Prior-Art Scout · Rule 56 compliance
Precision prior art, without the over-disclosure liability.
Get AI-grade prior art without inheriting a Rule 56 problem. Scout is a secure, algorithmic air gap between the global patent corpus and your drafting desk.
Click any step to jump to it
Independently benchmarked
The #1 commercially-available prior-art search.
We benchmarked Scout's retriever against every prior-art engine that can be independently tested, and it came out ahead of every engine a firm can license. That leadership is exactly what makes a capped result set trustworthy: the references you receive are genuinely the strongest, not a sample.
Best you can deploy
On our claim-to-prior-art benchmark, Scout's retriever led every commercially-licensable engine, and edged past the leading non-commercial research model.
Strongest where it's hardest
Its widest lead is on cross-domain prior art: relevant references from unrelated fields, the non-obvious "analogous art" that is easiest to miss.
The best, not the most
Because the engine leads the field, the references within your cap are the strongest available, so a capped, Rule-56-friendly output costs you nothing in quality.
Measured honestly
Tested on a large, leakage-controlled holdout: the evaluation set is held separate from training and the searchable corpus predates the claims, so nothing wins by memorizing.
The problem
AI search created a new liability: over-disclosure.
Run every unfiled idea through a broad AI search and you suddenly possess thousands of references, and a duty of candor over all of them. The result is Jumbo IDS filings, willful-blindness worries, and client-confidentiality risk.
Push every unfiled idea through a broad AI search and you now possess thousands of references, with a duty of candor over each one.
helps Scout's algorithmic air gap explores in the background but surfaces only the most-correlated references, up to the cap you set, so the set you review and disclose stays small.
"Jumbo IDS" filings and willful-blindness worries are the new occupational hazard of AI-assisted prior-art search.
helps You receive a focused, capped list. Not thousands of marginal hits you would have to triage and weigh for disclosure.
Putting unfiled invention disclosures into AI tools risks novelty, privilege, and client confidentiality.
helps Zero-data-training governance: your queries, drafts, and uploads are used only to run your search. Never to train our models, and never shared with third parties.
How Scout works
From the global corpus to a safe, filable set.
The algorithmic air gap
See only what matters. A set you control.
Scout explores thousands of global patents and non-patent literature in the background, then surfaces only the most-correlated references, up to a cap you set before the search. The broad exploration stays in the engine; only your capped set reaches your desk for review.
- Capped output: you set how many references surface
- Keeps your Rule 56 duty of candor to a manageable, relevant set
- You review and disclose only the references you receive
- Background exploration informs ranking; it never lands on your desk
Examiner intelligence
See what your examiner cites.
Look up the examiner on your case and see their allowance rate, interview receptiveness, difficulty, and the prior art they tend to rely on, so your search and your response anticipate their playbook instead of reacting to it.
- Allowance rate, interview, and difficulty scores
- Frequently-cited references and rejection mix
- Art Unit and Technology Center context
- Working-style and RCE-tolerance signals
Your IP stays yours
Never training data, never shared.
Your queries, draft claims, and uploaded invention disclosures are used only to run your search. Never to train our models, and never shared with third parties. Access is scoped to your account.
- Zero-data-training: client IP never trains our models
- Your work is processed only to serve your request
- No third-party data sharing
- Account-scoped and access-controlled
Non-patent-literature search can use an optional third-party source. It is off by default; every patent-database search stays internal.
Connects with
Client art + your results → one SB/08.
Your capped Scout results flow straight into Toolbox, where they merge with the references your client already knows into one ready-to-file PTO/SB-08. No manual re-keying, no export and reimport.
- Client portal: inventors upload known references securely
- Smart categorization into US patents, foreign patents, and NPL
- Auto-formatted into the correct SB/08 columns
- Ready for instant review and signature
The outputs
The deliverables you actually want to see.
A short, defensible set. Not a data dump you have to wade through and disclose.
Your capped result set
Only the most-correlated references, up to the cap you set, so the set you review and disclose stays manageable.
Your IP stays yours
Queries and drafts are used only to run your search. Never to train our models or shared with third parties.
A ready-to-file SB/08
Client-known art plus your curated results, auto-categorized into a signature-ready IDS.
Who it's for
For everyone who answers to Rule 56.
Prosecution teams
Get the strongest references without inheriting a duty of candor over thousands of hits.
IP risk & compliance
Keep AI-assisted search inside Rule 56 and client-confidentiality lines.
In-house & licensing
Vet inventions while keeping your IP scoped to your own account.
FAQ
Common questions.
You set a result cap before searching, and Scout surfaces only that many of the most-correlated references, keeping your Rule 56 duty of candor manageable and avoiding "Jumbo IDS" over-disclosure. The engine still explores thousands in the background; you receive only the ones that matter.
The broad exploration is used to rank your results and is not presented to you for review or disclosure. You receive only the references within your cap.
No. Under our zero-data-training governance, your queries, draft claims, and uploaded disclosures are used only to run your search. Never to train our models, and never shared with third parties.
Client-uploaded known references plus your curated results are auto-categorized into US patents, foreign patents, and NPL, then formatted into a ready-to-file PTO/SB-08.
No. Scout surfaces and ranks the most-correlated references; materiality and the final IDS decision stay with you.
Precision prior art, without the liability.
See how Scout's algorithmic air gap keeps your duty of candor manageable under Rule 56, and turns your capped result set into a ready-to-file IDS.
