I

InnovationQ+

Retrieval in InnovationQ+ is model work: Semantic Gist combines a probabilistic language model with a deep neural network, and it is what turns a query into results. CompassAI sits above it as a separate reasoning layer assessing novelty, similarity and technical differences, so both finding and judging run through named models. The platform is also a content and analytics business: customers pay for coverage of 154 patent authorities, more than 190 million IP data points, full text IEEE access and IP.com's own Prior Art Database, none of which depends on the models.

Keyword, image and document search are offered alongside concept search, and the Portfolio Intelligence Report rests on a published fourteen factor framework measured against a peer group, not on a generative step. Without the AI, a customer still has a large licensed corpus with keyword search and structured scoring.

IP.com's published technical white paper frames the problem in retrieval terms, polysemy and synonymy in patent vocabulary, and reviews the alternatives it rejected, naming latent semantic analysis and its successors and citing evidence that they degrade at scale. It describes the design used: probabilistic language modeling for detailed document representation, combined with a deep neural network that compresses those into compact semantic signatures, the two used together for coarse filtering and fine ranking.

A further concept layer uses graph clustering and automatic recognition of technical phrases. The paper names the patents covering the work and cites the research it draws on. On the reasoning side, outputs are grounded in retrieved evidence, and every insight stays linked to its source documents. IP.com says it benchmarks retrieval and reasoning against real IP use cases on relevance, recall and expert validated outcomes, but publishes no result, including no recall figure.

In InnovationQ+, retrieval is separated from reasoning so that the reasoning layer can only work on evidence the retrieval layer produced, which IP.com describes as reasoning from verified evidence within defined constraints. CompassAI is said to support expert judgment rather than replace it, and outputs are described as guided and grounded in evidence rather than free form text. Insights stay linked to the source documents, users can inspect how conclusions were formed and check them against the original material, and private interaction histories are kept so an organization can revisit earlier analyses.

No threshold at which a step proceeds without review, confidence signal on a retrieved result or reasoning output, distinction between supervised and unsupervised modes, or account of what happens when an output is found wrong is published.

The published InnovationQ+ figures describe the product rather than deployments: 154 patent authorities searchable in one query, more than 190 million IP data points, and a claim of 30 percent faster per search. Buyers are named by function: engineering and R&D, IP strategy and IP professionals. IP.com says patent offices are among its customers and that its Prior Art Database is a required reference for international examiners under Rule 34 of the Patent Cooperation Treaty.

No customer is named on the site. The one published case study, on finding critical prior art at a Fortune 500 company, is anonymized and requires contact details to read. The 30 percent figure has no baseline, method or sample. The stronger outcome figures on the IP.com site, 90 percent lower patentability evaluation costs and 80 percent faster idea to report, belong to IQ Ideas+, a different product.

IP.com says customer data is never used to train IP.com's models or any third party models, which closes the route by which customer material would otherwise reach an outside provider of the reasoning layer. Confidentiality is independently attested: the SOC 2 Type 2 examination expressly covers Confidentiality of the Software Services System, including, in IP.com's words, protection of sensitive information from unauthorized disclosure across collection, creation, use, processing, transmission and storage.

Interaction histories are described as private and secure. Privilege and work product are not addressed by name, and retention is stated only as a fact, that histories are kept, with no period, deletion right or position at the end of a subscription. The query IP.com invites is a draft invention disclosure or full claim set submitted before filing, which destroys novelty if it escapes and is often privileged.

CompassAI is designed to support expert judgment, not replace it, IP.com states repeatedly, with the platform keeping experts in control with transparency at every step and outputs guided and grounded in evidence so a professional can check them. One sentence names a specific professional boundary: AI speeds analysis without obscuring judgment or overstepping the inventorship guidelines set by the USPTO. Inventorship is the question AI assisted invention raises for patent practice.

The audience is named plainly, covering R&D professionals, IP managers and legal counsel, so nonlawyers are intended users. Nothing addresses supervision or competence: who should run which capability, what review a supervising practitioner should apply to a prior art result, or what training is expected. No jurisdiction limit is stated, and the duty to disclose material prior art, which this product touches most directly, is not addressed.

IP.com presents separating retrieval from reasoning as the governance mechanism itself: the reasoning layer is limited to evidence already retrieved, which IP.com says reduces hallucinations and unintended outputs by design, with safeguards built in rather than added later. Around that sit a Responsible AI page, a Responsible AI white paper, and a stated improvement loop in which models are refined against performance data and expert feedback.

Data governance is concrete: training exclusion extends to third party models, and organizations are isolated on the platform. A measurement program is described, with relevance, recall and expert validated outcomes named as metrics. No named accountable person, committee or function, adopted external framework, testing regime before release beyond the general benchmarking statement, or result is published. Bias is not addressed, although the retrieval engine's synonym handling is learned from historical patent drafting.

Customer data is never used to train IP.com's models or any third party models, a commitment stated without qualification. Security is independently attested: a SOC 2 Type 2 examination covering Security, Availability and Confidentiality of the Software Services System, described as testing both the design and the operating effectiveness of controls. Its scope is stated to cover protection of information through collection, creation, use, processing, transmission and storage, and a public trust center carries that summary.

