Back to Blog
Published:
Last Updated:
Recently Updated
Engineering Quote AutomationChapter 6

Can an AI agent chase a supplier for a delivery date?

6 min read
1,216 words
high priority
Ali Ahmed

Ali Ahmed

AI Solutions Engineer, Cognilium AI

The printed grid of a paper calendar

TL;DR

Microsoft's agent never contacts a supplier. It computes a promised date from capable-to-promise setup, so it is only as good as the lead-time fields behind it.

Microsoft's agent does not contact suppliers at all. It answers the delivery-date question a different way: by calculating one from configuration already in Business Central. That is a much narrower thing than chasing, and the difference decides whether the date on your quote is a promise or an arithmetic result.

For a distributor who does not hold stock, this is the whole quote.

Can the agent email a supplier?

No — and the boundary is drawn in two places rather than one.

Its document scope excludes the purchase side. The FAQ states the agent "can't create or work with other sales documents (such as blank orders, invoices, or credit memos) or documents in other areas of the product (such as purchase or service orders)."

And its mailbox is inbound from customers. The agent monitors "a designated mailbox for incoming customer emails about item inquiries", and its dispatcher "triggers the agent to perform tasks and then sends results, such as prepared sales quotes or orders, to the customer in response."

A supplier is neither a document type it holds nor a party in that loop. Searched to establish this: the Sales Order Agent overview, the FAQ, and the item-availability page — the three pages that between them define the agent's scope, mailbox and availability behaviour. None describes outbound supplier contact of any kind.

So the honest framing is not that supplier follow-up fails. It is that supplier follow-up is not in the design, and a delivery date gets produced without it.

Then where does the delivery date come from?

From capable-to-promise, and Microsoft is specific about what that means:

"The Sales Order Agent includes an option to use Business Central's capable-to-promise (CTP) functionality that lets the agent calculate the earliest possible delivery date for items not currently in stock."

Note the scope: items not currently in stock — precisely the non-stocking distributor's entire catalogue. Microsoft continues:

"CTP evaluates production capacity, procurement timelines, and supply chain constraints to determine when an item can realistically be delivered. This capability applies even when items aren't currently in stock and must be produced or procured based on lead times, not just inventory."

The load-bearing phrase is "based on lead times". The date is derived, not confirmed. Nobody at the supplier was asked.

What does the calculation actually read?

Four inputs, all of them configuration. Microsoft names them, and says the logic "is based in part on the configuration defined in the Order Promising Setup page":

ParameterWhat it contributes
Work Date"the current work date as the starting point for all delivery calculations"
Order Promising Setup Offset (Time)"how far into the future Business Central should look before issuing a new purchase, production, or transfer order"
Inventory Setup Safety Lead Time"accounts for internal delays like picking, packing, and labeling"
Item Lead Time Calculation"the time required to procure, produce, or transfer the item when it isn't in stock"

Three of the four are static settings, and the fourth is a field on the item card. None of them observes a supplier. The calculation is honest arithmetic over whatever those fields say.

Microsoft's own worked example makes the mechanism clear: a customer requests ten units, six are in stock or on scheduled orders, and "CTP determines when the remaining four units can be produced or sourced."

What happens when the lead-time field is stale?

The quote is confidently wrong, and nothing in the run signals it.

This is our reading rather than Microsoft's, and it follows from the section above: if the promise is computed from Item Lead Time Calculation, then the accuracy of every promised date is the accuracy of that field.

A supplier who has quietly moved from two weeks to five does not change the field. The calculation keeps returning the old answer, in a well-formatted email, with a date the customer will plan around.

For a non-stocking distributor this is the highest-consequence field in the system, and it is usually the least maintained — it is master data nobody revisits, doing work everybody assumes came from somewhere more current.

The same shape as a quantity that arrives wrong: the document is structurally perfect, so review slides over it.

Does the customer get told the date?

Yes, directly: "The agent creates and sends an email response to the customer that includes the promised shipment dates for items."

Which raises the stake on the previous section. The derived date does not sit in an internal field awaiting confirmation — it goes out, in writing, to the person deciding whether to order. As with every outbound message, a reviewer approves it first; what the reviewer sees is a plausible date with no indication of how it was reached.

What would actually chasing a supplier require?

A second surface, and it is worth being precise about the shape because it is a common request.

Chasing means outbound contact to a party that is not the customer, on documents the quoting agent does not hold, with replies that arrive unstructured and have to be read back onto a line. Three things follow:

  • A different mailbox and a different identity. The quoting agent's identity resolution is built around matching an inbound sender to a customer contact, which is the wrong instrument for supplier correspondence.
  • A different permission surface. Purchase documents are outside what the quoting agent may hold at all, so this is a separate account with its own permission sets rather than a wider grant on the same one.
  • A written-back date with its provenance attached. A date a supplier confirmed and a date a formula produced should never look identical on a quote, and on Business Central's own promising path they would.

That third point is the one to design for even if you never build the chasing. Marking how a date was reached costs a field and turns an invisible assumption into something a reviewer can check.

If you quote items you do not hold, the delivery date is the promise that costs you. Book a 15-minute call — bring a recent quote and the lead-time fields behind it, and we will show you where the date actually came from.

Sources

This article is part of our work on the RFQ quoting engine.

Share this article

The work behind this series

How an agent reads an RFQ, matches each line to your catalogue and prices it, and what has to happen to the lines it cannot resolve.

