Logistics and haulage company records and AI training: a UK guide
What UK hauliers, warehouses and forwarders hold, what the law makes them keep, what must stay out, and what AI data buyers want from logistics firms.
Key points
- Logistics firms hold detailed records of how freight moves: quotes, dispatch, proof of delivery, exceptions and claims.
- Much of it is personal data: drivers' hours and tachograph records, telematics, and the names and addresses of the people receiving deliveries.
- The law sets minimum periods for keeping some records, from 12 months for drivers' hours records and 2 years for working time records to 6 years for VAT records. A licence does not change them.
- Reported deals name dispatch records, shipment documents, delivery exceptions and claims among the records licensed.
403
Transport and storage company insolvencies, January to July 2026 (England and Wales), down from 462 a year earlier
Trans.INFOFew businesses record their operations in as much detail as logistics firms. Every load produces a quote, a booking, a dispatch decision, tracking events, a proof of delivery and, sometimes, an exception or a claim. AI developers training systems to plan and manage work want records like these, because they show real decisions made under time pressure.
This guide covers road hauliers (SIC 49410), removal firms (49420), warehousing (52101 to 52103), cargo handling (52241 to 52243) and other transport support activities (52290), the code that includes freight forwarding (Companies House). For the basics, start with our UK guide to selling company data to AI developers.
What logistics firms hold
- Operational records. Quotes, bookings, dispatch and routing decisions, tracking events, proof of delivery, delivery exceptions and claims. Under the RHA Conditions of Carriage 2024, proof of delivery can be "a document or electronic record acknowledging receipt of the consignment" (HFW).
- Systems. Transport and warehouse management systems, telematics, customer portals, email, accounts and HR.
- Compliance records. The law sets minimum rules for several of them:
| Record | Minimum rule | Source |
|---|---|---|
| Vehicle unit tachograph data | Download "at least every 90 calendar days" | GOV.UK |
| Driver card data | Download "at least every 28 calendar days" | GOV.UK |
| Drivers' hours records | Be able to produce them to enforcement officers for 12 months | GOV.UK |
| Working time records | Keep for "2 years after the end of the reference period in question" | GOV.UK |
| Safety inspection records | Keep "for at least 15 months for all vehicles" | DVSA |
| Records of goods declared to HMRC | Keep for 4 years | GOV.UK |
| Commercial and transport documents for VAT | Keep for "at least 6 years" | GOV.UK |
Logistics UK says operators "rely on advanced data-based systems for managing the movement of vehicle fleets", and that "Commercially sensitive information must be robustly protected" (Logistics UK).
Who owns the records
A logistics firm owns the records its staff create in the course of their work (CDPA 1988, s11(2)), and it is the maker of the databases they build (1997 Regulations, reg 14(2)). The harder questions are about the information inside them.
- Your customers' information. Shipment records show who your customers trade with, what they move and what they pay. Read your customer contracts and trading conditions for confidentiality terms before you include them. Remove or combine named rates and volumes.
- Personal data you control. A forwarder can be the controller of the personal data in its shipment records. One forwarder's addendum to the BIFA terms says its affiliate "acts as Controller in relation to Personal Data provided to it under an Agreement" (Clarksons). A controller needs a lawful basis to reuse personal data, so the practical route is to anonymise it before you license anything.
- Data you hold for others. If you run systems for a customer on its instructions, you may be its processor. A processor "should only process personal data in line with a controller's instructions" (ICO). That data is not yours to license.
What must stay out
- Drivers' data. Tachograph and drivers' hours records, telematics, dashcam footage and performance records. The ICO's own example is "A logistics company needs to monitor driving time, speed and distance", and its guidance also covers "body worn devices to track the locations of workers" (ICO).
- Recipients' data. Names, addresses, phone numbers, signatures and doorstep photographs on proofs of delivery. For removal firms, inventories of people's homes.
- Customers' commercial data. Named rates, volumes and margins, unless they are anonymised or the customer agrees.
- Customs declarations. These hold your customers' commercial details, and often personal data.
For how anonymisation works, see our UK guide and the glossary entry on the motivated intruder test.
What buyers want
- Mercor names logistics and supply chain among the sectors it takes data from (Mercor).
- Labelbox Prism names "manufacturing, logistics, and telecom" among its industries (Prism).
- Ooak Data shows an example dataset from a fleet logistics firm with 15 or more people and 4 years of history (Ooak Data).
