Try Claude
Academic ResearchSearch papers, authors, citations
AI Tools for Patent Attorneys
AI tools that help patent attorneys conduct prior art searches, track competitor filings, analyze claim language, and monitor patent litigation trends.
Get started for free
Works in Chat, Cowork and Code
Papers found
43 potentially relevant papers and preprints
Key prior art
Christiano et al. "Deep Reinforcement Learning from Human Preferences" (NeurIPS 2017)
Key prior art (2)
Stiennon et al. "Learning to summarize with human feedback" (NeurIPS 2020)
InstructGPT paper
Ouyang et al. 2022 — RLHF applied to GPT family; pre-dates 2024 application
Claim differentiation
Recommend narrowing to specific reward model architecture — appears novel
Prior art search in academic literature
Search millions of peer-reviewed papers, conference proceedings, and technical reports for prior art before filing or responding to office actions. Academic publications are often the most damaging prior art — and the hardest to find manually.
Search for academic papers and conference proceedings from 2010–2022 disclosing methods for detecting objects in video streams using convolutional neural networks. I need prior art for claim 1 of application 17/845,231.
Found 47 potentially relevant papers. 3 most relevant: Redmon et al. "You Only Look Once" (CVPR 2016) — real-time object detection; Ren et al. "Faster R-CNN" (NIPS 2015) — region proposal networks; Girshick et al. "Rich feature hierarchies" (CVPR 2014). Recommend reviewing claims against YOLO family — extensive disclosure overlaps.
ToolRouter search_papers
Papers found
47 potentially relevant papers identified
Key prior art 1
Redmon et al. "You Only Look Once" (CVPR 2016) — real-time object detection
Key prior art 2
Ren et al. "Faster R-CNN" (NIPS 2015) — region proposal networks
Key prior art 3
Girshick et al. "Rich feature hierarchies" (CVPR 2014)
Recommendation
Review claims against YOLO family — extensive disclosure overlaps Claim 1
Patent litigation case research
Search Federal Circuit and district court decisions on claim construction, obviousness, and patent eligibility. Build arguments on how courts have treated similar claim language and technology in prior cases.
Find Federal Circuit cases from 2020–2024 where courts reversed obviousness findings by the PTAB in IPR proceedings. I need to understand the current standard for "reasonable expectation of success."
Found 11 Federal Circuit reversals of PTAB obviousness findings. Key case: Becton Dickinson v. B. Braun (Fed. Cir. 2023) — PTAB's motivation to combine was conclusory; reversed. Current standard: obviousness requires articulated reasoning with rational underpinning. PTAB "hindsight bias" reversals up 40% since 2021.
ToolRouter search_cases
Cases found
11 Federal Circuit reversals of PTAB obviousness findings
Key case
Becton Dickinson v. B. Braun (Fed. Cir. 2023) — PTAB motivation was conclusory; reversed
Current standard
Obviousness requires articulated reasoning with rational underpinning
Trend
PTAB "hindsight bias" reversals up 40% since 2021
"Reasonable expectation of success"
Must be specific to claimed result — general optimism insufficient
USPTO examination guideline research
Look up current USPTO examination guidelines, MPEP sections, and Alice/Mayo framework analysis for subject matter eligibility. Know the examiner's playbook before prosecuting AI, software, and biotech applications.
What is the current USPTO guidance on patent eligibility for AI-generated inventions where an AI system substantially contributed to the invention? What claim structures are being allowed?
USPTO February 2024 guidance: AI-assisted inventions are patentable if a human made a "significant contribution" to each claim. "Significant contribution" = more than directing AI to solve a problem. Allowed claim structures include: method claims specifying human-defined technical parameters; system claims where human designed the AI architecture. Pure AI-as-sole-inventor: not patentable under Thaler v. Vidal (Fed. Cir. 2022).
ToolRouter research
Patentable threshold
Human must make "significant contribution" to each claim — more than directing AI to solve a problem
Allowed: method claims
Specifying human-defined technical parameters — allowed at examination
Allowed: system claims
Human designed the AI architecture — allowed at examination
Not patentable
Pure AI-as-sole-inventor — barred per Thaler v. Vidal (Fed. Cir. 2022)
Guidance source
USPTO February 2024 guidance — effective immediately for all pending applications
Competitor patent portfolio monitoring
Track when competitors file new patents in your client's technology space. Monitor SEC disclosures for patent licensing deals and infringement notices. Stay ahead of freedom-to-operate risks.
Has Qualcomm disclosed any new patent licensing agreements or infringement actions in the 5G standard essential patent space in their last 3 10-Ks?
Qualcomm 2022–2024 disclosures: 2022 10-K: Renewed 5G SEP license with Apple ($1.8B upfront). 2023 10-K: Active license negotiations with 3 Chinese OEMs disclosed. 2024 10-K: New ITC investigation filed against Arm Holdings for ADAS chip claims. Total royalty revenue from SEPs: $5.2B/year.
ToolRouter search_filings
2022 10-K
Renewed 5G SEP license with Apple — $1.8B upfront payment disclosed
2023 10-K
Active license negotiations with 3 Chinese OEMs disclosed
2024 10-K
New ITC investigation filed against Arm Holdings for ADAS chip claims
Annual SEP royalty revenue
$5.2B/year from standard essential patents
Technical literature analysis for claim drafting
Research the state of the art in a technology field to draft claims that distinguish over the prior art while covering the broadest scope. Find the boundaries of what was known vs. what your client actually invented.
