tekom/tcworld 2018

tekom/tcworld conference, Stuttgart, 13-15 November 2018

This was my fourth consecutive tekom/tcworld conference and I must admit I hesitated before booking, as I wondered if I would discover anything new this year or whether I had heard it all before. I need not have worried, as the conference lived up to the high standards of previous years and once again provided an opportunity to learn about and discuss a wide variety of current topics and future trends.

The sheer scale of tekom/tcworld can make it an overwhelming experience and there is a definite sense of being spoilt for choice. In planning my personal conference programme, I decided to concentrate on a few broad themes: terminology, taxonomies and ontologies, machine learning and human/technology interaction (chatbots featured heavily in this year’s programme). In addition, I chose some sessions which simply looked interesting – serendipitous discovery is also an appealing feature of this conference.

Among my personal highlights was France Baril’s Chatbots: like a cold shower in the middle of the Canadian winter, which, as the title suggests, was a somewhat sceptical view of the hype surrounding chatbots. Chat usage is increasing, with some figures suggesting that 53% of consumers prefer to use online chat than phone support and in particular expect chat to work on mobile devices. Because chat is popular, there is an expectation that chatbots should work just as well – but we need to know what it is that people like about chat in order to develop better chatbots. France gave examples of poorly designed chat systems in which the user was presented with limited options or made to feel they were not in control of the exchange. It is important to choose the right channel for the right use case: if you are going to have a bot, keep it small unless you have access to vast amounts of data – it takes 1 terabyte of data for a bot to learn something from scratch! Ideal uses for a bot are upselling, entertaining the customer during wait times and answering basic FAQs. As Anikar Haseloff pointed out in another session on the importance of controlled language, a chatbot is, like any other employee, a representative of the company and needs to communicate in an appropriate way with customers: as products become increasingly complex, communication needs to become clearer. Ken DeWachter of Collibra made a similar point in his talk on big data terminology, giving an example of how a user complaint provided the motivation to improve internal processes and take ownership of terminology within the organisation.

The theme of chatbots and their role was continued in other sessions, including Alex Masycheff’s The long and winding road to smart chatbots. Most chatbots are still rule-based and are using the same technology as the original ‘Eliza’ in 1966. Rule-based chatbots have their limits but are useful in cases where the subject domain is narrow, there is little variation and the amount of content is small and unlikely to grow. It is important to be clear about what you want the chatbot to do – support, marketing or sales? Where the user context is more complex, a more sophisticated chatbot is needed: the content needs to be as granular as the user context, which requires inference. Alex showed an example of this using a knowledge graph to enable content-based navigation and expanded on this in a fascinating workshop the following day on building a chatbot that understands user needs.

The final day of the conference featured the Information Energy track, which focused on the concept of Information 4.0 and the roles and responsibilities needed to manage a new era of human/machine interactions. Issues to consider include the ethical limits of personalisation, management of data protection in a dynamic environment and what kind of information model we need for the Information 4.0 environment. The concept of digital twinning, where changes modelled in and made to the digital copy are transferred to the physical version, was discussed both in this track and in Markus Drenckhan’s session on smart content and its role in Industry 4.0 and the Internet of Things. The boundaries between technical communication and learning technologies are blurring and we now need to design learning for machines as well as humans. The role of the technical communicator is still heavily invested in the publication model and the concept of the document: what happens when there is no document? For those of you interested in these questions, the next Information Energy conference on Empowering Intelligent Technologies, will take place on 11-12 April 2019 in Amsterdam.

I’m aware that this is a very personal account of tekom/tcworld and that those who attended other sessions and workshops will have a different story to tell, but I think it is safe to say that most of us appreciated the flawless organisation of the conference, the wide range of high-quality speakers and exhibitors and of course the constant supply of coffee! Next year’s conference will be 12-14 November 2019 – so mark your calendar now.