Findest
Findest is a Dutch AI research platform for corporate R&D teams. Its IGOR^AI engine searches 250 million scientific papers and patents, groups results by technology, and links every claim back to its source document.
What is Findest?
Findest is a Dutch technology-scouting company that sells two interlocking things: The Universe, a research workspace for R&D teams, and a scouting service run by PhD scouts. The engine behind both is IGOR^AI, a proprietary system the vendor positions against general-purpose LLM tools, working over a curated base of more than 250 million scientific papers and patents that Findest says it owns, updates and indexes itself. The method is deliberately old-fashioned before it is computational. A question is first decomposed the TRIZ way into function, objectives, constraints and means, with the technologies a user happens to name held aside as means so they are not mistaken for requirements. The question is then checked against seven framing traps the vendor has catalogued: keyword dump, goal-framed, solution-locked, overconstrained, conditions an abstract will not show, wrong source, and two questions in one. Only then does the search run. Five versions of the question go out in parallel: the original, a minimal one, a detailed one, a functional rephrasing, and one aimed at an adjacent field. Each makes five to seven passes, re-reading and re-scoring up to 3,000 documents per pass; across all five that is on the order of 150,000 abstracts read and judged, with the best 100 to 200 returned, in about two minutes. An LLM judge scores every document as highly relevant, relevant but thin, or wrong despite appearances, and its verdict steers the keywords of the next round; it was calibrated on how the in-house scouts used to pick papers. You get a short answer drawn from the top 30, then a full extraction grouped by technology, with the constraints pulled out group by group. The distinguishing claim is traceability: every statement links to the paper, patent or internal document behind it, verified before it is shown, which the vendor sums up as zero hallucinated references. Searching by function rather than by name is what surfaces answers from other industries. A Chrome extension captures findings in one click and warns when a colleague has already covered the ground, and everything stays in the project's Universe, so an engineer who joins six months later still finds the reasoning. Findest claims 3,000 R&D projects for more than 400 companies since 2017. Literature monitoring is announced as still being built.
What it does
- Ask an R&D question in plain language instead of assembling a list of keywords
- Have that question broken down into function, objectives, constraints and means before anything is searched
- Run five parallel reformulations against a curated base of 250 million papers and patents
- Get 100 to 200 retained documents back, grouped by technology, in about two minutes
- Generate a synthesis report in which every claim links to the source behind it
- Capture a web page, a paper or a patent in one click through the Chrome extension
- Share the project and its sources with the team, who can pick the thread up months later
When to use Findest / When not to
A quick filter to help you decide if Findest is the right fit.
When to use Findest
- R&D managers scoping a new project who need a defensible map of what already exists before committing budget and engineering time
- Technology scouts and innovation managers working on materials, ingredients or process innovation, where broad exploration matters more than deep specialisation
- Corporate R&D teams in the sectors the vendor claims as its own: electronics, semiconductors, agrotech and food, chemistry and materials, energy, consumer goods, pharma and health
- Engineers who must defend a technical direction in front of senior peers and need every statement traced back to a paper, a patent or an internal document
- Large organisations losing time to duplicated research across sites and projects, which the Chrome capture extension flags when a colleague has already studied the subject
When not to use Findest
- Individuals, freelancers and small teams: there is no self-service signup, no published price and no free plan, and the only way in is a sales demo
- Anyone with an occasional, one-off research need, since the commercial route is a negotiated subscription rather than pay-per-search
- Research outside the natural sciences: the vendor limits its claimed scope to topics that follow the rules of physics, chemistry or biology
- Users whose real question is a supplier search, a market question or competitive intelligence, which the vendor itself lists among the seven framing traps its literature engine is not built to answer
- Developers and mobile-first users: no API and no documentation are published, there is no iOS or Android app, and the capture extension exists for Chrome only
How to use Findest
A typical end-to-end flow, from setup to results.