Ali Ahmed

Ali Ahmed

AI Solutions Engineer, Cognilium AI

Ali Ahmed is an AI Solutions Engineer at Cognilium AI.

Applied AI AgentsAgentic SystemsRetrieval-Augmented Generation (RAG)LLM Product Engineering
In short

Key takeaways

  • Microsoft's Sales Order Agent does not contact suppliers. Purchase documents are outside its scope and its mailbox is inbound from customers.
  • Delivery dates are calculated, not confirmed — capable-to-promise derives them from lead-time and promising configuration.
  • The promised date is only as accurate as the item's lead-time field, which is master data that ages quietly.
  • The derived date is emailed to the customer, so a stale field becomes a written commitment.
  • Chasing a supplier is a separate build — different party, different documents, different permissions — not a wider setting on the quoting agent.
What goes wrong

Common mistakes to avoid

  • Reading a promised date as a supplier commitment. Nobody asked the supplier; a formula answered.
  • Leaving lead-time fields to age. For a non-stocking distributor that field carries every delivery promise the system makes.
  • Showing calculated and confirmed dates identically. A reviewer cannot approve what they cannot distinguish.
  • Expecting supplier follow-up to be a configuration option. It is a different surface, and the quoting agent holds none of it.

Frequently Asked Questions

Find answers to common questions about the topics covered in this article.

No — and the boundary is drawn in two places rather than one.
From capable-to-promise, and Microsoft is specific about what that means:
Four inputs, all of them configuration. Microsoft names them, and says the logic "is based in part on the configuration defined in the Order Promising Setup page":
The quote is confidently wrong, and nothing in the run signals it.
Yes, directly: "The agent creates and sends an email response to the customer that includes the promised shipment dates for items."
A second surface, and it is worth being precise about the shape because it is a common request.

Still have questions?

Get in touch with our team for personalized assistance.

Contact Us

Still have a question this did not answer?

The person who wrote this article answers these. Describe your setup and what you are stuck on — you will get a straight answer, including where we think the approach is wrong.

Related Articles

Continue exploring related topics and insights from our content library.

Sales Order Agent or your own quoting agent — which request goes where?
6 min
1
Ali Ahmed
September 9, 2026

Sales Order Agent or your own quoting agent — which request goes where?

Microsoft's agent handles the shape it was built for and its limits are published. The dividing line is your RFQs: line count, line types, where answers live.

words
Read Article
When is AI quote automation the wrong build?
5 min
2
Ali Ahmed
September 9, 2026

When is AI quote automation the wrong build?

Four cases where the honest answer is to build nothing: the shipped agent covers it, the catalogue cannot support matching, or quoting is not the bottleneck.

words
Read Article
How does an AI agent turn an RFQ into a quote?
8 min
3
Ali Ahmed
September 9, 2026

How does an AI agent turn an RFQ into a quote?

An agent reads the request, identifies the customer, matches each line and prices it. What it does with the lines it cannot resolve is the real engineering.

words
Read Article

Explore More Insights

Discover more expert articles on AI, engineering, and technology trends.

Personalized for You

Curated content recommendations based on your current reading and interests.

Plain white envelopes laid out overhead
For You
115%

Can an AI agent price a quote line without knowing the customer?

Same topic cluster: Engineering Quote Automation
Same author: Ali Ahmed

It can produce the document but not the price. The agent resolves identity from the email address alone and is designed not to discuss discounts at all.

Ali Ahmed
7 min read
Engineering Quote Automation
Read Article
Trending
39%

What a Dynamics 365 F&O database refresh does not copy, and what breaks after it

High priority content
Similar reading time

A refresh copies production into a sandbox, then disables every user, withholds every batch job, clears every secret and drops every endpoint. The checklist.

Mudassir Marwat
5 min read
D365 Platform Decisions
Read Article
Stacked sheets of paper seen edge on
For You
115%

How does an AI agent turn an RFQ into a quote?

Same topic cluster: Engineering Quote Automation
Same author: Ali Ahmed

An agent reads the request, identifies the customer, matches each line and prices it. What it does with the lines it cannot resolve is the real engineering.

Ali Ahmed
8 min read
Engineering Quote Automation
Read Article
Wooden dowels arranged in even rows
For You
115%

How does RFQ line-item matching actually work?

Same topic cluster: Engineering Quote Automation
Same author: Ali Ahmed

Three layers in order: exact identifier, then attributes, then global fuzzy search. Microsoft documents all three, and the third runs on the hardest lines.

Ali Ahmed
7 min read
Engineering Quote Automation
Read Article
A path splitting in two across fresh snow
For You
115%

Sales Order Agent or your own quoting agent — which request goes where?

Same topic cluster: Engineering Quote Automation
Same author: Ali Ahmed

Microsoft's agent handles the shape it was built for and its limits are published. The dividing line is your RFQs: line count, line types, where answers live.

Ali Ahmed
6 min read
Engineering Quote Automation
Read Article
Overlapping panels of frosted glass
For You
115%

What can a quoting agent read and write in Business Central?

Same topic cluster: Engineering Quote Automation
Same author: Ali Ahmed

It reads and writes as a named user, not through an API. It holds permission sets, works only on quotes and orders, creates no master data, and posts nothing.

Ali Ahmed
7 min read
Engineering Quote Automation
Read Article

Smart Recommendations

Our AI analyzes topic clusters, author expertise, and content relationships to surface the most relevant articles for you.

Topic Matching
Author Expertise
Content Quality
Personal Interests

Questions on this?

Talk to an engineer

Schedule a call