Juristat Data Layer is now available inside Solve Intelligence, so patent teams can draft and prosecute against more than 20 years of real patent prosecution history from inside the tool they already use.
By Lauren Bonner, CEO, Juristat. Current as of September 2026. Rejection-rate figures below are drawn from Juristat's USPTO prosecution dataset; date window pending a data-team pull.
Juristat Data Layer (MCP) is now available inside Solve Intelligence, the purpose-built AI for patent work, allowing patent professionals to pair Solve's prosecution workflows with more than two decades of Juristat's best-in-market, fiduciary-grade patent data. Customers can activate Juristat through a single MCP connector, available immediately.
Patent prosecution is not one size fits all. Two nearly identical applications can follow very different paths depending on the examiner assigned, that examiner's history, and the art unit. Until now, getting that context into an AI workflow meant switching tools and carrying the data back by hand.
Through this partnership, Solve's AI drafts and analyzes against the actual prosecution record, not a general model's guess, through a seamless integration that eliminates the startup time of learning a new tool. Juristat provides patent practitioners the advantage of accurate, defensible AI built on the real USPTO record versus patent prosecution built on a scrape of the open web.
FROM THE COMPANIES
“Anyone can pull a basic allowance rate. The sophisticated move is to look at what has actually worked for a given examiner, on a given rejection, in a given art unit. That is the prosecution record we are bringing into Solve.”
Lauren Bonner, CEO, Juristat
“Juristat's examiner analytics data provide Solve's AI agents with some of the deepest and most valuable proprietary IP data, enabling practitioners using Solve to make the highest-quality strategic decisions across the patent process. The Solve and Juristat teams work extremely well together, and we know this partnership will bring patent workflows in Solve to a new level.”
Chris Parsonson, CEO, Solve Intelligence
The integration shows up as concrete jobs in day-to-day prosecution. Here is where grounding the AI in the real record changes the work.
With Juristat Data Layer (MCP) connected inside Solve, you can draft an office action response against the examiner's actual prosecution record instead of a general model's guess. Same drafting tool, materially better input.
Without grounding, a general model has no reliable access to structured, complete data from the USPTO prosecution record, so when asked about a specific examiner, general AI generates a plausible answer from scraped, incomplete sources. That produces confident, but wrong statements about allowance rates, past arguments, and examiner history, the kind of hallucinated citation that erodes trust and introduces risk in AI-drafted work.
Grounded in Juristat, Solve drafts from how this examiner actually ruled on this kind of rejection, what argument overcame it, and how many rounds it took. Because that history comes from the real record, you are not switching to another tool to pull it yourself, and you are not double-checking the draft against numbers the model may have invented. The model could always do the reasoning. What it was missing is the record, and that is what grounding supplies.
Juristat data shows examiner interviews can increase allowance rates by as much as 9 percentage points, and Juristat flags which examiners are receptive based on similar application trends, so you know before you schedule whether an interview is worth the time.
The published USPTO benchmark set gives the reference points: the average application draws 1.6 office actions, runs 0.32 RCEs, takes about 28 months to disposition, includes 0.42 interviews, and reaches an 83 percent allowance rate. When an examiner sits well outside those numbers, that's a signal worth acting on, and grounded AI surfaces it inside the workflow rather than in a separate tab.
STAT
Up to 9 percentage points of allowance lift from examiner interviews, when you know which examiners are receptive.
See what grounded patent AI looks like in your workflow.
A basic allowance rate tells you if an examiner is more or less permissive, not how to win the case in front of that examiner. Use Juristat to find the winning strategy: whether to amend, request an interview, file an RCE, or hold firm.
The context that matters is what has actually worked: the claim language that overcame a Section 101 rejection with this examiner, the argument style that got the broadest claims allowed in this art unit. With Juristat inside Solve, that context sits next to the draft instead of in a separate tool, so the response strategy is grounded in what the record shows rather than a guess.
The prosecution record shows where applications actually get stuck; it is not uniform across technology, so you can anticipate the likely rejection by art unit before you file.
Obviousness under Section 103 is the most common rejection in every technology center, appearing in about half to three quarters of office actions depending on the field, consistently well ahead of anything else. After obviousness, the second most common substantive rejection shifts by field: subject-matter eligibility under Section 101 in software, and novelty under Section 102 in semiconductors and electrical.
The point holds regardless of which version you run: the requirement you are most likely to collide with is knowable before you file, by art unit and by examiner, if you have the record to read it from.
Grounding through the Juristat Data Layer (MCP) uses roughly 85 percent fewer tokens than feeding raw office action PDFs to a model. Grounded data is not only more accurate, it is cheaper to run.
Juristat hands the model a clean, structured record, with registration numbers already resolved to the real attorneys and firms, so it reads the answer instead of parsing PDFs to reach it. What that saves you depends on how your AI tool charges for usage.
Enable Juristat with a single MCP connector from inside Solve, the same shape as connecting Data Layer in a tool like Claude through settings and connectors. Connecting Juristat to Solve help article. It is a configuration change, not a project, and it is available immediately with no waitlist.
The connection is read-only and request-response, stores no conversations, writes nothing back to your systems, shares nothing with third parties, and runs under SOC 2 Type II certification.
Juristat Data Layer (MCP) also runs with Claude, ChatGPT, Microsoft Copilot, and Perplexity, inside the Juristat application, and now inside Solve. You pick the workflow you want; Juristat supplies the record underneath it.
FROM A FIRM USING IT
“Data Layer turns AI from a promising concept into a practical, day-to-day patent tool.”
Rob Burger, COO, Sterne Kessler
Solve executes the work: purpose-built AI to draft and prosecute faster and with greater precision. Juristat supplies the record underneath it: decades of cleaned, structured prosecution data that show what has actually worked, by response type, rejection, and art unit. Whatever AI tool a patent team picks, it is only as good as the data underneath it, and Juristat is that layer, now inside Solve.
Juristat subscribers can connect their account to Solve, enable Juristat as a source, and bring examiner-level analytics directly into AI-assisted prosecution workflows.
If you use Solve and want to add Juristat, email solvepartnership@juristat.com. If you use Juristat and want to explore using Solve, email partnerships@solveintelligence.com.
Solve provides a chat-based workflow where you attach matter documents and ask targeted questions. Enabling Juristat adds examiner analytics as another source available inside that same workflow.
Add Juristat from inside Solve as an MCP connector. It is a configuration change rather than a project, and it is available immediately.
The connection is read-only and request-response. It stores no conversations, writes nothing back to your systems, shares nothing with third parties, and runs under SOC 2 Type II certification.
Juristat Data Layer (MCP) runs with Claude, ChatGPT, Microsoft Copilot, and Perplexity, inside the Juristat application, and now inside Solve Intelligence.
New to Juristat or Solve? See the integration in action.