EcoIQ Microsoft Ecosystem Core Stack

Use this layer for EcoIQ agents that read company reports, compare industrial processes, recommend upgrades, generate investor briefings and coordinate multi-agent workflows.

• microsoft/agent-framework — Production-grade AI agents and multi-agent workflows in Python/.NET
• microsoft/semantic-kernel — Plugins, workflows, memory and enterprise AI orchestration
• microsoft/promptflow — Prompt testing, output comparison and report quality evaluation
• microsoft/graphrag — Knowledge graph from company reports, ESG files, ministries, suppliers and industrial assets
• microsoft/markitdown — Convert PDFs, Word, PowerPoint, Excel and webpages into clean Markdown for AI ingestion
• microsoft/LLMLingua — Compress long reports before sending to models
• microsoft/TaskWeaver — Data-heavy analysis for emissions, costs, ROI, CAPEX and efficiency

EcoIQ use case: Upload annual report → extract data → build company profile → score climate and industrial readiness → generate investor-ready recommendations.

Make EcoIQ answers evidence-based, not AI guesses.

• Azure-Samples/azure-search-openai-demo — RAG template using Azure OpenAI and Azure AI Search
• Azure-Samples/azure-search-vector-samples — Vector search across filings, reports, regulations and supplier documents
• Azure-Samples/document-intelligence-code-samples — Extract tables, forms, text and scanned documents
• Azure-Samples/azure-ai-search-power-skills — Custom enrichment for emissions, industry tags and risk labels
• Azure-Samples/azure-ai-studio-secure-bicep — Secure Azure AI deployment pattern

EcoIQ use case: Every EcoIQ recommendation shows source, confidence, evidence link, data age, risk level and missing data.

Use this layer for gas, uranium, oil, mining, boilers, metals, agriculture, electricity, factories and district heating systems.

• Azure/Industrial-IoT — Connect industrial equipment, OPC UA, factories and energy assets
• Azure/iotedge — Run monitoring at factories, mines, boilers, farms and remote sites
• Azure/azure-iot-sdk-python — Python device data ingestion
• Azure-Samples/azure-digital-twins-samples — Model factories, boilers, mines, farms and energy grids
• Azure/azure-digital-twins-explorer — Visualise industrial assets and relationships
• Azure/opendigitaltwins-energygrid — Electricity, substations and grid assets
• Azure/opendigitaltwins-manufacturing — Production lines, machines and processes
• Azure/opendigitaltwins-building — Boilers, heating systems, buildings and sensors
• Azure/opendigitaltwins-agriculture — Farms, irrigation, soil, greenhouses and livestock

EcoIQ use case: Create a digital twin of a coal boiler system → compare with electric boiler, heat pump or district heating upgrade → estimate CO₂ reduction, cost, payback and health impact.

Use this layer for investors, akimats, ministries, industrial companies and Microsoft-style dashboards.

• microsoft/fabric-samples — Lakehouse, analytics and enterprise data platform
• microsoft/SynapseML — Large-scale ML on industrial and climate datasets
• microsoft/PowerBI-JavaScript — Embed dashboards into ecoiq.uk
• microsoft/powerbi-client-react — React dashboard integration
• microsoft/PowerBI-visuals — Custom visuals for emissions, ROI and company rankings
• microsoft/DataConnectors — Connect external datasets, ministries and ESG portals
• microsoft/sql-server-samples — Enterprise SQL architecture
• Azure/azure-sdk-for-python — Connect EcoIQ backend to Azure services

EcoIQ use case: Public dashboard: Kazakhstan industrial modernisation map. Private dashboard: company-level transition plan, ROI, emissions, Maqasid/Mizan score and investor readiness.

EcoIQ scores companies, governments and investment opportunities, so scoring must be explainable, auditable and safe.

• microsoft/responsible-ai-toolbox — Understand, assess and monitor AI systems
• microsoft/fairlearn — Check fairness of scoring models
• microsoft/interpret — Explain why a company got a score
• microsoft/presidio — Detect and remove personal/private data from documents
• microsoft/PyRIT — Red-team prompts and agents for unsafe outputs
• microsoft/garak — Scan LLM vulnerabilities

EcoIQ use case: Every company score should answer: Why this score? Which evidence? Which assumptions? Which data is missing? What requires human review?

Give EcoIQ a stronger climate-tech foundation.

• Green-Software-Foundation/carbon-aware-sdk — Emissions-aware scheduling and carbon-aware software
• microsoft/Cloud-for-Sustainability-API-Samples — Connect sustainability data and reporting workflows
• microsoft/PlanetaryComputerExamples — Climate, satellite and geospatial data
• microsoft/torchgeo — Geospatial ML for land, agriculture, mining and environment
• microsoft/Project15 — IoT for conservation and environmental monitoring

EcoIQ use case: Satellite + company data + emissions + industrial assets → climate risk and transition opportunity map.