In short
- Start with work that repeats and follows clear rules: retyping orders, sorting documents and drafting messages to carriers. Leave decisions about shipments to people.
- Pick one process, measure how much time it takes today, and roll it out with a person checking every output.
- Find out whether you can exchange an electronic consignment note with carriers and customers. The Additional Protocol to the CMR Convention gives it the same weight as a paper note. It entered into force for the Czech Republic in 2011, and you should check the position in the countries you operate in.
- Around 2027, be clear on how your firm will work with eFTI. EU member states' authorities will have to accept electronically submitted transport data.
Where AI takes manual work out of a logistics firm
AI saves the most work where people retype, sort and reply to messages by the same rules every day. In logistics that means three places: taking in orders, handling documents and communicating with carriers. AI prepares the draft. The decision on who gets a shipment and at what price stays with a person.
In this article, artificial intelligence (AI) means software that reads text, sorts it and writes replies from your instructions. You don't need to build anything yourself. You only need to choose one process and test it.
Where firms stand: the Czech example
According to the Czech Statistical Office (CZSO), 18% of Czech enterprises used at least one of eight monitored AI technologies in 2025. The European Union average was 20%. The figures cover enterprises with ten or more employees, so firms of roughly your size.
The CZSO also reports that 13% of Czech enterprises used AI to generate text or code, against 9% on average in the EU. Generative AI is AI that creates new content from instructions or input data. In practice that means writing emails, summaries and filling in forms. The Czech figures are a useful benchmark: mid-sized firms in a mid-ranking EU country are already using these tools, so you wouldn't be starting from zero.
Orders, documents, carriers: what AI can handle
The table shows three areas, a typical task in each, and what to watch. These are illustrative examples, not results from specific firms.
| Area | What AI can do | What to watch |
|---|---|---|
| Orders | Reads a customer email or PDF, pulls out the loading and unloading addresses, goods and dates, and prepares a record in your system | A person checks the record before it is saved. A misread date or address costs more than the time saved |
| Documents | Sorts incoming documents, matches them to a shipment, flags a missing signature or field | Don't put sensitive documents (contracts, prices) into public tools without rules |
| Carrier communication | Drafts a request for quote, a reply to the question 'where is my shipment?', or a reminder for a missing document | A dispatcher approves sending until the drafts have proven reliable |
In all three cases AI prepares and a person confirms. That is a low-risk way to begin. You can then relax the checks gradually, but only for tasks where AI has been shown to make few mistakes.
How to start in a typical logistics firm
- Choose one process. Ideally one that several people do every day and that follows clear rules, such as entering orders from email.
- Measure the current state. Find out how much time it takes, how many errors occur and who does it. Without this you won't know later whether AI helped.
- Decide what data AI will see. Set what may enter the tool. Prices, drivers' personal data and contracts need rules.
- Have every output checked. A person reviews each output first. Keep a log of the errors.
- Evaluate after a few weeks. Compare against the original state and only then decide whether to extend the process.
The electronic consignment note: paper is no longer required
The CMR consignment note can already be electronic. The Additional Protocol to the CMR Convention of 2008 defines an electronic consignment note. It provides that, if the note meets the protocol's conditions, it has the same evidentiary value and the same effects as a paper note.
The Czech Republic has acceded to the protocol. According to the Czech Collection of International Treaties, the instrument of accession was deposited with the UN Secretary-General on 14 April 2011, and the protocol entered into force for the Czech Republic on 13 July 2011. Other states have joined at different times, so check whether it applies to the countries on your routes.
The protocol sets conditions:
- A reliable electronic signature. The parties authenticate the note with a signature that is linked only to the signatory, makes it possible to identify them, is under their control and detects any later change to the data.
- The same data as on paper. The electronic note contains the same particulars as a note under the Convention.
- Data integrity. Any addition or amendment must be traceable, and the original data must be preserved.
- Agreement between the parties on procedures. The parties must agree how the note is issued and delivered, how authority is proven, how delivery is confirmed and how the note may be replaced by a paper one.
A scanned paper note is therefore not enough. You need a system that secures the signature and preserves the data, plus an agreement with your partner.
eFTI: what is coming at EU level
European Union Regulation 2020/1056 on electronic freight transport information, known as eFTI, lets companies give authorities transport information electronically. It covers road, rail, air and inland waterway transport. It applies in general from 21 August 2024.
The key date is 9 July 2027. According to the European Commission, from then on Member State authorities must accept information that companies share electronically through certified eFTI platforms.
This obligation falls on authorities, not on companies. The regulation aims at simpler information exchange and less administration.
According to Trans.INFO, the regulation provides for digitising these documents:
- documents on the transport contract and the goods: CMR, CIM/SMGS, air waybill, bill of lading, customs declaration,
- vehicle documents: vehicle registration certificate, insurance,
- driver documents: driving licence, driver card, certificates and licences.
According to Trans.INFO, under eFTI an electronic consignment note must have a structured, machine-readable form (for example EDI messages, XML or an API) and precisely defined data. These requirements belong to eFTI. The CMR protocol does not require machine readability. The data are: carrier, place and date of taking over and of delivery, consignee, description of the goods, number of packages and weight, transport costs, instructions for customs formalities, and a statement that the carriage is subject to the CMR Convention.
Machine-readable data is much easier to capture from orders and documents than from paper. A firm that can already get order data into its system without retyping will be better placed for the electronic consignment note.
Checklist before you decide
- Which process in our firm is the most repetitive and follows clear rules?
- How much time and how many errors does it cost us today? Have we measured it?
- Who will check the AI output, and who is responsible for a mistake?
- What data may AI see, and who decided that?
- Can our transport or warehouse system pass data on in machine-readable form, or does it only print documents?
- With how many partners could we try an electronic consignment note and sign an agreement on procedures?
- Who in the firm is responsible for preparing for eFTI by 9 July 2027?
If you can answer most of these questions, you can start with a small pilot project. If not, first put your data and responsibilities in order. Both AI and electronic documents rest on them.
AI agent use cases in logisticsData security when you deploy AI in your company
Sources
Prepared by AI from the listed sources and checked automatically against them. Not legal advice.