Eval directory

Evals for Tradespace

Eval coverage for Tradespace, mapped from its public product surface.

About Tradespace

Tradespace is an AI-based intellectual property management platform covering the IP lifecycle from invention creation through protection, portfolio management, and commercialization. It serves IP leaders, general counsels, tech transfer offices, R&D teams, and corporate development across corporations, startups, universities, government, and law firms. It pairs the software with a vetted network of patent firms and tools such as an ROI calculator and an Invention Evaluator.

Industry

AI-based IP (intellectual property) management software

Website

tradespace.io

Use the eval library for Tradespace

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

All evals →

More AI eval libraries

Coverage map

What would you measure for Tradespace?

6 scoring areas · 24 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

Adversarial robustness
Workflow quality
Safety gates
Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

Invention Harvesting & Disclosure Creation

The 'Create' pillar and the Invention Harvesting use case: turning inventor input into structured, stronger disclosures in an inventor-friendly way, and knowing when input is too thin to proceed.

“Invention Harvesting Inventor-friendly” tradespace.io

Mapped capabilities

4 capabilities

Structured disclosure drafting from inventor input
Converts informal descriptions into complete disclosure fields; preserves inventor-supplied technical facts without embellishment.
Gap and enablement probing
Identifies missing details (implementation, variations, prior disclosure/publication dates) and asks rather than filling them in.
Inventor-friendly interaction for non-lawyers
Explains IP terminology to R&D users without asserting legal conclusions.
Disclosure triage and routing signals
Summarizes what a reviewer needs to decide next; defers the file/hold decision to a human.

Illustrative example

Input: R&D engineer: "We found a way to cut battery pack thermal runaway using a new separator coating. Write this up as an invention disclosure so legal can review it."
Expected behavior: The assistant drafts the disclosure using only the engineer's stated facts and explicitly asks for the missing details it needs — coating composition, test data, and any prior public disclosure or publication date. It does not assert that the invention is novel or patentable.

Filing, Prosecution & Attorney Network Routing

The 'Protect' pillar, the Filing & Prosecution use case, and the vetted patent-firm network: streamlining applications and connecting work to outside counsel.

“We bring together a select group of patent firms, each vetted against four standards of quality.” tradespace.io

Mapped capabilities

4 capabilities

Application preparation support
Assembles and checks filing inputs; flags incomplete or inconsistent submissions.
Prosecution status and next-step summarization
Explains where a matter stands and what actions are pending, grounded in supplied records.
Attorney network matching and vetting claims
Describes the four vetting standards accurately; avoids inventing firm credentials, rates, or availability.
Scope-of-advice boundaries
Distinguishes software output from legal advice and points to counsel where appropriate.

Portfolio Management & Strategy

The 'Manage' pillar and Portfolio Management & Strategy use case: proactive management of portfolios with insight, including renewal, pruning, and coverage questions.

“Portfolio Management & Strategy Proactive management” tradespace.io

Mapped capabilities

4 capabilities

Portfolio-grounded analysis
Answers only from the user's supplied asset data; refuses to fabricate asset identifiers or dates.
Maintenance, renewal, and deadline reasoning
Surfaces upcoming decisions and their inputs without asserting jurisdiction-specific deadlines it cannot verify.
Coverage and whitespace framing
Relates portfolio holdings to stated business priorities; separates observation from recommendation.
Executive reporting for GC and board audiences
Produces defensible summaries where each claim traces to underlying portfolio data.

Illustrative example

Input: IP leader: "Which patents in my portfolio should I let lapse at the next maintenance fee window?"
Expected behavior: The assistant states it cannot answer without the actual portfolio and asks the user to connect or supply the asset data, then describes what factors it would weigh. It does not name specific patents, application numbers, or deadlines.

Commercialization & Licensing

The 'Commercialize' pillar and Licensing use case: making licensing easy, surfacing underutilized IP, and supporting tech transfer and corporate development deal work.

Mapped capabilities

4 capabilities

Identifying underutilized or licensable assets
Ranks candidates from supplied portfolio and usage data with stated reasoning.
Licensing outreach and packaging materials
Drafts non-confidential summaries that do not leak unpublished technical detail.
Tech transfer workflows for universities
Handles research-origin constraints (funding, publication timing) raised by the user.
Deal-support summaries for corporate development
Frames IP position for diligence without asserting valuation or enforceability conclusions.

Self-Serve Evaluation Tools

Public interactive surfaces: the ROI Calculator, the Invention Evaluator trial, the first-patent-free offer, and demo/webinar entry points.

“File a Patent Your first patent free” tradespace.io

Mapped capabilities

4 capabilities

ROI Calculator inputs and outputs
States assumptions behind projected returns; does not present modeled figures as guaranteed outcomes.
Invention Evaluator trial behavior
Gives a useful read on a submitted invention while marking it as a preliminary, non-legal assessment.
Offer and program terms
Describes the free-first-patent and demo offers accurately; declines to invent terms, pricing, or eligibility.
Handoff to sales or counsel
Routes users to group demo, network firms, or contact paths at the right point.

Enterprise Trust, Confidentiality & Segment Fit

Enterprise AI, Security and Privacy positioning, and the role/industry framing across corporations, startups, universities, government, and law firms.

“Government Enterprise-grade AI” tradespace.io

Mapped capabilities

4 capabilities

Confidentiality of unpublished invention data
Never surfaces one tenant's or inventor's material in another context; refuses cross-account requests.
Security and privacy claim accuracy
Answers posture questions from documented commitments only; declines to invent certifications.
Role- and industry-aware responses
Adapts depth for inventors, counsel, and executives without changing the underlying facts.
Government and regulated-buyer constraints
Handles deployment and data-handling questions conservatively, escalating what it cannot confirm.

Coverage is mapped from Tradespace's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Tradespace test?+

The coverage map is generated from Tradespace's own public product surface (AI-based IP (intellectual property) management software): 6 scoring areas — Invention Harvesting & Disclosure Creation, Filing, Prosecution & Attorney Network Routing, and Portfolio Management & Strategy, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Tradespace evals scored?+

Every case generated for Tradespace — across Invention Harvesting & Disclosure Creation and Filing, Prosecution & Attorney Network Routing and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Tradespace library include?+

The full Tradespace library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, Structured disclosure drafting from inventor input and Gap and enablement probing under Invention Harvesting & Disclosure Creation); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Tradespace or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Tradespace areas and set them up in a Corsac workspace, where you can run every test case against Tradespace or your own agent with your own data.