Patents are not an archive.
They are a body of knowledge, a map of your technology field, a source of ideas — and above all a springboard: you don’t need to reinvent what has long been solved elsewhere. Most tools stop at the results list. inventMAP starts there — and turns patents into something your engineers reach for every day.
Find → Understand → Use
Running a patent search is the easy part. The value comes after: when documents become conclusions, and conclusions become a decision. inventMAP walks that path to the end — close enough to daily engineering work that you don't save it up for a project, you do it along the way.
Five building blocks, one way of working
This is how you work inside the module — from the first search to the report you hand on. Everything listed here is part of the patent module; anything beyond it is set out separately further down.
From a sentence to a search strategy — without a query language
Every search is an object of its own: named, reopenable at any time, with notes, flags and history — private to you or visible to the project team. How you start is up to you: describe what you are looking for in your own words and the assistant builds the search boxes from it. Or simply type a term, a patent number, a classification. Or build the boxes by hand — AND, OR, NOT, word proximity, title or full text, spelling variants grouped onto the term. The two ways interlock: the assistant makes the first throw, you fine-tune. A count preview tells you beforehand whether you are heading for 12 or 12,000 patent families, and the “what's in there” column shows the applicants, classifications, years and countries of your current result set — click to narrow down. And when a search comes up empty, a diagnosis tells you which box caused the zero instead of leaving you guessing.
A quick overview — and keep the good ones
What matters is getting a picture quickly. There are several views and filters for that: by year, country, applicant and legal status, sorted by relevance or date, optionally one row per patent family rather than per publication. Flag interesting hits with a star and attach notes directly to the patent — private or shared with the team. You can also rate hits as good or as irrelevant: the search sharpens itself on the next pass without you having to rebuild the boxes. Every patent opens in a detail card with abstract, claims, description, drawings, patent family, legal-status indicator and citations in both directions. Full texts are fetched wherever they exist in structured form — once obtained they stay stored and are available to every further analysis. That is the difference between a reading based on half an abstract and one based on the claims.
The heart of it: not counting, but understanding
Most tools count patents — by year, by applicant, by country. That tells you who was busy, but nothing about what is inside the documents. inventMAP reads every hit once, methodically, and draws out what matters for engineering: which problem was solved? By which working principle? Which technical contradiction sits behind it — which property was improved, and at whose expense? Where does the solution sit on the evolution lines, and what would come next? Which material replaces which, and on what grounds? Out of this come analyses that give your own development real pointers: which functions should your product have? How has the field solved which problems? Which kinds of solution are well covered — and which would be conceivable but nobody uses? Those gaps are precisely the interesting places. And the reading is stored across your whole account: paid for once, reused by every further analysis.
Talk to your result set
Instead of reading four hundred abstracts, just ask. “How is temperature sensing solved in these patents?”, “Are there solutions without a spring?”, “What is the recurring core problem in this field?” — the chat answers from the patents it has read, names them, and every mention is clickable: one click and the hit is open in front of you. As a sparring partner it also takes the uncomfortable questions — what have I overlooked, where do the sources contradict each other? It always draws on your result set, not on a language model's free guesswork.
Stay on it — and pass it on, licence or not
Any search can become a watchdog: re-run weekly through quarterly, and only what is genuinely new gets reported. The email is merely the messenger — behind it sits a properly designed report page with drawing, title, applicant, date and summary for each new filing, plus the watchdog's observations. The report is frozen: it shows the state of that run for good, which makes it usable as evidence. And you can pass on either one — the report as well as the live search with its analyses as a read-only view. Recipients need no inventMAP licence: their own link, optionally password-protected, revocable at any time, with a view counter. That way your management gets a readable page every four weeks instead of a results list nobody attends to.
See an example report →See an example alert →
What the other modules do with patents
So the boundary is clear: the two tools below work with patents but are NOT part of the patent module. They come with Product development and Strategy respectively — and both packages include the patent module in full.
A body of knowledge that keeps growing
We continuously turn relevant patents into problem-solution networks: each patent is held as a chain of problem, contradiction and solution rather than filed away as a document. That stock keeps growing — and with it the benefit for everyone working in the network (Product development module).
The everyday tool of the development engineer
Patent work has a reputation as a chore: once at the start of a project, then never again. We are building the opposite — a tool you open because it helps while you design. Anyone who has once seen how much solved problem sits inside a patent specification enjoys working with them.
Three typical ways in
Are your patents currently producing nothing?
Start with a single search — and see what comes after the results list.