AI use case · Finance

AI supplier quotation comparison for UAE procurement teams.

For procurement teams that copy three or more supplier quotes into a spreadsheet before every award. AI reads each quote into the same structure, lines items up against your request and flags what is missing, so your buyer compares like with like.

Ahmed Khaled, FounderLast updated:

Discuss this use case

A short note is enough. Phone and email let us reach you.

About: AI & Intelligent Automation

Tell us about the workflow, who uses it and what is difficult today. Please leave out passwords, confidential records and sensitive personal information.

Add company and other details (optional)

After you send this, choose a time for a free 30-minute discovery call. If no time suits, we get back to you as soon as possible to arrange one.

What is AI supplier quotation comparison?

AI supplier quotation comparison reads quotes from several suppliers, whether PDF, Excel or email, into one structure. It matches each line to your request for quotation, aligns units and quantities, and lists exclusions, validity and payment terms side by side. Missing items and unclear lines are flagged. Your buyer reviews the comparison and makes the award.

The problem

Every quote arrives in a different shape.

A request for quotation goes out with one item list. The answers come back as a PDF from one supplier, an Excel sheet from another and an email body from a third, some in Arabic and some in English. Item names change, units switch between pieces, boxes and metres, and some lines are simply left out.

Before anyone can compare prices, a buyer rebuilds every quote into one spreadsheet by hand. The real differences often sit in the small print: delivery terms, validity, whether VAT is included, and exclusions written in a note at the bottom. Those details are the first to be missed when a comparison is rushed to meet a site or project date.

  • Rebuilding each quote

    Every line is retyped into a comparison sheet before the first price can be compared.

  • Unlike with unlike

    Different units, pack sizes and descriptions can make one quote look cheaper than it really is.

  • Terms in the small print

    Exclusions, validity, delivery and payment terms are read late or not compared at all.

How it works

From the request to a reviewed comparison.

An illustrative workflow. The channels, comparison rules and target system are agreed for each project.

  1. Request as the baseline

    Your request for quotation, with its item list, quantities and units, becomes the structure every quote is read into.

  2. Quote intake

    Quotes arrive by email or upload, as PDF, Excel or scanned pages, in Arabic, English or both.

  3. Line matching

    AI matches each quoted line to a requested item, keeps a link to where it appears in the quote, and marks lines it could not match.

  4. Units and terms

    Units and pack sizes are converted using rules you approve. VAT treatment, validity, delivery, payment terms and exclusions go into their own columns.

  5. Gaps and outliers

    Missing items, extra items, prices far from the others and quotes past their validity are flagged with the reason.

  6. Buyer review and award

    Your buyer reviews the comparison, resolves the flags and records the award in your ERP or procurement system.

Who does what

What the AI does · What a person checks.

What the AI does

  • Reads each quote into the structure of your request, in Arabic or English

  • Matches quoted lines to requested items and links each value to its source

  • Converts units and pack sizes with rules you approve

  • Lists VAT, validity, delivery, payment terms and exclusions side by side

  • Flags missing lines, unmatched items and outlying prices

What a person checks

  • Every line the AI could not match, or matched with low confidence

  • Alternative products a supplier offered instead of the one requested

  • Technical compliance with the specification

  • Commercial judgement and the award itself

First pilot

A first pilot on quotes you have already compared.

We test on closed requests, so the result can be checked against your team's own comparison. Scope, timeline and price are confirmed in a written proposal.

  • One category

    Pick one purchase category with frequent requests, such as building materials, spare parts or office supplies.

  • Past requests and quotes

    A set of closed requests with their quotes and your team's comparison sheets, shared safely after the first call.

  • Agreed comparison rules

    Unit conversions, how VAT is shown and which terms must be compared, written down before the test.

  • Line-level results

    We report how many lines were matched correctly and how many needed a person, against your team's comparison.

Data, hosting and Arabic

Where the data goes, and how Arabic is handled.

  • Hosting

    The application and stored data can be hosted on Azure UAE North, on private servers managed by Vloud, or in your own cloud account. The AI model is used through a provider agreed per project, and its processing region may be outside the UAE.

  • Access

    Quotes and comparisons stay with the buyers and approvers you name for each request. Suppliers never see each other's prices.

  • Arabic

    Arabic and mixed-language quotes, including item names each supplier writes differently, are tested as their own sample.

  • Audit trail

    Each comparison keeps the original quotes, the matched lines and every change a buyer made, so the award can be explained later.

Honest limits

Where it does not work yet.

  • Technical equivalence is a judgement. If a supplier offers an alternative product, a person decides whether it meets the specification.
  • Unit conversions only work with rules you approve. Where a pack size is unclear, the line is flagged, not guessed.
  • The lowest price is not a recommendation. The comparison shows the differences; the award stays with your team and your approval rules.
  • We do not quote accuracy or time savings before testing on your quotes. Results are measured on a held-out set first.
  • If you only compare a handful of quotes, a better spreadsheet template may be enough, and we will say so.

What drives cost and timeline. The number of categories and suppliers, quote formats and languages, and how many unit and term rules are needed. Cost also depends on where the request data comes from and the system the award goes to. We don't publish a price list; scope, timeline and price come in a written proposal after the discovery call.

Evidence

What we can show you.

Every example is labelled: delivered work, our own product, a demonstrator, or an illustration with synthetic data.

  • Our product

    ArkanERP procurement

    ArkanERP, the ERP we build and own for construction and project-driven companies, includes procurement and BOQ control. It is our product, not a delivered AI comparison project, and owning it is not evidence of client adoption.

  • Delivered work

    Odoo for Ashbilia Contracting

    We implemented Odoo purchase, accounting, employees and inventory for Ashbilia Contracting LLC, replacing a fully manual process. It is ERP work, not an AI quotation project.

  • Illustrative

    How to test AI document extraction

    The evaluation method we use on a pilot: development set, held-out set, field-level results and review rate reported separately.

Before you start

Questions buyers ask.

Can it compare quotes in different formats?

Yes: PDF, Excel, scanned pages and email text. Each quote is read into the structure of your request, so the original format does not decide the layout of the comparison.

Does it pick the winning supplier?

No. It shows where quotes differ and what is missing. The award, and any trade-off between price, delivery and quality, stays with your buyer and your approval rules.

How does it handle different units?

With conversion rules your team approves, such as pieces per box or metres per roll. Where a quote does not say enough to convert, the line is flagged for a person.

Can it work with our ERP?

It can read requests from, and write awards to, a system with a supported API or an agreed import route. We review your ERP before confirming the integration scope. ArkanERP, our own ERP, includes procurement.

Does it handle Arabic quotes?

Yes, Arabic, English and mixed quotes. Supplier item names vary more than the language does, so matching is tested on your own quotes during the pilot.

Where is the data hosted?

The application and stored data can be hosted on Azure UAE North, on private servers managed by Vloud, or in your own cloud account. The AI model is used through a provider agreed per project, and its processing region may be outside the UAE.

Who owns the code?

You own the source code and IP of the custom work we build for you. Our own products stay ours, and any licence terms are agreed separately.

Next step

Test it on your own documents.

Tell us what the work looks like today. We reply with the questions we need answered before a pilot can be scoped.

Discuss this use case

A short note is enough. Phone and email let us reach you.

About: AI & Intelligent Automation

Tell us about the workflow, who uses it and what is difficult today. Please leave out passwords, confidential records and sensitive personal information.

Add company and other details (optional)

After you send this, choose a time for a free 30-minute discovery call. If no time suits, we get back to you as soon as possible to arrange one.