generate through the platform belong to you. Cauliflower keeps ownership of the site, the software and the underlying technology, and reserves a single carve-out — the right to use anonymised and aggregated data to improve its services. On the text analysis side the vendor states that your codebook stays yours. That carve-out sits in tension with the security page, which promises that research data never trains Cauliflower's models or the foundation models behind them. The two statements are not strictly contradictory, but the boundary between them is defined nowhere on the site.
Reuse rights
Customers keep control of what they produce. Outputs can be exported without asking permission as native editable PowerPoint, PNG, PDF, Excel or CSV, shared internally, or published through password-protected, view-only links that expire. On the vendor's side the use is narrower than the terms alone suggest. AI runs on OpenAI models via Microsoft Azure in an EU region, and the security page commits that research data is not used to train Cauliflower's models or the foundation models beneath them. The assistant answers only from your own research model and dashboards, never the open web, and cites the base size, filter and chart behind every answer. For the website itself the privacy policy names Amazon Web Services EMEA SARL and Amazon SES, both in the EU Frankfurt region, as processors of contact and demo enquiries, and states that the site sets no cookies at all. The one reservation the terms retain is the use of anonymised and aggregated data for service improvement.
Data retention & training
Hosting summary
All infrastructure runs in EU data centres, and the vendor names the regions rather than gesturing at them: Frankfurt in Germany (eu-central-1), Gävle in Sweden (eu-north-1) and Dublin in Ireland (eu-west-1). Data residency is stated as EU only, and customer data processing is said to stay in the EU. Data is encrypted at rest, in transit with TLS 1.2 or above, and backups are encrypted as well. AI inference is not an exception: it runs on OpenAI models through Microsoft Azure in an EU region. A data processing agreement is signed with every customer, and Enterprise customers may deploy into a private tenant or on their own infrastructure on-premise. One qualification is worth carrying: the ISO 27001 certification mentioned on the security page describes the data centres, which are operated by Amazon Web Services, not Cauliflower itself, which publishes no certification of its own.
Things to keep in mind
Risks and trade-offs to weigh before adopting Cauliflower.
- Smart Captions and the research agent draft the takeaway for you; a team that stops reading the underlying chart slowly loses the habit of interrogating its own data
- Source-cited answers feel authoritative, and that confidence is itself the risk: a citation proves where a number came from, not that the question asked of it was the right one
- Automatic aspect coding replaces days of manual work, but the codebook and the judgement behind it still need a human owner, or categories drift silently between waves
- The terms reserve the use of anonymised and aggregated data to improve the services while the security page promises no training on research data; the boundary between the two is defined nowhere
- The site's own structured data advertises an entry price of EUR 19, which is in fact the Chat-with-the-Data add-on; the real starting point is EUR 49 per seat per month
- The site is inconsistent about its own customers: the same 'under 7 days' figure is credited to Tchibo on the homepage and to XING on the customers page
- Registered addresses differ between sources — the imprint gives Fischertwiete 2, the commercial register gives Grimm 17 — so confirm the contracting entity's address before signing
Setup & Integrations
Technical difficulty
Low for everyday use, once you are in. There is no self-service sign-up: access is arranged after a demo. From there, uploading SPSS, Excel or CSV and working through the guided setup — assigning columns, roles, weightings and recodings — needs research literacy, not development skills, and standard onboarding is included in the seat price. One customer reports a workflow live in under seven days. Technical work only appears at the edges: integrating the Surveybot into existing survey software goes through an API, and SSO, on-premise installation and custom integrations are scoped and quoted separately.
Deployment
Integrations
Supported languages
Behind Cauliflower
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What is an Insights Operating System?
Which data formats can I upload?
Does it analyse closed-ended questions too, or only open ends?
How is it different from a BI tool or a general-purpose LLM?
Is there a free trial or a free plan?
What does it cost, and what is a seat?
Is my research data used to train AI models?
Where is the data hosted, and is a DPA available?
Does Cauliflower offer an API?
When do we actually need the Enterprise plan?
Should you pick Cauliflower?
Cauliflower is a narrow tool that knows exactly how narrow it is, and that is its main strength. Rather than selling a generic dashboard builder with an AI layer bolted on, it takes the position that survey methodology is the hard part — weights, waves, base sizes, significance, item batteries — and builds everything else on a research model that encodes it. For an insights team that rebuilds the same tracker every quarter and hand-codes thousands of open ends, that is a credible answer to a genuinely expensive problem, and the native editable PowerPoint export attacks the single step research teams complain about most.
The European positioning is coherent end to end rather than decorative: a Hamburg company, hosting in Frankfurt, Gävle and Dublin, AI inference kept inside an EU Azure region, a data processing agreement signed as standard, and an unconditional statement that customer research data does not train any model. Pricing is published in unusual detail, down to nine credit top-up bands and a plain definition of what a seat is — and the vendor even corrects its own earlier marketing on that page.
The reservations are commercial rather than technical. Nothing can be tried alone: there is no free plan and no self-service trial, so evaluating the product means booking a sales call. The full REST API sits behind the third tier and has no public documentation, and the security artefacts buyers usually diligence — subprocessor list, technical measures, DPA — arrive on request under NDA rather than on the page. The vendor holds no security certification of its own, and the terms of use exist only in German. This is a young, small publisher selling to large procurement departments; the product looks ready, the paperwork asks for a conversation.
- Choosing a selection results in a full page refresh.
- Opens in a new window.