- Start on the vendor's site: there is no self-service signup, so the entry point is Book a demo or Talk to us
- Fill in the contact form with your name, business email, phone number, the size of your R&D team and the technology challenge you are facing
- Pick a slot in the vendor's HubSpot scheduler and take the demo
- Once a licence is signed, sign in through the browser on the Universe platform; every authorised user must be named to Findest and credentials cannot be shared
- Install the Universe by Findest extension from the Chrome Web Store to capture pages, papers and patents in one click
- Write your R&D question in plain language and let the platform break it down into function, objectives, constraints and means before you launch it
- Run the search: roughly two minutes of computation for about twenty minutes of your own time
- Read the quick answer first, then work through the technology groups, keeping the documents that count and discarding the rest
- Dig deeper, regenerate the report, and share the project so the questions, answers and sources stay available to whoever joins later
- If the scouting service is part of the deal, follow the six-week sprint (about two weeks of scoping, two of landscape and deep dive, two of validation and delivery), with platform access continuing for up to six months after the project ends
Pros & Cons
Pros
- Full traceability: every claim links back to the paper, patent or document behind it, verified before display
- Depth no team could reach by hand: on the order of 150,000 abstracts read and scored per search, in about two minutes
- The method is published step by step on a dedicated page of the vendor's own site, which is rare in this market
- Functional search by what a technology does rather than what it is called, which brings back answers from outside your own industry
- The scientific base is owned, updated and indexed by the vendor rather than being a thin wrapper over a third-party engine
- Support by PhD scouts is presented as included in the software licence, and projects persist so work can be picked up months later
- European publisher established in Amsterdam since 2017, Dutch law and Amsterdam courts, with 3,000 projects and more than 400 client companies claimed
Cons
- No public pricing at all: no pricing page, no amount, no tier, purchase by quotation only, and neither a free plan nor a free trial is announced
- The legal pages are unfinished Webflow templates: the terms-and-conditions page and the legal-pages page are still Lorem ipsum, and the real terms exist only as two linked PDFs
- The published privacy document is an internal compliance audit rather than a structured privacy policy, and the site contradicts itself on its own registered address and on whether third parties ever get access to data
- No API and no mobile app; the capture extension exists for Chrome only
- No subprocessor list is published (available on request only), no hosting country or provider is named, and no retention period or minimum age is documented
- No email address on the site's own domain: the published addresses sit on findest.eu, and there is no dedicated legal or GDPR contact, nor any link to a social network
- Visible editorial neglect: a misspelled URL for the legal pages, a menu item reading Electonics, buttons pointing at # or at an unrelated industry page, and a drafting note left in place on the homepage
Pricing & Plans
Findest publishes no price. There is no pricing page, an absence confirmed against the site's 59-URL sitemap, no published tier, and neither a free plan nor a free trial is announced. Access is by quotation: a demo, then an Order Confirmation Form and a Subscription Fee agreed in the contract. The SaaS terms state that all amounts are exclusive of tax, that payment falls due within 30 days, and that Findest may raise its rates on 30 days' notice, the client then being free to terminate; the scouting terms cap the validity of agreed prices at 24 months from signature. Additional user licences can be bought mid-term, and support by the scouts is presented as included in the software licence. One caution: the figure of about EUR 7 quoted on the site is the vendor's claimed execution cost of a single guided search, set against roughly EUR 1,000 for the same job done by hand. It is not a subscription price, not a plan and not a starting price.
Data, GDPR & hosting
A consolidated view of how Findest handles your data.
GDPR overview
Findest is established in Amsterdam, so no Article 27 EU representative is required or named. What the site publishes as its privacy page is not a conventional privacy policy but a four-chapter internal GDPR/AVG audit, last updated 06/06/2025, whose conclusion is the vendor's own: we conclude that Findest is GDPR compliant. That is a self-assessment, not third-party certification. The document limits personal data to name, job title and business email held in a CRM, names LinkedIn and Rocketreach as sources, and lists the usual rights: access, rectification, erasure, objection, restriction, portability and withdrawal of consent. The only channel given for exercising them, complaints included, is the phone number +31 20 893 2700; there is no GDPR email address and no DPO. No supervisory authority is named. The SaaS terms, last updated 10 August 2025, cover transfers outside the EEA with Standard Contractual Clauses.
Who owns the data?
Under the IGOR^AI SaaS terms, Data means anything the client owns, uses or creates in the service, and it stays the client's; search queries and search terms are expressly treated as the client's confidential information, and Findest may access them only to perform its obligations. Two carve-outs matter. Findest owns the anonymised, aggregated Analytical Data derived from client use, and that ownership survives termination. Any feedback a client sends, including suggestions and feature requests, is licensed to Findest perpetually, worldwide, free of charge and with the right to sub-license. Up to one month after termination, the client may request a copy of its data, at its own reasonable cost, or its deletion.
Reuse rights
The client keeps its own content and may reuse the reports, extracts and summaries it produces without asking Findest for permission; the SaaS terms impose no downstream restriction on the client's own Data. On the vendor's side, Findest may access client data to perform the contract and may reuse anonymised, aggregated data for internal research and product development, and article 14.3 lets it use the client's name and logo in its own communications unless the client says otherwise. Nothing anywhere states whether customer data trains models, in either direction. Two first-party documents contradict each other: the privacy page states that Findest gives no access to third parties, while article 9 of the SaaS terms explicitly provides for third-party subprocessors and for transfers outside the EEA under Standard Contractual Clauses.
