tender-assessment

Tender Assessment Skill

Automated tender discovery, alignment scoring, and pursuit package generation for marcov's government tender opportunities.

Overview

This skill provides an end-to-end tender assessment workflow:

1. Scrape - Fetch current open tenders from tenders.vic.gov.au
2. Score - Apply weighted alignment matrix against marcov's capabilities
3. Shortlist - Identify tenders ≥80% aligned
4. Deep Assess - Generate full pursuit packages for shortlisted opportunities

What This Skill Delivers

• Shortlist report with all assessed tenders and recommendations
• Full pursuit packages for high-alignment opportunities including:
  • Go/No-Go recommendation with rationale
  • Competitive positioning and win themes
  • Draft response outline
  • Team requirements
  • Risk matrix
  • Pricing considerations

---

Reference Files

Load these before processing:

• Company Profile: references/company-profile.md - marcov's capabilities, industries, constraints
• Scoring Matrix: references/scoring-matrix.md - Weighted alignment criteria
• Pursuit Package Template: references/pursuit-package-template.md - Deep assessment output format
• Shortlist Report Template: references/shortlist-report-template.md - Summary report format

---

Discovery Process

Before running the assessment, confirm the following with the user:

Required Context

1. Scraper location: Where is vic_tenders_scraper.py located?

   • Default: Current working directory
   • May need full path

2. Keyword filtering: Use default marcov keywords or custom?

   • Default keywords cover asset management, reliability, rail, etc.
   • User may want to add/remove keywords

3. New tenders only?:

   • --new-only flag for last 24 hours
   • --track-new to compare against previously seen
   • Or all currently open tenders

4. Output preferences:

   • Shortlist report only
   • Shortlist + pursuit packages for all ≥80%
   • Shortlist + pursuit packages for top N only

Skip Discovery If

User provides explicit instructions like:

• "Run the tender assessment on all current tenders"
• "Assess new tenders from today"
• "Generate pursuit packages for the top 3 aligned tenders"

---

Workflow

Phase 1: Scrape Tenders

Execute the Victorian tenders scraper to get current open opportunities.

# Default: All open tenders with marcov keywords
python vic_tenders_scraper.py --all-keywords --pretty

# New tenders only (last 24h)
python vic_tenders_scraper.py --all-keywords --new-only --pretty

# Track genuinely new (vs previously seen)
python vic_tenders_scraper.py --all-keywords --track-new --pretty

Output: JSON array of tender objects with:

• id, rfx_number, title, issuer
• categories (UNSPSC)
• date_opened, date_closing
• url (link to tender detail page)
• matched_keywords (if keyword filtering applied)

Phase 2: Initial Assessment (Scoring)

For each tender, calculate alignment score using the weighted matrix:

Dimension	Weight	Assessment Basis
Domain Fit	30 pts	Keywords, title, categories vs marcov services
Industry Match	25 pts	Issuer, sector vs marcov target industries
Service Type	20 pts	Tender type, scope vs marcov offerings
Strategic Value	15 pts	Client tier, sector priority, reference potential
Competitive Position	10 pts	Relationship, incumbent status, differentiators

Scoring Guidelines

Domain Fit (30 points):

• 30: Direct match to core service (RCM, ISO 55001, asset strategy, predictive maintenance)
• 25: Adjacent service (CMMS, condition monitoring, reliability)
• 20: Related domain (data analytics for assets, maintenance optimisation)
• 15: Tangential but achievable (general consulting in our industries)
• 10: Requires stretch or partnering
• 0: Outside domain entirely

Industry Match (25 points):

• 25: Rail, rolling stock, tram, public transport
• 22: Water, wastewater, utilities
• 20: Energy (generation, T&D)
• 18: Mining, resources
• 15: Healthcare facilities
• 12: Local government
• 10: Defence, federal
• 5: Other
• 0: Outside scope (IT, construction, retail)

Service Type (20 points):

• 20: Advisory/strategy (consulting, review, assessment)
• 18: Technical analysis (RCM, FMEA, reliability studies)
• 15: Implementation support (change management, training)
• 12: Hybrid advisory + operational
• 8: Managed services
• 5: Labour hire
• 0: Not applicable (construction, manufacturing)

Strategic Value (15 points):

• 15: Opens new Tier 1 client, reference project, growth sector
• 12: Strengthens existing relationship or sector presence
• 8: Profitable but no strategic uplift
• 4: Keeps team busy, no lasting benefit
• 0: Could damage brand or distract

Competitive Position (10 points):

• 10: Known to client, no incumbent, our niche
• 8: Some relationship, weak incumbent
• 6: Open competition, level field
• 4: Strong incumbent, we're unknown
• 2: Incumbent has advantage, no differentiator
• 0: Wired for competitor, missing credentials

Phase 3: Categorise Results

Based on total score:

Score	Category	Action
≥80	Shortlisted	Generate full pursuit package
60-79	Flagged	Include in report for manual review
<60	Declined	Log with reason, no further action

Override Triggers

Upgrade to review even if <80:

• Strategic target client (Tier 1 government agency)
• Value >$500K
• Opens door to growth sector (water, defence)

Flag for review even if ≥80:

