Artificial intelligence in transport, or the practical applications of conversational AI in the TSL industry

Applications of artificial intelligence in transport

AI (or, to be more precise, machine learning) is on everyone’s lips these days. ChatGPT has ushered in a real revolution in how we do things but, above all, in how we think: it has shown us that AI solutions are easily accessible and applicable in business. However, it has not solved all our problems and phrases like “AI-driven optimisation” or “automation” may sound good, but continue to be nebulous. AI tools promise to generate time savings and increase effectiveness. But how can we use AI-based software to get there?

Supply chain management

The answer is not clear, because AI can have applications in many different areas of transport, including route planning and optimisation, technical vehicle monitoring and data analytics. In this post, however, we will focus on supply chain management, and in particular, on the immense work that goes into creating orders, answering messages and presenting quotations.

All the above are tasks that can be easily handled by dedicated software, with little to no human intervention.

Communication automation software

Let’s make it clear. Most conversations in shipping revolve around the same old issues: repetitive messages, repetitive parameters. From a purely human perspective, all of this screams a waste of time, but no transport order can proceed before certain details are agreed upon in advance.

Practically every order requires clear answers to questions such as: what, where, when and for how much? If these questions remain unanswered (or the data is imprecise), the shipping agent needs to make further inquiries. This is where conversational AI can step in; as a result, the shipping agent will only need to join the quotation process when a precise order appears in their inbox or is imported directly into the TMS system.

E-mail automation

Conversational AI, on which the entire e-mail automation approach is based, focuses on natural interactions between machines (systems) and people. Its main goal is to understand natural language, generate responses and provide users with information or assistance in an intuitive and effective way.

So, can you really use AI for correspondence around orders, deadlines, deliveries, vehicles, transports or goods?

You bet.

Natural language in a transport app

To give you a better picture of how a natural language-based app or system works, let’s take an example: in the TSL industry, people frequently need to communicate in several different languages all in one day. Of course, multiple online translators have been around for years, but their translations often sound unnatural or full of mistakes, because they lack context.

The advantage of conversational AI is that it does not translate messages word for word, but adapts the text to the context at hand. For instance, AI will know that it is replying as a shipping agent and, accordingly, the Polish word firanka should be translated as tautliner rather than lace curtain. When messaging with the client, it remembers the entire history of the conversation and conducts itself in line with industry standards.

How does this software work?

Before conversational AI-based software can be applied in a specific field, it first needs to be additionally trained; this is known as fine tuning. First of all, we need to have an accurate and thorough understanding of the needs, challenges and specific issues of the transport industry. We need to study different applications, interactions with customers, typical issues and queries, as well as industry standards and regulations.

In the training process, we will need the right kind of training data, e.g. historical data from actual conversations, which can be used to train and test our AI model. The data can include conversation histories, road traffic data, supply chain management operational data, customer feedback, etc.

Fine tuning is usually an iterative process. This means that the honing process can be repeated over and over based on the results of model testing and assessment; tweaks and changes can be added to make sure that the app matches the needs of the industry or company even more closely.

Does the app recognise intentions?

By teaching AI to recognise context, we also train it to recognise the customer’s intention. This feature means that our conversational system will be able to identify client needs and problems just as a person would. Of course, conversational AI can also make mistakes or fail to resolve customer doubts or dilemmas; in such cases, the conversation is passed on to a human employee along with its entire history.

This guarantees that no customer is ever left unattended and the AI-assisted company retains its market credibility.

Historical data

An AI-based transport app can take over not just repetitive tasks but also:

• monitoring and data collection,
• data analytics,
• reporting.

Based on conversation histories, we can, for instance, conduct customer segmentation or scoring analysis and then prepare a fully personalised offer. For example, AI-based tools allow the identification of customers who show a preference for fast delivery and low-cost solutions, or those who expect an extra service, such as, e.g. real-time parcel tracking.

Using historical data, we can score customers based on their activity, loyalty, total transaction value and interactions with the company. These scores may then be used to classify them in terms of their potential value to the company. For instance, higher-scoring customers could be given preferential terms, discounts or special offers that would encourage them to work with us again and again.

So why is it that, despite all these advantages, conversational AI and other AI-based or machine-learning solutions are not yet the gold standard in the transport industry?

AI vs transport: data, figures, prospects

EY-Parthenon reports that:

“Business leaders recognise that artificial intelligence helps optimise processes and streamline repetitive tasks, which translates into greater employee efficiency. However, they do not expect these processes to bring any revenue growth just yet.

[…] This cautious approach to artificial intelligence is also reflected in business decisions. Applying AI tools to boost efficiency and improve financial results ranks fourth among the priorities of Polish CEOs for the next 12 months. This option was chosen by 27% of respondents, a significantly lower proportion than in the global survey, where 4 out of 10 CEOs declared they were aiming to implement AI solutions. This is a warning cry for Polish business to avoid losing its competitive edge.”

_{Source: EY-Parthenon CEO Outlook Pulse – business leaders accelerate transformation despite a difficult macro environment.}

AI in TSL

In contrast, a McKinsey & Company report indicates that although the potential for AI applications in the TSL industry is not as high as in banking, the opportunities are not essentially different than in other sectors, such as retail or insurance.

And these applications already tend to follow the path set out by banks.

How will conversational AI be used in the future?

According to a Trans.eu report entitled “Transport and Logistics Market in Europe”, growing labour costs and a lower availability of specialised workforce, such as drivers and shipping agents, mean that the use of chatbots and autoresponders to automate customer queries ranks as one of 11 possible AI applications in shipping or transport companies.

These predictions are in line with the growth forecasts for conversational AI. AI handled c. 3% of all contact centre interactions globally in 2023; by 2027, that proportion is expected to rise to 14%. (Source: Gartner, Forecast analysis, Contact center, Worldwide; 2023).

Tapping the potential of AI in transport. Some takeaways

As you can see, artificial intelligence, especially conversational AI, is becoming a very promising tool in transport. The daily grind of the shipping industry consists of repetitive manual communication processes, but recent advances and increasing access to technologies open up new opportunities for a more efficient use of available time and resources.

Conversational artificial intelligence allows many repetitive tasks to be automated, facilitating customer service and speeding up the quotation process. It should be noted, however, that despite its many advantages, business continues to be somewhat wary of AI, which suggests there is still a need for raising awareness of the great potential of these technologies.

Fortunately, many reports now show a slowly increasing trend for automation in transport, and the use of conversational AI is expected to increase in the future.

Interested in this topic? Find out about a new transport app that taps the potential of generative AI for business communication: Altkom AI Assist. A generative AI-driven transport app for customer service and data analytics.