The design intent is documented: safeguards built into the architecture, the reasoning layer limited to retrieved evidence, and private interaction histories kept under access control. No subprocessor list is published and no processor is named, though the training commitment expressly contemplates third party models. No retention period or deletion commitment is published for queries or interaction histories, and no incident or breach notification commitment to customers is published, though the SOC 2 scope implies an incident program exists.

The only published allocation of risk is IP.com's website Terms and Conditions, updated 28 June 2024. They provide the services and content as is and as available, exclude every implied warranty including fitness for purpose and noninfringement, and say IP.com makes no commitment about the content or the reliability of the services. They exclude liability for lost profits, revenue or data and for indirect or consequential damage, and cap total liability for any claim at what the user paid in the twelve months before it arose.

The user indemnifies IP.com for any claim arising from use of the content. The same terms grant IP.com a worldwide license to use, reproduce, publish and distribute content a user uploads, written for defensive publishing. Nothing addresses the AI features: no warranty or remedy for a missed prior art reference or a wrong novelty assessment from CompassAI, no vendor indemnity and no insurance position. Enterprise terms are handled through the account team and are not published.

IP.com presents the Innovation Power Suite as products sharing the Semantic Gist retrieval layer, so InnovationQ+ sits alongside InnovationQ, IQ Ideas+ and the report products. The Patent Vitality Report runs self service inside InnovationQ+ with results in minutes. Content integration is real: full text IEEE is delivered in the workflow alongside the Prior Art Database and patent data from 154 authorities. Output leaves the platform as structured reports for sharing with stakeholders.

No connector catalog is published, no outside system is named as supported, and no API or developer documentation is published for this product. Nothing describes how results, saved work or portfolio data would move into a docketing, IP management or document management system.

Full isolation of customer data from other organizations on the platform is a stated governance commitment, reinforced by the SOC 2 Type 2 examination covering Confidentiality and by the description of interaction histories as private and secure. Delivery is a hosted browser platform, with the Innovation Power Suite at ip.com and enterprise deployment handled through the account team. No data center location, country of processing or region choice is stated, and no cloud provider is named.

The tenancy model itself is not described: IP.com states that organizations are isolated without saying whether through single tenant deployment or logical separation.

IP.com has undergone a SOC 2 Type 2 examination, which it describes as confirming both the design and the operating effectiveness of controls. The scope is stated as controls relevant to Security, Availability and Confidentiality of the Software Services System. It covers protection of information during collection, creation, use, processing, transmission and storage, access to information and services, and protection of sensitive information from unauthorized disclosure.

A public trust center, hosted on a third party trust platform, carries this without a sales conversation, and IP.com says it maintains certifications and has regular third party audits. No report, examination period, date or auditor is published, so the scope is stated but not the evidence or how current it is. How the underlying documents are accessed is not stated.

IP.com names Semantic Gist and describes its design in technical terms: probabilistic language modeling for detailed document representations, and a deep neural network compressing those into compact semantic signatures. The two work together, with signatures driving coarse filtering and the language model driving precise ranking, plus a concept layer built on language modeling, graph clustering and automatic phrase recognition.

Its origin is stated plainly: the engine was built entirely in house and is not based on open source search frameworks or licensed third party components. Ownership is evidenced by United States Patents 8,539,000 and 8,548,951, with applications pending. Nothing says whether CompassAI, the reasoning layer, is built in house or runs on a third party foundation model, no version is identified, and no change notice is offered; IP.com describes continuous improvement instead.

No pricing information is published at any level, including the unit of charge. No pricing page appears among IP.com's Why IP.com pages, IP Suite products, professional services lines, industry and use case pages, resources, company pages, trust center or legal pages. Every commercial route leads to Book a Demo, Request a Demo, Contact Sales or a lead form, and even the white papers and sample reports require contact details.

Nothing states the charging model, so whether InnovationQ+ is licensed per named user, per concurrent seat, by search volume, by module or by corpus access is not stated, nor what IEEE full text access costs relative to the base platform, or whether the Patent Vitality Report is included or metered. The one relative price published, that defensive publishing costs as little as two percent of a patent filing, concerns InnovationQ, the free publishing interface, not this product.

Coverage spans 154 patent authorities searchable in one query and more than 190 million IP data points across patent and nonpatent literature, by IP.com's count. Full text IEEE access sits alongside IP.com's own Prior Art Database of technical disclosures. Search modes are listed as concept, keyword, hybrid, image and whole document, so image based prior art and search by disclosure are both supported. Buyers are named as engineering and R&D, IP strategy and IP professionals, with patent offices stated as institutional customers.

Use cases cover prior art and patentability, freedom to operate, invalidity, landscape mapping, competitor analysis, portfolio analysis, licensing opportunities and merger and acquisition diligence. No technology area is named as better or worse covered, no jurisdiction among the 154 is identified as out of scope, nothing says which languages the retrieval engine handles well, and nothing separates what InnovationQ+ covers from the other suite products.