• Closing <7 days (resource risk)
• Estimated value <$20K (effort vs return)
• Requires unverified certifications

Phase 4: Generate Shortlist Report

Create summary report using references/shortlist-report-template.md:

1. Dashboard with counts and metrics
2. Shortlisted tenders with quick assessments
3. Flagged tenders with review rationale
4. Auto-declined tenders with reasons
5. Upcoming deadline warnings
6. Recommended actions

Phase 5: Deep Assessment (Pursuit Packages)

For each shortlisted tender (≥80%):

5.1 Fetch Full Tender Details

Use WebFetch to retrieve the tender detail page:

URL: {tender.url}
Prompt: Extract all tender details including:
- Full description and scope of work
- Mandatory requirements
- Evaluation criteria
- Contract value (if stated)
- Contract duration
- Key dates (briefing, Q&A, submission)
- Required certifications or qualifications
- Incumbent information (if mentioned)
- Any attachments or documents referenced

5.2 Generate Pursuit Package

Using references/pursuit-package-template.md, create comprehensive assessment:

1. Executive Summary: Recommendation, win probability, strategic importance
2. Opportunity Overview: Scope, dates, deliverables
3. Go/No-Go Analysis: Reasons to pursue, reasons for caution
4. Competitive Positioning: Differentiators, win themes, competitor analysis
5. Response Strategy: Approach, key points, weaknesses to address
6. Draft Response Outline: Proposed structure, proof points
7. Team Requirements: Proposed team, capacity check, subcontracting needs
8. Risk Assessment: Bid risks, delivery risks, commercial risks
9. Pricing Considerations: Preliminary estimate, pricing strategy
10. Next Steps: Immediate actions, response timeline

---

Output Formats

Option 1: Shortlist Report Only

Generate tender-shortlist-[DATE].md with:

• Summary dashboard
• Categorised tender list
• Quick assessments
• Recommended actions

Option 2: Full Assessment Package

Generate:

1. tender-shortlist-[DATE].md - Summary report
2. pursuit-[RFX_NUMBER].md - One file per shortlisted tender

Option 3: HTML Report

Generate interactive HTML with:

• Filterable tender table
• Expandable assessment cards
• Deadline countdown timers

---

Example Usage

Basic Assessment

User: "Run tender assessment"
Action:
1. Run scraper with default keywords
2. Score all tenders
3. Generate shortlist report
4. Generate pursuit packages for ≥80% aligned

New Tenders Only

User: "Check for new tenders today"
Action:
1. Run scraper with --new-only flag
2. Score new tenders only
3. Generate shortlist report

Specific Tender Deep Dive

User: "Generate a pursuit package for tender RFT-12345"
Action:
1. Fetch tender details from URL
2. Score against matrix
3. Generate full pursuit package regardless of score

---

Company Profile Summary

marcov (SAS Asset Management) specialises in:

Core Services

• Asset management strategy (ISO 55001, GFMAM)
• Reliability engineering (RCM, FMEA, RCA)
• Predictive maintenance and analytics
• CMMS/EAM advisory (Maximo, SAP PM)
• Condition monitoring strategy

Target Industries (Priority)

1. Rail & rolling stock
2. Public transport (tram, bus)
3. Water & wastewater
4. Energy (generation, T&D)
5. Mining & resources
6. Healthcare facilities
7. Local government
8. Defence

Constraints

• Team: ~11 FTE
• Sweet spot: $50K-$500K contracts
• Geographic: Victoria primary, national capability
• Capacity: 8-12 concurrent projects

Differentiators

• Practitioner-led (worked in maintenance roles)
• Technology/vendor agnostic
• Sovereign AI capability (edge/on-premise)
• Deep ISO 55001 and GFMAM expertise
• Rail specialism

See references/company-profile.md for full details.

---

Scoring Matrix Quick Reference

Dimension	Weight	Top Score Criteria
Domain Fit	30	Direct match to core service
Industry Match	25	Rail/transport sector
Service Type	20	Advisory/strategy work
Strategic Value	15	Opens Tier 1 client or growth sector
Competitive Position	10	Known to client, our niche

Pass threshold: ≥80 points

See references/scoring-matrix.md for full rubric.

---

Integration Notes

Scraper Location

The skill expects vic_tenders_scraper.py to be accessible. If not in the current directory, ask the user for the path.

State Management

The scraper supports --track-new with a state file (seen_tender_ids.json) to identify genuinely new tenders. This persists across runs.

API Usage

Deep assessment uses Claude API (via the skill execution context) to:

• Interpret tender requirements
• Generate pursuit strategy
• Draft response outlines

No external API keys required beyond the Claude Code session.

---

Error Handling

Scraper Fails

• Check network connectivity
• Verify tenders.vic.gov.au is accessible
• Review scraper logs for specific errors

WebFetch Fails

• Tender detail page may require authentication
• Fall back to public information only
• Flag for manual document download

Insufficient Information

• If tender details are sparse, generate partial pursuit package
• Flag gaps requiring manual research
• Note limitations in output

---

This skill is designed for marcov's Victorian government tender pipeline. Adjust the company profile and scoring matrix for different organisations or jurisdictions.