Public Procurement Intelligence

Find government contract opportunities and qualify them before your team burns time on bad bids. This Actor aggregates procurement notices from SAM.gov, UK Contracts Finder, and TED, then scores each opportunity against your company profile.

Best for: founders, business development teams, bid managers, consultants, and agencies that need a ranked pipeline instead of another raw tender dump.

What This Actor Does

• Searches multiple procurement portals in one run
• Normalizes opportunities into one schema
• Scores each opportunity by:
  • keyword relevance
  • hard relevance gates that suppress obvious non-software work
  • NAICS / CPV code match
  • budget fit
  • preferred geography
  • set-aside compatibility
  • deadline urgency
• Optionally uses OpenAI to re-rank the strongest candidates and demote false positives
• Produces a ranked pipeline with clear next actions
• Exports JSON, CSV, and a Markdown management report

Why This Is Useful

Most tender tools stop at collection. The problem is not finding notices. The problem is deciding:

• Which ones are actually relevant?
• Which ones fit your contract size?
• Which ones should be reviewed now?
• Which ones should be ignored?

This Actor turns procurement search into go / no-go triage.

AI Review

The Actor can optionally use OpenAI to review the best rule-scored opportunities and apply semantic judgment to questions like:

• Is this actually software work, or just a generic procurement that happened to match one weak term?
• Is this a physical-goods or construction notice that should be demoted?
• Does the notice genuinely fit the company profile?

This is controlled by input flags. The Actor reads OPENAI_API_KEY from the environment, so users do not need to pass API credentials in Actor input.

Supported Sources

Source	Region	Notes
SAM.gov	United States	Federal procurement opportunities
Contracts Finder	United Kingdom	UK public sector contracts
TED	European Union	EU-wide procurement notices

Input Examples

Basic Search

{
  "query": "software development",
  "countries": ["us", "uk", "eu"],
  "maxItems": 60,
  "useAiReview": true
}

Company-Profile Qualification

{
  "query": "cybersecurity",
  "countries": ["us", "uk"],
  "onlyQualified": true,
  "minFitScore": 65,
  "companyProfile": {
    "companyName": "SecureOps",
    "keywords": ["cybersecurity", "SOC", "incident response", "threat detection"],
    "excludedKeywords": ["construction", "catering"],
    "naicsCodes": ["541512", "541519"],
    "preferredCountries": ["us", "uk"],
    "minContractValueUsd": 100000,
    "maxContractValueUsd": 5000000,
    "targetAgencies": ["Department of Defense", "NHS"]
  }
}

Small Business Focus

{
  "query": "data analytics",
  "countries": ["us"],
  "onlyQualified": true,
  "companyProfile": {
    "keywords": ["data analytics", "dashboard", "business intelligence"],
    "setAsideEligibility": ["small business", "8a"]
  }
}

Input Parameters

Parameter	Type	Default	Description
query	string	""	Search keywords. If empty, profile keywords are used.
countries	array	["us","uk","eu"]	Which procurement sources to search.
postedWithinDays	integer	30	Only include recently posted notices.
deadlineWithinDays	integer	60	Only include deadlines within N days when available.
status	array	["open"]	Opportunity statuses to include.
valueMin	number	none	Minimum contract value in USD.
valueMax	number	none	Maximum contract value in USD.
maxItemsPerSource	integer	30	Raw fetch limit per source.
maxItems	integer	90	Final ranked result limit.
onlyQualified	boolean	false	If true, only keep opportunities above minFitScore.
minFitScore	integer	60	Qualification threshold used when onlyQualified is enabled.
sortBy	string	fit_desc	fit_desc, deadline_asc, posted_desc, value_desc.
includeAnalytics	boolean	true	Include summary analytics.
outputFormat	string	full	full, summary_only, opportunities_only.
useAiReview	boolean	false	Use OpenAI to re-rank the strongest rule-scored opportunities.
aiReviewModel	string	gpt-5-mini	OpenAI model used for semantic review.
aiReviewMaxItems	integer	25	Maximum number of opportunities to send to AI review.
aiReviewMinRuleScore	integer	40	Minimum rule score required before AI review is attempted.
companyProfile	object	{}	Business profile used for qualification scoring.
proxy	object	Apify Proxy	Proxy configuration for browser-based sources.