Data retention & training
Hosting summary
Findest names no hosting country, no region and no provider. The privacy page says only that data is stored on external, highly secured servers meeting the right ISO security standards: no standard number, no certificate, no data centre. A targeted search of the whole retrieved site for AWS, Azure, Google Cloud, data centre, ISO 27001 or SOC 2 returned nothing, so the empty hosting fields record a documented absence rather than a gap in collection. What can be established is jurisdictional rather than physical. The publisher is Findest B.V., established in Amsterdam, and both sets of terms are governed by Dutch law, with Amsterdam as the competent court. Article 9.2 of the SaaS terms states that personal data transferred outside the European Economic Area will be covered by Standard Contractual Clauses where required, which confirms that non-EEA transfers are possible without saying where to; article 5.3a promises administrative, physical and technical safeguards in general terms. A buyer who needs the hosting location in writing will have to obtain it during contract negotiation.
Things to keep in mind
Risks and trade-offs to weigh before adopting Findest.
- Over-reliance is the quiet risk: 150,000 abstracts read in two minutes cannot be checked by hand, so a team that stops reading the underlying papers ends up trusting the engine's grouping and its zero-hallucination claim on faith. Treat the output as a well-sourced starting map, not as proof that nothing else exists.
- The site contradicts itself on its own registered address: the privacy page gives De Boelelaan 1085, 1081 HV Amsterdam, while the footer of every page and the NDA page give Hekelveld 8-10, 1012 SN Amsterdam. The Dutch commercial register confirms Hekelveld; De Boelelaan appears to be inherited from a 2023 PDF and out of date.
- It also contradicts itself on third parties: the privacy page states that no access is given to third parties, while article 9 of the SaaS terms explicitly provides for third-party subprocessors and for transfers outside the EEA under Standard Contractual Clauses. The subprocessor list is not published, only available on request.
- The terms that actually bind you are not on the terms-and-conditions page, which is an unfilled Webflow template: they live in two separate PDFs, one for the IGOR^AI SaaS (last updated 10 August 2025) and one for the scouting service. Read the PDFs, not the page.
- The GDPR compliance statement is a self-assessment. The privacy page is an internal audit in which Findest is both author and subject, and its conclusion that Findest is GDPR compliant carries no third-party certification.
- Derived and incidental rights run further than buyers expect: Findest keeps ownership of the anonymised Analytical Data drawn from your use, and that survives the contract; any feedback you send is licensed to Findest perpetually and sub-licensable; and your name and logo may be used in its communications unless you object.
- Commercial fine print worth reading before signing: no refund of sums already paid on termination whatever the cause, liability capped at the amount paid over the year, and a one-month window after termination to request a copy or the deletion of your data. No hosting country, retention period or minimum age is documented anywhere.
Setup & Integrations
Technical difficulty
Technically, close to none. The platform runs in a browser, and the only thing to install is the Universe by Findest extension from the Chrome Web Store. There is no API, no connector and no named third-party integration to wire up. The friction is commercial and administrative rather than technical: no self-service signup, entry through a sales demo, every authorised user named individually to Findest and no credential sharing allowed. Onboarding help comes with the deal, since PhD scouts are included in the software licence and a customer success manager handles training.
Deployment
Behind Findest
Fundraising
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Findest do?
How does a search actually work?
How does Findest avoid invented references?
How much does Findest cost?
Is there a free trial or a free plan?
What is the figure of about EUR 7 quoted on the site?
Is there an API or a mobile app?
How long does a scouting project take, and what do you get?
Is customer data used to train models?
Who is behind Findest?
Should you pick Findest?
Findest is a niche tool and does not pretend otherwise. It is built for upstream R&D in the hard sciences, inside companies, and it does one thing with unusual rigour: it turns a research question into a traceable map of what already exists, every claim tied back to a paper, a patent or one of your own documents. The method is published step by step on the vendor's own site, which is rare in this market and lets a buyer judge the machinery rather than the marketing. Behind it stands a European company established in Amsterdam since 2017, with 3,000 projects and more than 400 client companies, most of them European, and PhD scouts included in the software licence. The reservations are as concrete as the strengths. Nothing about the price is public: no pricing page, no tiers, no free plan or trial announced, and the only way in is a sales demo. The legal surface is careless: the terms-and-conditions page and the legal-pages template are still filled with Lorem ipsum, the real terms live in two linked PDFs, and the published privacy document is an internal audit in which Findest is both author and subject. The site also contradicts itself on its own registered address and on whether third parties ever get access to data. Ask for a demo if your team runs materials, ingredient or process research, needs answers it can defend in front of senior engineers, and can negotiate an annual subscription without a public price to anchor on. Look elsewhere if you are an individual, need an occasional one-off search, want an API, or expect to compare prices before talking to sales. In that first conversation, ask for the subprocessor list, the hosting jurisdiction and the retention period: none of the three is published.
- Choosing a selection results in a full page refresh.
- Opens in a new window.