- Pewter, a US firm that runs data-licensing deals for owners of US firms, lists a "Freight broker · 100 people" deal at $650,000 for "Rate quotes, shipment documents, dispatch records, delivery exceptions and claims", a third-party logistics firm at $500,000 and a logistics and distribution firm at $350,000. It labels the freight broker figure as reported by its licensing partner, and the other two as reported by a licensing firm. It says none of these figures are independently verified (Pewter).
In the US, a large sale of airline records is going through the courts. In August 2026, Google won the auction for Spirit Airlines' internal records with a $10m bid. The records include about 100 million emails and 500 million Microsoft Teams messages (TIME), and data on 763,391 flights. The sale leaves out passenger profiles, loyalty records, credit card information, call recordings and direct contact details (Business Traveller). The sale still needs US court approval, and the hearing is now set for 6 November 2026 (Chapter 11 Cases).
Exceptions are where the value is. A delivery that went wrong, the messages that followed and how the claim was settled show the judgement an AI system needs to handle the same problem.
If the firm is closing or insolvent
Insolvency Service figures for England and Wales, as reported by Trans.INFO, show 403 transport and storage company insolvencies from January to July 2026, down from 462 a year earlier. Road freight and removals (SIC 494) accounted for 226 of them (Trans.INFO).
Keep what the law requires for its full period, even while you close. The records are company property, so a liquidator can realise them, but the personal data in them must be dealt with first. Our guide to company records in insolvency explains how.
Before you talk to buyers
- Map your systems. Transport and warehouse management, telematics, tachograph analysis, email and accounts, with the years each one covers.
- Separate the records you must keep. License copies, never the originals.
- Remove drivers' and recipients' personal data. Or have it properly anonymised.
- Check your customer contracts. Look for confidentiality and data clauses.
- Remove or combine customer rates and volumes. Named prices are the most sensitive commercial detail.
- Check your role. Where you act as a processor for a customer, leave that data out.
Tydex values a logistics firm's records from basic facts, finds buyers whose criteria fit, and brings you written offers. Get a free valuation.
Buyers that take these records
Questions
Can a haulier license its tachograph data?
Not as it stands. Tachograph and drivers' hours records are personal data about each driver. You must also keep them for their legal minimum periods. Leave them out, or have them properly anonymised first.
Do we own our shipment records?
You own the records your staff create and the databases they build. But shipment records also hold your customers' commercial information and other people's personal data. Check your customer contracts, and remove or combine named rates and volumes.
What records do logistics buyers value most?
Records of decisions: dispatch and routing choices, delivery exceptions and how claims were settled. Pewter lists a US freight broker deal, reported by its licensing partner and not independently verified, covering rate quotes, shipment documents, dispatch records, delivery exceptions and claims.
We act as a processor for some customers. Can we license that data?
No. A processor may only use personal data on its customer's instructions. License only your own operating records.
Sources
- Condensed SIC codes (Companies House)
- RHA Conditions of Carriage 2024 update (HFW, 29 May 2024)
- Drivers' hours, goods vehicles: responsibilities of vehicle operators (GOV.UK)
- Drivers' hours, goods vehicles: working time rules (GOV.UK)
- Guide to maintaining roadworthiness (DVSA)
- Archiving your trade documents (GOV.UK)
- Response to the Transport Committee's Future of transport data inquiry (Logistics UK, 11 August 2023)
- Copyright, Designs and Patents Act 1988, section 11 (legislation.gov.uk)
- Copyright and Rights in Databases Regulations 1997, regulation 14 (legislation.gov.uk)
- GDPR addendum to BIFA terms and conditions (Clarksons)
- What are controllers and processors? (ICO)
- Data protection and monitoring workers (ICO)
- Data partnerships (Mercor)
- Prism (Labelbox)
- Ooak Data: partner
- Pewter
- Why Google bid $10 million for a failed airline's data (TIME, 25 August 2026)
- Google Wins $10 Million Auction for Spirit Airlines Data to Train AI (Business Traveller, 18 August 2026)
- Spirit Airlines case update: the data sale slips to November (Chapter 11 Cases, 14 October 2026)
- UK hauliers going bust (Trans.INFO, 19 August 2026)
- UK business: activity, size and location, 2026 (ONS)
Our research is general information, not legal, tax or financial advice.