Company Profile Fields

Field	Type	Description
companyName	string	Your company name for context
description	string	Short description of what you sell
keywords	array	Positive fit terms
excludedKeywords	array	Terms that count against fit
naicsCodes	array	Target US industry codes
cpvCodes	array	Target EU procurement codes
preferredCountries	array	Preferred geographies
minContractValueUsd	number	Smallest target contract size
maxContractValueUsd	number	Largest target contract size
setAsideEligibility	array	Eligible US set-aside programs
targetAgencies	array	Buyers you actively want

Output

The Actor writes two record types to the default dataset.

Summary Record

{
  "type": "summary",
  "summary": {
    "totalOpportunities": 42,
    "qualifiedOpportunities": 11,
    "highPriorityOpportunities": 3,
    "averageFitScore": 58.7,
    "deadlinesNext7Days": 4,
    "quickTake": "11 of 42 opportunities look viable. Best fit: \"Enterprise SOC Services\" (86/100).",
    "recommendedActions": [
      "Review 3 pursue-now opportunities immediately.",
      "Prioritize 4 opportunities closing within 7 days."
    ]
  }
}

Opportunity Record

{
  "type": "opportunity",
  "tenderId": "12345",
  "title": "Cybersecurity Monitoring Services",
  "country": "us",
  "source": "SAM.gov",
  "agency": {
    "name": "Department of Defense",
    "country": "us"
  },
  "value": {
    "cents": 250000000,
    "display": "$2,500,000",
    "currency": "USD"
  },
  "qualification": {
    "ruleFitScore": 71,
    "fitScore": 84,
    "qualification": "strong_fit",
    "recommendedAction": "pursue_now",
    "matchedKeywords": ["cybersecurity", "soc"],
    "matchedCodes": ["541512"],
    "riskFlags": ["Deadline is within 14 days."],
    "aiReview": {
      "model": "gpt-5-mini",
      "relevanceScore": 92,
      "decision": "strong_fit",
      "confidence": "high",
      "summary": "This is a genuine software/security services procurement."
    }
  }
}

Dataset Views

• Overview: one-row management summary
• Ranked Opportunities: all scored opportunities
• Priority Pipeline: best opportunities with signals and risks

Key-Value Store Outputs

File	Format	Description
OUTPUT	JSON	Main structured output
opportunities.csv	CSV	Spreadsheet export
REPORT.md	Markdown	Human-readable report

Qualification Logic

Each opportunity gets a fitScore from 0-100.

Positive signals:

• keyword matches
• software-delivery evidence in title/category/description
• NAICS / CPV code matches
• preferred geography
• budget within target range
• target buyer match
• compatible set-aside

Negative signals:

• excluded keywords
• construction / furniture / physical-goods language
• placeholder values such as £0 or £1
• weak generic matches like development without software evidence
• budget mismatch
• incompatible set-aside
• missing value data
• weak description
• very short deadline
• expired deadline

Typical Workflow

1. Define what your company actually sells.
2. Add contract size and geography preferences.
3. Run daily or weekly.
4. Review Priority Pipeline.
5. Export qualified opportunities to CRM or bid tracker.

Limitations

• Source websites can change HTML structure at any time.
• Some notices do not publish contract value or clean classification codes.
• Set-aside compatibility is heuristic and should be manually verified.
• This Actor ranks opportunities, but it does not replace bid/no-bid review by a human.

Apify Integration Notes

This project includes:

• .actor/input_schema.json
• .actor/output_schema.json
• .actor/dataset_schema.json
• .actor/actor.json
• typed TypeScript source
• Dockerfile for Apify build/runtime

Development

npm install
npm run build
npm run start

Local AI setup:

$cp .env.example .env

Then add:

$OPENAI_API_KEY=your_key_here

Example local input:

{
  "query": "software development",
  "countries": ["us", "uk", "eu"],
  "onlyQualified": false
}

Positioning

This is not just a tender scraper. It is a procurement qualification layer for companies that need a faster answer to:

Should we spend time on this bid?