What data does AI dispatching need, and where do dispatcher decision records live?

What does AI dispatching actually need to learn from?

AI dispatching software needs records of real dispatcher decisions and what happened next: which driver or technician got the job, what alternatives existed, what the customer or load looked like, and whether the outcome was on time, profitable and complete. Those records sit inside trucking, logistics and field service companies, mostly in dispatch boards, transportation management systems and the email and chat threads around them.

Public data cannot supply this. A map, a rate index or a manual explains the problem, but it does not show how a veteran dispatcher reshuffles a day's routes after a driver calls out. That judgment is written down only in the company's own operating records.

How does AI dispatch work, and where do the training records come from?

Dispatch software vendors describe products that propose assignments and flag exceptions, not only apply fixed routing rules. To propose a good assignment, a model has to see many past assignments together with their results.

A single table of assignments is a start. The value rises when assignments are linked to the notes and messages that explain them, and to outcomes months later.

Which companies hold dispatcher decision records?

Companies that route people or vehicles every day for years and keep the history. Look for a few practical signals:

• Trucking and freight brokerage firms with a TMS in use for many years and dispatcher notes on loads.
• Field service businesses (HVAC, electrical, telecom, equipment service) running a scheduling platform with technician assignment history.
• Delivery, waste, towing and last-mile operators with route and exception logs.
• Third-party logistics firms that keep customer exception emails next to shipment records.

The company baseline still applies. It needs 50+ full-time employees at peak (contractors excluded), several years of documented operations, the right to license the records, and an authorized sponsor. Owner-operator networks that rely on contractors may fall below that line even when the fleet is large.

What should a partner check before making an introduction?

Use a four-question screen:

1. History: Has the company run dispatch on a system of record for several years, with archives still available?
2. Context: Are decisions explained in notes, chat or email, or only stored as final assignments?
3. Rights: Do customer contracts, driver agreements and broker-carrier terms allow the company to license operational records? Shipper data may belong to the shipper.
4. Personal data: Do the records include driver or technician personal details, home addresses or customer contact data that will need redaction agreed with the company?

If the answers are yes, yes, probably, and manageable, the company is worth a conversation. The data inventory builder helps the company list its systems without describing any record contents.

How do dispatch records compare with other decision data?

Dispatch is one of several record types that pair a decision with an outcome. Freight claims records show what went wrong after delivery. Bid/no-bid decision records show how a contractor chose work before it started. Support escalation records capture tiered handoffs. Architecture decision records do the same for software teams. Dispatching sits in the middle: a fast, repeated decision under constraints.

A company that dispatches and also runs a customer service queue may hold two of these at once, which is worth noting in the introduction.

When is dispatch data not worth an introduction?

Skip the conversation when the company runs dispatch on a customer's portal and keeps no history of its own, when records were purged after each job to save storage, or when the only copy sits with a software vendor that will not export. Also pause if a court, trustee or assignee controls the assets, or if the data has already been licensed for AI training.

What are the limits and open questions?

• Vendor claims about AI dispatch performance are mostly marketing and vary by product. This page makes no performance claim.
• A dispatch platform vendor may hold some data under its own terms. The company should confirm what it controls.
• Telematics and location traces can identify individual drivers. Whether and how they are included is a matter for the company and the agreement.
• Not every buyer wants dispatch data; demand depends on what AI developers are building, and nothing is binding until the company agrees price and terms and signs.

What should you say to the owner?

Keep it to a question about fit. Do not ask for samples, screenshots or exports, and do not mention reward amounts.

How does the introduction work?

1. You introduce the company using the referral form or your referral link, with basic fit information only.
2. SourceX qualifies the company on size, history, data breadth and rights.
3. The company completes a data inventory.
4. Price and terms are agreed, and buyers review.
5. If the deal closes, data is delivered after an executed agreement and the company's authorization, and the company is paid.
6. Your reward is paid after SourceX receives payment.

You never export, upload or describe confidential records.

How do partner rewards work?

Partners earn 25% of the eligible platform fees SourceX actually collects from the referred company's licensing deals, capped at $100,000 per referred company. The reward is paid only after the buyer pays and SourceX receives its fee; an introduction, meeting or signed agreement alone does not trigger payment, and no reward is guaranteed. The reward is a share of SourceX's fee and is never deducted from what the company receives.

Next step

If you know a logistics or field service operator with years of dispatch history, register as a partner and use the referral link. Check the who qualifies page first if you are unsure about size or rights, and see how specialty contractor roll-ups are screened when a sponsor owns several service businesses.