Stage 2 of 7 - AI extraction

What did AI read from the presolicitation?

The AI extracted grounded facts and 14 evidence phrases from the public package.

Time saved by this AI pass

AI used the large-opportunity path to convert a dense public package into a normalized capture-review packet before humans started reading.

4h opportunity screening avoided

~2d manual document review avoided

13 AI chunk reviews

55 pages normalized

14 source phrases grounded

Processing path

Large-opportunity path selected

The package was too dense for a single safe review pass, so AI FlowOps inventoried the source files, classified document purpose, reviewed page-aware chunks, reconciled the results, and generated one normalized master packet before department routing.

3 source documents inventoried

55 pages normalized

13 page-aware chunks reviewed

1 normalized packet generated

Large path selected for production-realistic validation

Stage

Presolicitation Prepare for next solicitation step

Findings routed

14 Grounded findings routed downstream

Primary route

Legal Eligibility, data rights, OCI, export-control gates

Decision frame

Pursue? Lower bar than final bid/no-bid

Risk flags identified (8)

! This is a normal presolicitation gate, not a no-pursue blocker: capture should staff ... Implementation

! Classification handling is still a presolicitation unknown because DD254 guidance is ... Security

! Award vehicle and terms discretion Legal

! Implementation staffing is a serious pursuit risk if the delivery plan depends on non... Implementation

! OCI screening required Legal

! Funding and zero-award uncertainty Security

! DISA cloud authorization path Security

! Foreign participation restriction Legal

Show extracted data and source phrases

Structured fields passed downstream

This is the operational form of the AI extraction: values the workflow can route, audit, and hand off instead of leaving the RFP as unstructured prose.

{
  "ai_recommendation": "pursue_with_conditions",
  "bd_ops_decision": "pursue_with_conditions",
  "case_id": "FA875026S7002",
  "chunks_reviewed": 13,
  "complexity_score": 47,
  "decision_question": "Should BD/Ops pursue and prepare for the next solicitation step?",
  "departments_routed": [
    "legal",
    "security",
    "finance",
    "implementation"
  ],
  "evidence_items": 14,
  "normalized_packet_generated": true,
  "opportunity_stage": "presolicitation",
  "processing_mode": "large_normalized_packet"
}

Source evidence phrases (14)

The AI extracted 23 structured facts from the package. These are the cited source phrases used to ground the major findings and reviewer packets.

Large-path artifacts

These static artifacts were produced during local AI processing and exported into the read-only demo wrapper.

processing_profile.local.json
document_inventory.local.json
document_classification.local.json
chunk_reviews.local.json
extracted/ai_normalized_packet.local.md

<- Previous Next: recommendation ->