Research the state of the art in quantum error correction as of December 2021. Specifically: surface code implementations, logical qubit thresholds, and any publications on real-time decoding algorithms.
State of the art Dec 2021: Surface codes dominant (Fowler et al. 2012 foundational). Error threshold: ~1% physical error rate. Real-time decoding: Union-Find decoder (Delfosse 2020) — fastest known. Gap in prior art: parallel hardware decoders for >1000 qubits — this appears novel as of your priority date.
ToolRouter search_papers
Dominant approach
Surface codes (Fowler et al. 2012) — most cited implementation
Error threshold
~1% physical error rate — well established as of Dec 2021
Real-time decoding
Union-Find decoder (Delfosse 2020) — fastest known decoder at priority date
Gap in prior art
Parallel hardware decoders for >1,000 qubits — appears novel as of your priority date
ToolRouter research
Google Quantum AI
Surface code experiments on Sycamore — "Beyond Classical" (Nature 2019)
IBM Quantum
Falcon/Eagle processors · topological code research ongoing
Academic leaders
Preskill (Caltech) · Aaronson (UT Austin) · Fowler (Google)
Notable gap
Real-time decoding at scale (>1,000 qubits) — not commercially demonstrated
Ready-to-use prompts
Prior art literature search
Search academic papers and conference proceedings from 2015–2023 disclosing methods for natural language processing using transformer architectures, specifically attention mechanisms for machine translation. Return titles, authors, publication year, and key technical contributions.
Alice eligibility research
Research how courts and the USPTO have applied the Alice two-step framework to claims involving computer-implemented financial instruments. What claim structures survived eligibility challenges in 2022–2024?
IPR invalidity case search
Find PTAB IPR decisions from 2021–2024 where claims in the semiconductor manufacturing technology space were found obvious over a single prior art reference. Include case numbers and claim language patterns that were invalidated.
Competitor IP disclosure review
Search Google's last 3 annual reports (10-K) for disclosures about patent licensing revenue, ongoing IP litigation, and any risks related to open-source software patent exposure.
Patent damages case research
Find Federal Circuit cases from 2019–2024 applying the Georgia-Pacific factors for reasonable royalty damages in smartphone patent cases. What royalty base calculations have courts approved?
State of the art search
Research the state of the art in CRISPR gene editing as of January 2020, specifically base editing techniques that achieve single nucleotide precision without double-strand breaks. Include key papers, inventors, and institutions.
Trademark coexistence research
Search TTAB decisions from 2020–2024 involving likelihood of confusion between technology company trademarks in the SaaS software category. What factors did the Board weigh most heavily?
Technology standard patent research
Research which companies hold the most 5G NR standard essential patents declared to ETSI as of 2024, and whether any have been subject to FRAND licensing disputes in US or EU courts.
Tools to power your best work
SEC FilingsSEC EDGAR filings & financial data
Deep ResearchAI research reports with citations
PDFRead, analyze & merge PDFs
Company LookupCompany data for UK, US & EU
Compliance ScreeningAML, sanctions, PEP & due diligence investigations
Legal ResearchCase law, statutes, regulations & legislation worldwide
165+ tools.
One conversation.
Everything patent attorneys need from AI, connected to the assistant you already use. No extra apps, no switching tabs.
Pre-filing prior art clearance
Before filing a patent application, conduct a thorough prior art search across academic literature, patents, and case law to identify the closest prior art and refine claim scope.
Freedom-to-operate analysis
Before a product launch, assess whether the product design clears active competitor patents and any relevant litigation risks.
Frequently Asked Questions
Does the academic research tool cover patent databases or just journals?
Academic Research covers peer-reviewed journals, conference proceedings, preprint servers (arXiv, bioRxiv), and technical reports. It does not directly search USPTO or EPO patent databases — for patent-specific searches, use it for non-patent literature and combine with dedicated patent search tools.
Can the case law tool search PTAB decisions and ITC rulings?
Legal Research includes PTAB IPR and PGR decisions, Federal Circuit opinions, and district court patent decisions. ITC Section 337 proceedings are included where public records are available. For the full PTAB database, cross-reference with the USPTO's PTAB e-filing portal.
How useful is the deep research tool for claim drafting support?
Deep Research is most useful for understanding the state of the art and identifying claim differentiation opportunities. It synthesizes technical and legal information into structured summaries. Draft claims should always be reviewed by a registered patent practitioner before filing.
Can I use these tools to monitor competitor patent filing activity?
SEC Filings and Company Lookup can surface disclosed patent deals and litigation. For real-time competitor patent filing monitoring (new applications at USPTO/EPO), combine Deep Research with the web search tool and set up recurring searches on specific assignees.
Are these tools useful for international patent prosecution?
Deep Research covers EPO, PCT, and major national patent office guidelines. Case Law includes UK, German, and EU court decisions. For prosecution-specific guidance in non-US jurisdictions, verify findings against official patent office sources.
Give your AI superpowers.
Get started for free
Works in Chat, Cowork and Code