Our Analysts' Insights

Five or 10 years ago, most enterprises produced support documents, manuals, marketing, websites, videos, and other types of materials they then pushed out to the world for consumption by their audience. Although they might hear back from people who called in to support lines, this content usually disappeared into the ether where it either served its purpose or did not. CSA Research’s data points to a new trend: Content is becoming conversational. It is moving away from a publish-and-forget model to one that engages users in two-way communication.

Since that time, the expectation of audiences has shifted dramatically. They now expect to be able to engage with organizations and see them respond publicly and in a matter of hours or days – if not minutes – rather than whenever the next product release comes out. We can attribute the start of this shift to two early factors: the emergence of wikis and the shift to social media, particularly Twitter. In the first case, the public now expects to be able to correct errors in content and contribute back to it, reversing the center-out model of corporate communications. In the latter case, Twitter has enabled customers to bypass consumption-only models to engage directly with companies, including corporate officers.

In our briefings and interviews, we kept hearing that content production is moving upstream, but also that its source is being diversified. Ten years ago, Adobe Systems was a pioneer in promoting locally authored, third-party content in some markets – for example, a tutorial on Illustrator written in Chinese for the China market – and began translating some of that back into English or other languages. This approach – and other community-based activities – has become more widespread and has helped break down the walls that separate organizations from their customers.

At the same time, web applications have developed increasingly sophisticated capabilities to maintain the state of a support or chat session across multiple devices or log-ins. When enterprise websites can remember where individuals were and what they have done over multiple days and screens, they create an opportunity to view that support content as an interactive, ongoing conversation.

For example, if a farmer in Kazakhstan has trouble with her German grain harvester and goes online to find a solution, the following sequence might occur:

1. Search. The support webpage analyzes her search terms – in Kazakh – and the service records for the harvester to identify her problem. It offers her an appropriate solution tailored to her needs, translated on demand from German. Already, at this stage, the Germany-based manufacturer has adapted its support to her needs and begun a dialogue. 
    
2. Customized troubleshooting. The webpage can diagnose her problem by posing a set of questions – such as “Can the cutterheads move freely?” The manufacturer’s site may generate these questions based on her harvester’s serial number and reference to a knowledge base of known issues for her model then render them in Kazakh using machine translation, rather than simply selecting static questions that might or might not apply to her situation. She might start this troubleshooting on a desktop computer, and then continue it with an augmented reality-equipped tablet or phone in the field that projects outlines of parts onto images of her equipment. A successful user experience will maintain continuity across these devices and adapt at each step of the way to what our Kazakh farmer does. 
    
3. Chat. If she continues to have trouble, she can initiate a chat with the support team in which she writes in Kazakh and the agents write in German, but each sees the communication in their own language. Language technology helps the enterprise absorb content from other countries, tie it to specific products and knowledge bases, better inform the technical support agents about the situation at hand, and interact with the farmer in her language with information about her harvester. Previously, the German manufacturer might have pushed generic support content, thus missing the opportunity to drill down to address the actual problem at hand. 
    
4. Phone. If chat fails to fix the problem and the farmer switches to the phone, she won’t have to start from scratch with someone new. Instead, the manufacturer’s support system picks up the thread of her previous interactions based on a case number, pulls up information about her problem, and reviews what she has already tried. Today, the manufacturer might need to have the customer call a dealer in Kazakhstan to have this conversation, but as machine interpretation becomes more robust, she might have access to a call center within the manufacturer’s contact center itself. 
    
5. Updating corporate knowledge. Based on this conversation across multiple devices, the manufacturer can update its support materials to reflect the solution for her problem – and thus benefit other customers. She might even contribute her own notes directly to the organization, which could help in future diagnosis, regardless of where in the world support is required.

What is essential to this emerging conversational content model is that the systems at each step maintain a dialogue with the customer and keep the conversation alive across multiple channels. Robust language technology ensures that the support interactions take place in a form that both the customer and the support representatives understand. This is one of the primary drivers that CSA Research saw behind the largest language industry acquisition ever in 2016, when French call center provider Teleperformance acquired over-the-phone interpreting provider LanguageLine for US$1.5 billion. 

As content increasingly shifts away from monolithic blocks of text with a long lifespan, language technology will become more and more vital as a way to mediate conversations that flow in both directions. Augmented translation and interpretation will be essential for cases that go beyond what automated solutions can provide alone. LSPs that can help their clients navigate this transition across multiple languages will be able to offer significant added value. Enterprises that succeed in engaging in conversational support will find greater customer satisfaction and an improved ability to help users regardless of which device they happen to use.

The shift to conversational support will have profound implications for support as voice interfaces and augmented and virtual reality become more common. CSA Research is currently looking for individuals who would be willing to be interviewed about their multilingual knowledge bases and how they are responding to the ongoing shifts to mobile devices and new types of support content. If you would be interested in participating in this research, please contact Rebecca Ray at rray@csa-research.com.