A virtual fireside chat about AI and Quality

Data Quality is often criticized or seen as an obstacle to shifting to Artificial Intelligence (AI) solutions. “Garbage in/garbage out” as Ashok Rajan (AR), Senior Vice President – Global Head – Cargo & Logistics Solutions at IBS Software, politely paraphrased it, is often the comment encountered in discussions or panels on the topic. Yet is that really still the case in air cargo? CargoForwarder Global (CFG) wanted to know from a man who has been on the pulse of digital developments for the past 25 years. A virtual fireside chat, courtesy of online web-conferencing, looked at AI (and human) evolution.

Humans and AI will evolve together. Image: IBS Software/Perchance.org/CFG

CFG: How clean is data these days? Is that the first hurdle to AI application, and how can change management be encouraged in companies?

AR: Everybody loves an outcome and few care about how you get to the outcome. It takes a lot to get there, however, and the biggest hurdle is data. Missing, incorrect or incomplete data hinders unleashing the full potential of AI. However, cleanliness has come a long way in the last few years helped by that fact that companies have invested in standardized systems and platforms that have enforced cleaner data.

Whilst the one push is a technical one, a change in mindset and the improved perception of value of data is also required. Clean data must be given the same importance as issuing an accurate invoice, for example. No company or customer would accept an incorrect invoice. Data quality needs to be lifted to that same level of awareness, and the topic needs to become part of stakeholder conversations – between airlines, handlers, customers, and others in the chain. It should also be part of executive reviews; how good is our data? How much value are we making out of this data? Data quality has to go up, and it is moving forward. We are no longer at the garbage in/out state. We have moved far beyond that, but still have some way to go.

The approach companies had a few years ago still, was to see how they could allow customers to tender cargo first, and only looking at data at a later stage. Also, systems 10 years ago did not validate data rigorously. Because iCargo and similar newer generation data systems now insist on a far higher standard of data quality since they have to work with it, this is pushing for better quality from the start. And with that, quality awareness in increasing. The next leap will come with mental change. My recent experience is that when Heads of Cargo see the opportunities that come with clean data, this awareness has then influenced mindset changes in the broader organization.

CFG: Where are the barriers when it comes to AI? How does IBS help to overcome / or how does it address multiple messaging standards in the industry?

AR: Quality of data continues to be biggest barrier to really taking the power of AI. Added to that, is the fact that AI-related stuff tends to happen in silos in companies. While companies are now starting to have digital strategies, what is also needed, is an overall data strategy. There is currently still a lack of a coherent strategy across the organization, and data is still something of a specialized subject. There is a shortage of qualified talent who can make sense of data, and this challenge will remain until you have tools that can abstract the complexity and quality. Not enough of the right kind of talent is currently being put onto the topic, however.

From an IBS perspective: we have been trying, in our own software, to insist on the right quality of data. Very little bad data is let in. This slows processes down to begin with, perhaps, but good data is the outcome. We also try to make it bad-data-proof. iPartner, for example, is a set of platforms that sits between systems of the ground handler – who are effectively an extension of the airline and the airlines themselves. Whereas, previously, communication would be via archaic FBL messaging, iPartner Handler enables rich connectivity. Similarly, iPartner Customer, too – offers deeper connection across new age platforms, marketplaces and ONE Record. An API First approach builds bridges with better quality of data and information exchange, instead of relying on an old messaging standard as the only means of data information.

CFG: How can and does AI support quality in processes?

AR: Quite simply: clean input drives clean processes and leads to clean outputs. AI also helps when starting to automate processes that are manual in nature; it takes the human error element out of the equation, and it improves processes by pointing people in the right direction and automating repetitive “scripts”.

CFG: How long has AI been part of IBS Software solutions – where – and what is planned for 2024?

AR: I don’t look at AI as an independent topic, – it is simply a more intelligent way of working with data, and data has been part of our DNA from the start. How we work with data has evolved into different shapes, such as Blockchain, Big Data, Analytics, etc. AI is a natural evolution of this. Currently, what we have done at IBS, is to increase the focus on AI in all our products. We now have a horizontal AI center of excellence, getting the right talent to work on our models. Our cargo line of business takes what this horizontal does and makes it relevant for cargo. Our Revenue Management system in collaboration with Korean Air, is one of the first solutions using AI. Similarly, we are putting our platforms on AI data boosters. For example, booking with an AI booster will enable upselling, offering the right price recommendation. AI should become part of day-to-day operations in everything we do, as we move from data capture to intelligent systems.

We also now use AI in our own processes – for example in development and recording. We work with Microsoft CoPilot, which enables much of the repetitive work to be pre-done, enhancing our own productivity.

CFG: AI versus human relationships – is it a competition or complement to human interaction?

AR: As you wrote in one of your articles: People make you successful.

We sometimes underestimate the human capacity to evolve. 40 years ago, people went on strike because they feared job losses when computers began. And Hollywood often exaggerates the AI vs Human situation. Yet, what will happen is collaborative intelligence where AI and humans work and evolve together. We will learn to make better use of the tools produced. Just like in a Formula 1 race: the best combination is the right driver with the best machine, not an average driver with a great machine or a great driver with an average machine. It’s when both come together, that it results in a win – and it will always be like this.

Therefore, this myth that AI will become more intelligent and take over is just a Hollywood idea. No, we [AI and humans] will both get more intelligent, and we will find better value, too, as manual processes are automated.

CFG: What are the fears about AI, in your experience?

AR: The fears are not the ones portrayed in the movies centered around AI. [Laughs]

Rather they are concerned with the biases in decision making. For example, if companies potentially have a bad way of selling cargo, then AI learns it and does same.

And a fear of implicitly trusting the black box. You send something into the system and something else goes out, which you then don’t question. I fear the adaption to ‘That’s what the machine told me.’ Collaborative intelligence comes into play here; don’t say you’re not responsible for what comes out!

We need to evolve with AI and understand it. It should not be allowed to simply go off on its own – we must actively work towards improving it, at all times. So, ensure that AI is doing the right things and make it improve in the direction you want.

These are the challenges, certainly, but don’t be scared of something taking over the world, as film producers often like to portray as the outcome of AI.

CFG: And wherein lies the excitement? For example, is there such a thing as ChatGPT for air cargo?

AR: My team would love to have ChatGPT in air cargo, reducing thousands of screens into one! [Laughs.]

These are things that evolve. What would be an exciting outcome, is if (or when) the freight industry becomes a truly digital enterprise, and everyone can be connected seamlessly. One where prices are a true reflection of market, where AI can reduce ground time and lead to other, greater efficiencies. The steps towards all this, are exciting as we encounter many shiny, new tools along with way, such as Blockchain, for example: a solution looking for a problem – not the other way around. The stages moving us in this direction, are where the excitement lies.

CFG: As a participant in the recent IATA ONE Record HaQathon in Doha, what is your take on Hackathons?

AR: What hackathons do, is keep the conversation alive. We have an IBS Cargo Forum every 6 months, which features a dedicated track on ONE Record. As a group, together with airlines and handlers, we can move this initiative forward. We are active in our own community, too. ONE Record needs to become part of daily conversation – moving one step at a time towards genuine value-add and driving processes, rather than simply another Cargo XML and ‘old stuff in new bottles’, as it were. Everything in Doha was aligned to value extraction rather than pushing standards.

And that is what AI and digitalization should all be about – not about implementing a standard, but about changing the way we operate – in terms of the right business. Does this help to ensure clean data? That is the first win and the basis for allowing us to truly operate seamlessly as opposed to half-blind.

CFG: Thank you, Ashok, for your insights and your time!



Please enter your comment!
Please enter your name here

See Also