Overview

/presales-analyze is a production command in the Presales category of the Prismatic Platform that performs comprehensive analysis of text, files, and URLs to assess presales opportunities. The command ingests unstructured business information from multiple source types -- including RFPs, tender documents, company websites, press releases, financial reports, and email correspondence -- and produces structured opportunity assessments with scoring, risk factors, and strategic recommendations.

The analysis engine applies natural language processing, entity extraction, and domain-specific heuristics to identify key opportunity characteristics: project scope, budget indicators, timeline constraints, technical requirements, competitive landscape signals, and decision-maker information. These extracted signals are then evaluated against configurable qualification criteria to produce an opportunity score that quantifies the likelihood of successful engagement and the strategic value of pursuing the opportunity.

This command operates under the L2+ authority level and is executed by the opportunity-analyzer agent. It is part of the platform's 216-command slash command registry, built on the AIAD (Autonomous Intelligence Agent Design) standard. The opportunity analyzer agent specializes in processing heterogeneous business documents and extracting actionable intelligence for sales strategy formulation.

Within the presales workflow, /presales-analyze typically serves as the first analytical step after an opportunity is identified. Raw materials such as tender documents, RFP PDFs, or prospect website URLs are fed into the analyzer, which produces a structured assessment that feeds into downstream commands like /presales-assess for technical evaluation, /presales-price for pricing strategy, and /presales-propose for proposal generation.

Architecture

The analysis architecture implements a multi-format ingestion pipeline that normalizes diverse input sources into a common document model before applying analytical extractors.

Input Sources                  Processing Pipeline              Output
┌──────────────┐         ┌──────────────────────┐        ┌─────────────┐
│ Raw Text     │────>    │ Format Normalizer    │        │ Opportunity │
│ PDF/DOCX     │────>    │ (Text Extraction)    │        │ Assessment  │
│ URL Content  │────>    │         │            │        │ Report      │
│ Email Body   │────>    │         v            │        │             │
│ CSV/Excel    │────>    │ Entity Extractor     │────>   │ - Score     │
└──────────────┘         │ (NER + Domain Rules) │        │ - Risks     │
                         │         │            │        │ - Strategy  │
                         │         v            │        │ - Entities  │
                         │ Signal Classifier    │        │ - Timeline  │
                         │ (Qualification)      │        │ - Budget    │
                         │         │            │        │ - Competitors│
                         │         v            │        └─────────────┘
                         │ Opportunity Scorer   │
                         │ (Multi-factor)       │
                         └──────────────────────┘

The Entity Extractor identifies and classifies key entities within the document: organization names, person names with roles, monetary amounts with currency, dates and deadlines, technology mentions, and geographic references. Each extracted entity is tagged with confidence scores and source position, enabling traceability back to the original document.

The Signal Classifier evaluates extracted entities and patterns against presales qualification criteria. Positive signals (large budget mentions, urgent timelines, technology alignment) increase the opportunity score. Negative signals (incumbent vendor mentions, budget constraints, scope creep indicators) decrease it. The classifier operates on configurable rule sets that can be tuned for specific market segments and engagement models.

Usage

Text Analysis

# Analyze raw text description of opportunity
/presales-analyze --text "Company XYZ is looking for a cybersecurity platform..."

# Analyze text from clipboard
/presales-analyze --clipboard

# Analyze with specific qualification criteria
/presales-analyze --text "..." --criteria enterprise-security

File Analysis

# Analyze RFP document
/presales-analyze --file /path/to/rfp-document.pdf

# Analyze multiple files for same opportunity
/presales-analyze --files "rfp.pdf,requirements.docx,budget.xlsx"

# Analyze with specific focus areas
/presales-analyze --file tender.pdf --focus technical-requirements

URL Analysis

# Analyze company website for opportunity signals
/presales-analyze --url "https://example.com/about"

# Analyze tender portal listing
/presales-analyze --url "https://tender-portal.gov.cz/listing/12345"

# Deep crawl with link following
/presales-analyze --url "https://example.com" --crawl-depth 3

Combined Analysis

# Multi-source opportunity analysis
/presales-analyze --file rfp.pdf --url "https://example.com" \
  --text "Additional context from meeting notes" --combine

# Generate structured opportunity brief
/presales-analyze --file rfp.pdf --output opportunity-brief.json --format json

Options & Parameters

Parameter	Type	Default	Description
`--text`	string	none	Raw text to analyze
`--file`	path	none	File path for document analysis
`--files`	string	none	Comma-separated file paths
`--url`	URL	none	URL to fetch and analyze
`--clipboard`	flag	false	Analyze text from system clipboard
`--crawl-depth`	integer	1	URL crawling depth for link following
`--criteria`	enum	default	Qualification criteria set
`--focus`	enum	all	Analysis focus: all, technical-requirements, budget, timeline, competitive
`--combine`	flag	false	Combine multi-source analysis into single assessment
`--case-id`	string	auto	Associate analysis with existing presales case
`--format`	enum	table	Output: table, json, html, pdf, markdown
`--output`	path	stdout	Output file path
`--language`	enum	auto	Source language: auto, cs, en, de
`--extract-contacts`	flag	false	Extract and structure contact information
`--competitive-analysis`	flag	false	Enable competitive landscape analysis
`--confidence-threshold`	float	0.6	Minimum confidence for entity inclusion

Execution Flow

The analysis pipeline processes input through four sequential phases with parallel sub-operations within each phase.

