Turn messy construction and infrastructure data into report ready intelligence.
Yeresi AI helps asset-management, facilities, and infrastructure teams clean, tag, reconcile, and query scattered project files, PDFs, photos, spreadsheets, inspection reports, and handover documents.
Built from firsthand experience organizing unstructured construction asset-management data for enterprise EAM workflows.
17 assets missing required inspection documents.
Sources cited across 42 files.
Asset data is everywhere. Useful knowledge is nowhere.
Construction and infrastructure teams generate massive amounts of data across project lifecycles. But critical information often lives in disconnected folders, PDFs, photos, spreadsheets, emails, inspection reports, and handover packages.
Scattered project records
Files live across SharePoint, local folders, email, PDFs, spreadsheets, and vendor systems.
Manual reporting pain
Teams spend hours searching, copying, reconciling, and validating information for reports.
Poor EAM readiness
AI and EAM workflows fail when the underlying asset data is incomplete, unstructured, or unreliable.
No source of truth
Teams struggle to trust what is current, complete, and tied back to original documentation.
From messy files to auditable asset intelligence.
Yeresi creates a structured knowledge layer from the documents and media your team already has.
Ingest
Bring in project folders, PDFs, spreadsheets, photos, inspection reports, handover packages, and asset records.
Extract
Identify assets, locations, dates, vendors, document types, inspection details, issues, and missing fields.
Structure
Tag, reconcile, and organize information into a searchable asset/project knowledge layer.
Query
Ask questions, generate reports, identify missing documentation, and cite every answer back to source files.
Yeresi does not replace your EAM, SharePoint, GIS, or ERP systems. It helps prepare messy operational data so those systems become more useful.
Start with one messy project folder.
We are running limited data readiness pilots with teams that want to understand how much operational value is trapped in their unstructured project and asset data.
2-Week Data Readiness Pilot
A focused engagement to map what's in your data — and what it could become.
- One messy project folder or asset dataset
- PDFs, spreadsheets, reports, photos, logs, or handover docs
- Any current reporting or search workflow pain
- Structured file and data inventory
- Extracted assets, locations, dates, vendors, reports, issues, and document types
- Searchable/queryable knowledge layer
- Source-cited answers back to original files
- Missing documentation and data quality findings
- Before/after workflow analysis
- Rollout recommendation
Best fit for construction handover, inspection docs, asset records, facilities documentation, infrastructure project files, or EAM migration prep.
Request a PilotBuilt for the messy middle between field data and enterprise systems.
Construction handover packages
Validate, organize, and query project handover documents before they disappear into static folders.
Inspection report search
Find reports, dates, asset references, issues, and source documents across messy inspection archives.
Asset documentation cleanup
Identify missing manuals, warranties, photos, reports, and metadata tied to specific assets.
EAM migration readiness
Prepare unstructured project and asset information before it enters Maximo, GIS, CMMS, or other asset platforms.
Facilities knowledge retrieval
Give operations teams faster access to the documents, records, and answers they need.
Project reporting support
Reduce time spent manually compiling evidence, summaries, and source references for status reports.
AI is only as useful as the data underneath it.
Asset-heavy organizations want AI, automation, and EAM modernization. But those systems depend on clean, trustworthy, structured operational data. Yeresi focuses on the upstream data-readiness problem: turning fragmented project and asset records into a foundation teams can actually use.
Unstructured data is blocking automation
Most operational records are not in a form that AI, analytics, or modern EAM workflows can actually consume.
Teams need source-cited answers
Confidence depends on traceability. Every answer must point back to the original file, page, and record.
EAM initiatives need cleaner data
Asset platforms only deliver value when the underlying documentation, hierarchies, and metadata are reliable.
Founder-led. Field-informed. Engineered for mission-critical data.
Yeresi AI is founded by Kobena Idun, a 2x founder and 2x YC startup engineer who has shipped production AI infrastructure and built mission-critical, data-heavy systems at teams under asset management at Fidelity Investments. Yeresi's vision is born out of firsthand experience structuring large volumes of unstructured financial, construction, and asset management data.
- 2× Early Engineer at YC-backed startups (CollectWise, Vellum)
- Software Data Engineer at Fidelity Investments
- Enterprise Asset Management Development at Symetri
- $10K IDEA Gap Fund recipient · Northeastern Venture Accelerator
- Llama-Node contributor · Ava Labs research lead at NEU Blockchain
- Northeastern Khoury — CS & Business
- Founding EngineerCollectWise (YC F24)
Building TypeScript infrastructure that injects context into voice agents driving $1M+ ARR in collections; shipped an AI paralegal workflow that cut document redaction from 28 days to 4 hours.
- Software EngineerVellum AI (YC W23)
Shipped a monitoring system inside Vellum's Python SDK for agentic workflow observability and cut event-emission latency 20–100ms via a Kubernetes orchestration migration.
- Software Data EngineerFidelity Investments
Mission-critical, data-heavy ops: automated Neo4j knowledge-graph deployments powering 100+ data APIs, built GPT-powered test pipelines hitting 80%+ coverage across 45+ APIs, and benchmarked AWS infrastructure that saved the team $100K.
- CEO & FounderFINEAS.AI
Built an AI financial literacy co-pilot with 1,200+ users and institutional partnerships with multiple student-run VCs.
- Enterprise Asset Management DeveloperSymetri
Built a Python/PySpark pipeline that organized hundreds of thousands of unstructured construction site files from Azure Data Lake into report-ready directories, reducing a 1-month manual cleanup into a 2-week build and overnight automated run.
Have messy asset or project data? Let's map the opportunity.
If your team is spending hours searching, cleaning, reconciling, or reporting from scattered documents, Yeresi can help evaluate whether a data readiness pilot makes sense.