Prosecution Insights
Prosecution Insights
Technology area: Communications
2 pending office actions • 1 art units • 2 examiners • 0 of 2 (0%) have an AI response strategy ready
Rohirrim Inc. has 2 pending office actions in the Communications technology area. Both actions are located in 1 distinct art unit. This total concentration in 1 art unit suggests that the company's current filings are focused on a very specific set of technical challenges within the communications field.
There are 2 distinct examiners assigned to these actions. CHEN, XUEMEI G is the busiest examiner with 1 pending action. Each action is being reviewed by a different examiner within the same 1 art unit. This distribution ensures that while the technical focus is narrow, the workload is split between two different individuals.
With 2 pending actions and 2 examiners in 1 art unit, the company has a very focused prosecution front. For a practitioner, this environment offers an opportunity to study the specific preferences of a single art unit, as both pending actions will be subject to the same internal policies. However, the use of 2 different examiners means that each action must still be managed as an independent matter.
Based on the USPTO statutory response window for each pending office action. 1 of the docket's apps have a known mailing date; the rest are excluded from the tile counts.
Difficulty is derived from the rejection statutes on the most recent pending office action. §101-driven and multi-statute cases are graded Hard; §112-only and obviousness-type double-patenting cases are graded Easy; everything else is Medium. "Unknown" means we have not yet parsed a statute for that office action.
| Bucket | Cases |
|---|---|
| §103 only | 1 (50%) |
| Double-patenting only | 1 (50%) |
How the docket's pending cases split across USPTO tech-center bands.
Manual office-action response work runs about 10 hours per case. The time-saved bands below show what IP Author's prosecution pipeline typically delivers — a conservative 20% on the low end, 35% in the middle, 50% on the high end.
| Examiner | Apps on this docket | Allow rate | Interview lift |
|---|---|---|---|
| CHEN, XUEMEI G | 1 | 77.0% | +25.6% |
| SHAH, PARAS D | 1 | 73.4% | +31.0% |
Cases in front of an examiner whose interview lift is 10 percentage points or more — i.e. interviewed cases historically resolve more favorably than non-interviewed ones. The top 2 ordered by deadline are shown.
| App # | Title | Examiner | Due in |
|---|---|---|---|
| 19266519 | EXTRACTING IMAGES AND DETERMINING THEIR MEANING FOR SEMANTIC IMAGE RETRIEVAL AND TRAINING A TRANSFORMER-BASED MULTI-MODAL LARGE LANGUAGE MODEL TO GENERATE DOMAIN-AWARE IMAGES BASED ON IMAGE MEANINGS | CHEN, XUEMEI G | 46d |
| 19171562 | COMPUTER-GENERATED CONTENT BASED ON TEXT CLASSIFICATION, SEMANTIC RELEVANCE, AND ACTIVATION OF DEEP LEARNING LARGE LANGUAGE MODELS | SHAH, PARAS D | — |
| Art Unit | Apps |
|---|---|
| 2661 | 1 |
| App # | Title | Examiner | Art Unit | Statutes | Status | Due in | AI | Filed |
|---|---|---|---|---|---|---|---|---|
| 19266519 | EXTRACTING IMAGES AND DETERMINING THEIR MEANING FOR SEMANTIC IMAGE RETRIEVAL AND TRAINING A TRANSFORMER-BASED MULTI-MODAL LARGE LANGUAGE MODEL TO GENERATE DOMAIN-AWARE IMAGES BASED ON IMAGE MEANINGS | CHEN, XUEMEI G | 2661 | §103 | Non-Final OA | 46d | Pending | Jul 11, 2025 |
| 19171562 | COMPUTER-GENERATED CONTENT BASED ON TEXT CLASSIFICATION, SEMANTIC RELEVANCE, AND ACTIVATION OF DEEP LEARNING LARGE LANGUAGE MODELS | SHAH, PARAS D | — | DP | Non-Final OA | — | Pending | Apr 07, 2025 |
IP Author helps IP teams respond to office actions faster with AI-generated responses, examiner analytics, and prosecution intelligence.
Start Free Trial