If Chat Is the Next Interface, Can Libraries Reestablish Their Place in the Research Workflow?
Silicon Valley observers are starting to wonder if an interface change is underway. Mobile apps displaced the dominance of the web over the last several years, threatening substantial disruption for advertising behemoths like Google and leaving libraries and publishers ill-prepared to support emerging user needs. If another interface displacement of this magnitude is gathering, there is a major opportunity for libraries to leapfrog ahead.
Today, some observers are asking if chat will be the new interface. They look to the Chinese service WeChat’s integration of shopping, doctor’s appointments, and a variety of other services and activities–not in apps or on websites but via a chat-like interface. Facebook is very clearly positioning its Messenger platform with such features; users can load on a variety of services that not only extend its basic communication purpose but even allow them to order an Uber.
Chat is not just texting and it is not about typing, even if most chat interfaces accept typed inputs. Chat is a conversation-based services interface. Amazon’s Alexa has grown to offer a sophisticated array of services even though it is implemented entirely screen-free through the use of verbal chat. This offers up an indication of how natural and powerful a human interface chat can be.
If chat is the next interface, then libraries have a running start. They have developed extensive chat reference services, which are sometimes run locally but also delivered collaboratively through shared platforms. Many libraries have substantial information about user needs and practices in this chat medium, and some are beginning to make use of this information asset strategically. These are assets that can be built upon. Notably, they are assets that many publishers and other content providers would seem to lack.
But if chat is the next interface, then libraries need to completely rethink their models for providing chat. Most importantly, chat-based reference is often seen as a teaching practice or referral service, rather than a more fully developed service interface. Chat and other forms of reference service are poorly linked to other service providers in most library organizations, so user requests can be lost within the library rather than pursued and resolved relentlessly. Rather than use the data from chat interactions to anticipate user needs and provide ever-better service to students and faculty, libraries keep chat-based reference interactions completely anonymous. Finally, chat-based reference is typically an entirely manual and therefore labor-intensive service point, so does not offer the efficiencies that other chat services are beginning to find to automate responses to simple needs or otherwise tier the levels of support provided.
New services and access preferences have resulted in research workflows that are challenging for libraries and publishers alike. Today, I am not aware of any library that has developed chat as a service interface beyond chat-based reference, let alone built a presence inside Facebook Messenger or another consumer chat service.
Even so, as a thought experiment, imagine some possibilities for the kinds of services that could be provided inside a conversation-based services interface:
- Discovery services respond to basic searches presented in the interface
- Basic information requests about hours, locations, policies, and so forth can receive automatic up to date responses
- More sophisticated queries can be routed automatically or on user request to a professional librarian
- When discovered through whatever means, digital resources can be delivered to users (authorized inside the service where appropriate) inside the same interface as more sophisticated reference queries are addressed
- Enough user information is present so that complex requests can receive follow-up after the current session is completed
- Alerts can be provided through the chat service and not just through email
I am not arguing that chat is the ideal interface for any of these types of services individually. But, as interfaces continue to evolve in unexpected ways, there may be opportunities for libraries to make choices about interfaces that would help to reestablish their place in the research workflow. Given the combination of existing chat assets and the growth of chat in the consumer space, I wonder if this particular interface might not be an especially interesting way for libraries to reassert themselves as a key starting point for their users.
Messaging wächst: Neue Chancen für den Instant-Kundenservice in Bibliotheken? | ZBW MediaTalk
I also commented on the spate of recent posts on chat and virtual assistants and connected it both to the mixed initiative literature and some things burrowed in my PIM research findings from years ago: http://christinaslisrant.scientopia.org/2016/02/13/bots-mixed-initiative-and-virtual-personal-assistants-4/
I didn't go quite where you did with chat reference, but why not? At another presentation (http://www.slideshare.net/cpikas/pikas-xtreme-reference-2010-final) I talked about how libraries can be better integrated into their users workflow...
I think there really is something here, but what's needed is a path from here to there. I think this: "Basic information requests about hours, locations, policies, and so forth can receive automatic up to date responses" does actually happen. Many systems query a knowledge base before submitting the request to a person.
Thanks for these additional references Christina and for pointing out that there is a little more automation in place through some services. The broader issue I was trying to raise is about a more fundamental interface transformation, which requires more than just an incremental strengthening of these existing tools but rather a rethinking of various other aspects of library operation. So, I agree that the question of how (and whether) to get from here to there is completely unclear. How would we figure out whether and how to move forward? Is this a discussion that should happen inside a single library or at some broader level?