Departments · Procurement AI in Procurement: Compare Quotes, Check Order Confirmations, Track Deadlines

Tasks in procurement: six workflows at a glance

These are typical workflows, not client projects. In each one the AI prepares and flags, and a buyer checks and decides. The individual uses and their limits are covered in the article compare quotes and check confirmations with AI.

1. Turn purchase requests into orders

Requests arrive as a quick word, a note, an email or a voice message from the shop floor, the warehouse and the departments.

• Input: the request in any form, with or without an item number.
• AI prepares: recognises item, quantity and requested date, matches the item to your item records and the preferred supplier and asks for what is missing.
• Purchasing checks: the match, and approves the order as far as your amount limit requires.
• Output: a draft order in the ERP.

2. Collect and compare supplier quotes

Quotes are rarely directly comparable: different units, tiered prices, delivery terms, payment terms and validity.

• Input: the quotes as PDFs or emails and your enquiry with the items you asked for.
• AI prepares: reads the lines, converts units and tiers to a common basis and puts price, quantity, delivery terms, lead time and payment terms side by side. What a supplier did not quote is flagged.
• Purchasing checks: the comparison, and decides who to negotiate with. The AI does not propose an award you cannot follow.
• Output: a comparison table with the source of every figure.

3. Check order confirmations against the purchase order

Suppliers like to confirm something slightly different from what was ordered: a later date, another price, a part delivery.

• Input: the order confirmation and the purchase order from the ERP.
• AI prepares: compares item, quantity, price, unit and date, highlights every difference and points to references to the supplier's own terms that could conflict with your purchasing conditions.
• Purchasing checks: every difference and settles it with the supplier or accepts it.
• Output: "confirmed" in the ERP or a query to the supplier as a draft.

4. Track delivery dates

Whoever is waiting for goods chases them, but usually only when time is short.

• Input: the confirmed dates, dispatch notices and shipment status.
• AI prepares: spots orders where the date is getting close or slipping and drafts the query to the supplier.
• Purchasing checks: the draft and decides whether it needs a query, an alternative or a note to production.
• Output: a date list with traffic lights and queries ready to send.

5. Supplier scorecards from existing data

How reliable a supplier is already shows in your data: delivery dates, complaints, price changes.

• Input: orders, goods receipts and complaints from the ERP.
• AI prepares: code calculates figures such as on-time delivery, complaint rate and price trend from the data. The AI writes a short description for each supplier and names the figures it relies on.
• Purchasing checks: the assessment, before it goes into talks or decisions.
• Output: a supplier overview for annual reviews.

6. Keep contracts, certificates and deadlines in view

Framework agreements expire, notice periods pass, certificates and insurance proofs lapse.

• Input: contracts, certificates and proofs from your suppliers.
• AI prepares: reads term, notice period, price lock and expiry date, enters them in a deadline list and reminds in good time.
• Purchasing checks: the contract text itself before any decision. The AI shows the passage.
• Output: a deadline list with reminders and the passages.

Purchase to pay: where the AI helps in the process

From need to payment, purchasing sits only in the middle. The table shows who works in which step and where the AI prepares.

Programs and data routes: connected to what you already use

In procurement the data sits in several places. The table lists the usual cases. Which route your system offers, we clarify before the quote.

The names are for orientation and do not mean a partnership. Where a system offers no interface, we work with export and import, a watched folder or email.

Connect, do not replace

You do not need new purchasing software or a new supplier portal, and we do not sell either. The AI works next to your ERP and reads what sits there and in the mailboxes.

• Your ERP stays in charge. Orders and master data change only there and only by your people.
• Your suppliers do not have to change anything. They send quotes and confirmations as before by email or portal.
• One workflow at a time. We start with one step, for example order confirmations, and run it alongside the existing work first.
• Supplier management without new software. Scorecards, deadlines and dates come from the data you already have, in an overview your purchasing team reads.

Approvals and rules

Procurement is about money and contracts. So before the start we agree what the AI may do.

• Order approval: by amount and cost centre, as you handle it today. The AI creates drafts, it never orders by itself.
• Supplier records: changes to bank details or address always with a second person and a call-back through a known contact route.
• Framework agreements and call-offs: The AI checks call-offs against quantity, term and price of the agreement and reports when a limit is reached.
• Tolerances: you decide which differences in quantity, price and date the AI only reports and which it puts forward for clarification.
• Terms and conditions: references to other terms in confirmations are reported by the AI. Purchasing decides whether to accept or object.
• Log: every suggestion is recorded with source, time and decision.

For public contracting authorities, procurement law applies as well: in Germany the GWB and the VgV, in Austria the Federal Procurement Act 2018, in Switzerland the procurement law of the Confederation and the cantons. What role AI may have in a tender procedure you agree with your procurement office. The GDPR also applies to personal data of contacts: a data processing agreement (art. 28) and processing in the EU. This overview is general information, not legal advice.

Data: suppliers, items and prices

The quality of the comparisons depends on the data the AI works with.

• Supplier records and item records: Duplicate suppliers, missing units and outdated item numbers distort every comparison. So we look at the master data first.
• Price history: Earlier prices and tiers are the yardstick for quotes and confirmations. Where they exist only in Excel, we read them from there.
• Purchasing conditions and framework agreements: as a knowledge source with the passage, so that every statement of the AI points to a place in the contract.
• Data protection: contacts at suppliers are personal data. We process them on servers in the EU, show the AI only what it needs, and agree retention and deletion with you. More under AI and data protection.

Your own programs for procurement

Some things are missing not in the ERP but in front of it or next to it. We build them as separate programs when standard products do not fit:

• Purchase request app for shop floor and warehouse: scan an item or say what is missing, and the request reaches purchasing in a structured form.
• Approval app: orders and invoices for approval with the differences at a glance, confirmed on the phone.
• Supplier portal for your smaller suppliers: confirmations and dates in a simple form instead of emails, without them having to use a large system.

These programs complement your ERP, they do not replace it. More under app development.

Where to start

These three questions help with the assessment:

1. How many order confirmations and quotes does your purchasing check a week, and how much of that by hand?
2. Which ERP or merchandise management do you use, and do price lists and contracts sit there or in files?
3. Which step costs the most time: comparing quotes, checking confirmations or tracking dates?

If you send us the answers, we tell you which workflow is worth doing first and what has to stay with your purchasing team. We begin with a trial run alongside the existing work.