Accelerating11 conversations · Small business

RFQ-to-Excel template kit

PDFs won't become structured data

  • Product forums
  • Freelancer.com requests
  • GitHub
  • Hacker News
62

WantRank · Strong

How much people want this, and how sure we are. Every independent voice counts for what its origin is worth — a paying customer more than a passing post — then the market speaks: funded companies on the problem, proof that people already pay.

Demand78
Money68
Market60
Momentum76
Conversations found per week, last 26 weeks
DayWeekMonth

Counts the conversations PeopleWant found, by the date they were posted. It is a lower bound, and the lighter last column is the period still in progress.

Conversations found per week
PeriodConversations
Week of May 40
Week of May 110
Week of May 180
Week of May 250
Week of Jun 10
Week of Jun 80
Week of Jun 150
Week of Jun 220
Week of Jun 290
Week of Jul 60
Week of Jul 130
Week of Jul 200
Week of Jul 270
Week of Aug 30
Week of Aug 100
Week of Aug 170
Week of Aug 241
Week of Aug 310
Week of Sep 70
Week of Sep 142
Week of Sep 213
Week of Sep 282
Week of Oct 50
Week of Oct 120
Week of Oct 190
Week of Oct 260

In short

  • Procurement managers receive supplier quotes in unstructured PDFs and emails; manual transcription into Excel is slow and error-prone
  • Existing tools fail on edge cases: layout variations, OCR errors, hallucinations on critical financial data, and inconsistent formatting across documents
  • Data validation against approved catalogues is manual; structured output with flagged anomalies could unlock automation for downstream processes

The problem

Procurement teams receive quotes and RFQs from multiple suppliers in inconsistent PDF formats—some scanned, some with mixed layouts, some with non-standard structures. They need to extract line items, normalize them, and validate against an approved catalogue before entering Excel for comparison or reporting. Today's solutions either require manual entry (slow, error-prone) or use generic AI that hallucinates numbers, skips items, or misinterprets digits—unacceptable for financial data. The edge cases—layout deviations, OCR failures, ambiguous…

Who has it: Procurement managers and teams in mid-market companies processing multiple supplier quotes.

What people say

  1. 01Product forumswillingness to pay
    “That is the same problem as yours: messy source documents in, structured normalized records out, validated against a source of truth, with the edge cases surfaced rather than swallowed.”
    Open the original conversation ↗
  2. 02Hacker Newsworkaround
    “Yes, remember that these are effectively random PDFs in various different designs and formats, some of them not editable or even OCR'd.”
    Open the original conversation ↗
  3. 03Freelancer.com requestsactive search
    “About two-thirds share one standard layout; the rest deviate slightly, so column mapping will change from time to time.”
    Open the original conversation ↗
  4. 04GitHubopen-webui/open-webuipain
    “I've tried many variations of chunk configs, top k, hybrid search, nothing results in a working conversion.”
    Open the original conversation ↗
  5. 05Product forumswillingness to pay
    “Sample (RFQ PDF to Excel with the review sheet) available on request.”
    Open the original conversation ↗

3 more verified conversations behind this opportunity

Every one links to its original source. Subscribers see all of them.

Unlock the full evidence

What this shows, and what it does not. 11 people described this problem in separate conversations, under separate accounts. That makes it real and recurring; it does not tell you how many would pay for your product, or how much. Talk to a few of them before you build.

Signals

Repetition66
Pain71
Money92
Momentum100

Computed from the evidence. “Conversations” is a lower bound: at least this many people described the problem, in separate threads, under separate accounts.

Business ideas

  1. 01Solopreneur

    RFQ-to-Excel template kit

    Pre-built Power Query or Zapier templates that map common RFQ layouts to Excel, with built-in validation rules against user-uploaded catalogues. Sold as a one-time download or template library subscription ($29–99/month).

  2. 02SaaS

    Procurement document ingestion platform

    Web app that accepts PDF uploads, extracts and normalizes line items, validates against a user-defined catalogue, and exports clean Excel or CSV. Flags ambiguous or anomalous rows for review. Subscription model ($199–499/month) based on documents processed…

  3. 03App / extension

    Quote parser browser extension

    Chrome extension that intercepts email attachments and PDFs, extracts structured data in real-time, and offers one-click export to Excel or CRM. Per-seat subscription ($9–19/month) or pay-per-parse.

  4. 04AI agent

    Autonomous RFQ processor

    AI agent that monitors an email inbox, extracts quotes, normalizes them against a catalogue, flags exceptions, and posts clean data to a shared spreadsheet or database daily. Retainer model ($500–2000/month) based on volume and integrations.

SaaS · First version: About a month

For

  • Strong money signal: teams already spend on tools, contractors, or staff time to solve this; validated data accuracy is non-negotiable for procurement
  • Repetition across many domains: RFQs, purchase orders, invoices, expense reports all face the same problem—unstructured PDFs into structured, validated records
  • Reachable buyers: procurement managers are identifiable, live in procurement software communities, and have budget allocation for process automation

Against

  • AI accuracy on financial data is still hard: even recent models hallucinate or skip rows; users will demand 99%+ accuracy, and the last 1% may require human review, limiting margin
  • Consolidation risk: major ERP and procurement platforms (SAP Ariba, Coupa, Jaggr) are adding native document parsing; if they ship good-enough versions, your TAM shrinks
  • Incumbent tooling: teams already use Zapier, Power Query, and generic document AI; switching costs are low but so is willingness to pay for a specialist tool unless pain is acute

Existing solutions funded by YC

3 companies from YC's last 10 batches sell a fix for this problem. Investors already bet on it: expect competition, and look at what they leave out.

  • Unsiloed AI ↗Fall 2025

    API for parsing multimodal unstructured data

    Artificial Intelligence · B2B · Infrastructure · APIs

  • Saldor ↗Summer 2024

    Making procurement easy

    B2B · Procurement

  • Pulse ↗Summer 2024

    Production-grade unstructured document extraction

From YC's public directory (open yc-oss dataset), matched to this problem automatically.

That was one fiche, in full.

2 more are open to everyone: Agent cost dashboard template · Inventory sync template or plugin.

Subscribers open every fiche on the radar — 662 today — with every conversation behind it, ranked every day by how much people want it.

View pricing