Hisham Alhussain
All work

Case study

Tender360

Automated tender intelligence: unstructured bid documents become searchable data and executive-ready summaries of key points, risks and comparisons.

For
NEOM
My role
Lead architect (30%) & hands-on builder (70%)
Stack
RAG, Embeddings, Semantic search, LLM summarization

Problem

Procurement relied on manual review of PDFs and email threads, causing missed deadlines and poor visibility during busy tender cycles.

What I built

An event-triggered document pipeline with chunking, embeddings and semantic search, plus an AI summarization layer feeding reports and dashboards.

Outcome

Tender processing went from weeks to hours with ~90% less manual effort. Other teams reused it as the reference workflow for document automation.

tender processing time
Weeks → hourstender processing time
less manual effort
~90%less manual effort

Architecture

  1. 01Tender documentsUploads & central repositories
  2. 02Auto ingestionTriggered as new files land
  3. 03Semantic processingChunking, embeddings, search
  4. 04AI summariesKey points, risks, bid comparison
  5. 05DecisionsReports & dashboards for committees
  • Event-driven
  • Auditable
  • Repeatable each cycle

The problem

Tender evaluation is time-sensitive and business-critical. Before Tender360, teams reviewed bid documents by hand, worked out of PDFs and inboxes, and pieced together status from scattered files. During busy procurement periods that meant missed deadlines and little visibility for leadership.

The approach

The flow is simple to say and took care to get right: tender documents → automated ingestion → structured data → insights.

  • Ingestion is triggered automatically as documents and presentations land in uploads or central repositories.
  • Semantic processing chunks and embeds the content and indexes it for semantic search, so it's queryable at scale instead of locked in PDFs.
  • AI summarization produces executive-ready summaries that surface key points, risks and side-by-side comparisons across bids.
  • Delivery happens through reports and dashboards that committees and leadership already use.

My role

Mostly hands-on: I designed and built the ingestion pipeline, extraction and semantic indexing, and the workflow automation, while aligning with procurement teams, tender committees and leadership.

Outcome

  • Tender processing went from weeks to hours
  • Manual effort dropped by ~90%, and evaluations became more accurate and consistent
  • It fed downstream analytics and became the reference workflow other teams reused for document-driven automation

Details are kept high-level because this is internal NEOM work.

Internal NEOM work, shared at a high level without confidential data.
Next case studyWhatsApp AI Front-Desk Agent