Phase 1 -- Ingestion and Normalization (1-10 seconds): Input sources are ingested and converted to a common text representation. PDF documents are processed through text extraction with layout preservation. DOCX files are parsed for structured content. URLs are fetched with JavaScript rendering support for dynamic content. Multiple sources are concatenated with source boundary markers for traceability.

Phase 2 -- Entity Extraction (2-5 seconds): The normalized text is processed through the entity extraction pipeline. Named Entity Recognition identifies organizations, persons, locations, and monetary values. Domain-specific extractors identify technology mentions, compliance references, timeline expressions, and competitive signals. Each entity is annotated with confidence score, source position, and entity type.

Phase 3 -- Signal Classification (1-3 seconds): Extracted entities and document patterns are evaluated against qualification criteria to classify opportunity signals. The classifier evaluates budget adequacy (does the indicated budget align with solution pricing?), timeline feasibility (can the solution be delivered within the stated timeframe?), technical fit (does the requirement align with platform capabilities?), and competitive position (is there an incumbent or preferred vendor?).

Phase 4 -- Scoring and Report Generation (< 2 seconds): The multi-factor opportunity score is calculated by weighting classified signals according to the selected criteria set. The final assessment report is generated with structured sections: executive summary, opportunity score with breakdown, extracted entities, risk factors, strategic recommendations, and recommended next steps (including which presales commands to execute next).

Integration Points

Component	Integration Type	Description
/presales	Parent Command	Top-level presales orchestration
/presales-assess	Downstream	Technical assessment of analyzed opportunity
/presales-price	Downstream	Pricing based on analyzed requirements
/presales-propose	Downstream	Proposal generation from analysis
/presales-case	Case Management	Associate analysis with presales case
/presales-research	Complementary	Competitor research to enrich analysis
Prismatic Agents	Execution	`opportunity-analyzer` agent
Telemetry	Observability	Analysis timing and quality metrics
Quality Gates	Validation	Analysis completeness checks

Best Practices

Provide Multiple Sources: The analysis is significantly more accurate and comprehensive when multiple source types are combined. An RFP document plus the prospect's website plus meeting notes will produce a much richer assessment than any single source alone.

Set the Right Criteria: Use --criteria enterprise-security for cybersecurity opportunities, --criteria data-analytics for data platform opportunities, and so on. The criteria set adjusts signal weights to match the specific engagement model, significantly improving scoring accuracy.

Extract Contacts Early: Use --extract-contacts during initial analysis to capture decision-maker information while it is fresh. This structured contact data feeds directly into follow-up workflows and CRM integration.

Associate with Cases: Always use --case-id to link analyses to existing presales cases. This creates a complete analytical history for each opportunity, enabling trend tracking and decision audit trails.

Review Confidence Scores: Pay particular attention to entities with confidence scores below 0.8. These may represent ambiguous extractions that require manual verification before strategic decisions are based on them.

Error Handling

Error Condition	Behavior	Recovery
Unsupported file format	Error with supported formats list	Convert to PDF/DOCX/TXT
URL fetch failure	Error with HTTP status	Verify URL accessibility
Empty document	Warning with no analysis	Verify file has extractable content
Language detection failure	Fallback to English processing	Specify `--language` explicitly
Extraction timeout	Partial results with warning	Reduce document size or scope
Low-confidence results	Warning with confidence distribution	Provide additional source material

Advanced Usage

Pipeline Integration

# Full presales pipeline from initial document
/presales-analyze --file rfp.pdf --case-id "CASE-2026-042" \
  | /presales-assess --technical \
  | /presales-price --model subscription \
  | /presales-propose --template enterprise

Batch Opportunity Screening

# Screen multiple tender documents
/presales-analyze --files "tender1.pdf,tender2.pdf,tender3.pdf" \
  --criteria enterprise-security --format json --output screening-results.json

# Rank opportunities by score
/presales-analyze --batch tenders/ --rank --top 5

Competitive Intelligence Enhancement

# Deep competitive analysis from RFP
/presales-analyze --file rfp.pdf --competitive-analysis \
  --cross-reference /presales-research --format pdf

Doctrine Compliance

All commands operate under the NO MERCY, NO DOUBTS doctrine:

NO MERCY: Zero tolerance for incomplete execution or quality violations. Every section of the input document is processed. No extraction failures are silently suppressed. Opportunity scores reflect all identified signals, including negative ones.
NO DOUBTS: Full investigation before action, evidence-based results. Every extracted entity includes source provenance and confidence scoring. The NABLA axiom of Provenance Mandatory ensures all analytical claims are traceable to specific source content. Contradictory signals (e.g., large stated budget but constrained actual spending patterns) are preserved and highlighted per the Contradiction Preservation axiom.

/presales - Presales intelligence for company analysis and opportunity identification
/presales-assess - Technical assessment of opportunities and cases
/presales-case - Presales case management for status tracking and updates
/investigate - Launch comprehensive OSINT investigation across 121+ sources
/email-osint - Email-based OSINT gathering with breach correlation and social profiling
/google-hacking - Google dorking and advanced search intelligence extraction

Connect & Contribute

Created by Tomáš Korcak (korczis) | Open Source under GHL

GitHub | GitLab | LinkedIn | Contact
Developer Portal | Architecture | Meet the Creator

/presales-analyze