All Posts
FeatureFeb 25, 20269 min read

Prior Art Comparator: Side-by-Side Claim-to-Citation Mapping for Patent Prosecution

The most time-consuming part of responding to a 103 rejection is finding where the examiner says each claim element is taught by the prior art. The Prior Art Comparator eliminates that search entirely.

The Manual Prior Art Mapping Problem

When an examiner issues a 103 rejection, the Office Action contains a claim chart mapping each claim element to one or more prior art references. The problem is that this mapping is buried in dense prose -- not organized in a way attorneys can quickly act on.

The typical workflow looks like this:

1
Read the OA rejection section15-20 min

Parse the examiner's prose to identify which reference is cited for which claim element. Often requires re-reading multiple times.

2
Open each prior art reference5-10 min per ref

Navigate to the specific pages, paragraphs, or figures the examiner cited. With 3-5 references, this is 15-50 minutes.

3
Cross-reference to claim language20-40 min

For each element of each claim, verify whether the cited passage actually teaches what the examiner says it teaches. This is where most arguments originate.

4
Identify gaps and arguments15-30 min

Find where the examiner's mapping is weak -- elements poorly mapped, passages that do not actually teach the limitation, combination motivations that are unsupported.

Total manual time: 55-140 minutes per rejection. For a single independent claim with 3 references. Multiply by the number of independent claims and rejection grounds.

How the Prior Art Comparator Works

The Prior Art Comparator extracts the examiner's claim-to-reference mapping from the Office Action text and presents it as a structured side-by-side view:

Prior Art Comparator View (example)
Claim 1, Element [a]

"a processor configured to receive a plurality of sensor inputs..."

Smith (US 10,xxx,xxx), col. 4, ll. 12-28

"the controller 102 receives data from sensors 104a-104d through bus 106..."

Confidence: 87%
Claim 1, Element [b]

"applying a machine learning model to generate a prediction..."

Jones (US 2024/0xxx), para. [0047]-[0052]

"neural network 210 processes the feature vector to produce output classification..."

Confidence: 62% -- Potential argument opportunity

Automatic Extraction

The AI parses the examiner's rejection text and identifies which specific passages from which references are cited for each claim element. No manual search required.

Side-by-Side View

Your claim language on the left, the examiner's cited passage on the right. Element by element. For every independent claim.

Confidence Scoring

Each mapping includes a confidence score. Low-confidence mappings indicate where the examiner's citation may not fully teach the claim element -- these are your argument opportunities.

Glass Box Verification

Every extracted passage is verified against the actual prior art document. If the AI cannot locate the exact passage, the mapping is flagged rather than fabricated.

Time Savings: Before and After

TaskManualWith ComparatorSaved
Parse examiner's claim chart15-20 minAutomatic15-20 min
Locate cited passages in references15-50 minAutomatic15-50 min
Cross-reference to claim elements20-40 minAutomatic (side-by-side)20-40 min
Identify argument opportunities15-30 min5-10 min (confidence scores)10-20 min
Total65-140 min5-10 min55-130 min

No Other Patent AI Tool Does This

Other patent AI tools provide rejection summaries, claim analysis, and amendment suggestions. None of them extract the examiner's specific citation passages and display them side-by-side with your claim language at the element level.

FeatureAbigailOthers
Side-by-side claim vs. citation view
Element-level mapping
Exact passage extraction from references
Per-mapping confidence scores
Argument opportunity identification
Glass Box verification of citations

See Your Prior Art Mapped Side-by-Side

Upload an Office Action with a 103 rejection and see the Prior Art Comparator extract and map every citation to every claim element.

Frequently Asked Questions

Related Guides