Libraries and IM bots

I was reading through my feed reader just now and I stumbled upon this from

Just as you often have phone trees with recordings covering the basics, I wonder if there is a way to set up a centralized Jabber server that all libraries could use to do a similar service to the Major League Baseball (MLB) “IM bot”?

My idea is that libraries could have a “bot profile” that they could customize and then patrons could IM for the automated information they want, like phone numbers, branch locations, hours, events, etc. So, one library runs the server, but it handles multiple libraries each of whom is responsible for logging in and customizing the responses as the information changes (like if you shift from summer to winter hours, etc.) through their account. Then you could set up one IM address for the entire system for the automated info – or each branch could have it depending on the situation and preference of the library.

The example he’s working from is a baseball statistics bot:

Look up Carlos Beltran’s postseason stats, Barry Bonds’ regular season numbers, historical stats and every mathematical marvel that matters to a baseball fan. Let’s say you’re at the office or a sports bar and need a quick answer for a statistical debate. Just IM “MLB” and choose 9 from the main menu and let your digits find the digits.

Now, communicating with cell phones is an interesting idea and certainly one to explore, and admittedly I’ve not spent a whole lot of time considering that potential route. (I expect to do so as soon as I make better use of that technology myself.) But my gut reaction to IM bots answering reference questions, no matter how basic, is please please don’t do it.

I should preface this by saying that I have spent many many months of my life tinkering with virtual bots. I am a keyword bot specialist, in fact. I have built bots for the sole purpose of responding to other bots to underscore a particular piece of information. I have built dramatic historical recreations using only bots. I have built bots that people sometimes mistake as real people, who never repeat themselves in the course of a conversation. I love programming keywords. I love bots.

If v-ref experiments across the continent are determining that people always ask the same kinds of questions (“What are your hours?” “What’s your phone number?” etc.), I’d say the key here isn’t to develop a bot to answer those questions. Take that knowledge and display that information more prominently on the library’s website instead. Make the information easier to find, or include it on the same page as the IM service in addition to its regular place. Add a FAQ if you can’t work out better ways to answer obvious questions on the front page. I think limiting v-ref by creating databases of generic answers is a really good way to kill the service rather than promote it.

The baseball database is interesting and probably works well; but it works only on the assumption that no real person will ever respond. The user realizes they are sending queries to a database. This is a fancy way of scanning an index, and one I think has potential and is interesting, but telling a patron that their reference question is in some way generic and does not require a real person’s attention is not a good way to promote reference service in libraries. Yes, a person may ask the same question the last five patrons asked. But it’s the first time this patron has asked it, and since the one thing everyone loves about librarians is that their nice, replacing that nice person with a bot is probably not going to be popular.

In order to enable the bot on a standard v-ref service, you’d have to have the bot scan the questions as they are asked for keywords. In the baseball example, people are working from a “menu” and typing in specific, known, predicable queries on a very specific subject. How would we create a generic answer for the question “what are your hours?” Bots are keyword promoted, so what’s the keyword? “hours”? What happens when someone wants to know more about the film or the book The Hours? Or what if the patron says “I’ve been looking for this information for hours, I hope you can help me!” Is the keyword “what are your hours”? Then when do you do if the patron says “I can’t seem to find your hours listed on the website, when do you open on Saturdays?” I could go on, but you see what I mean. Patrons do not ask questions in a reliable, predictable way, so don’t try to use bots to spit out answers to ideally-worded questions. In the end, this offends more people than it helps.

There are two ways to think about technology and implementing it in a service context: either it’s a way to automate your universe and make everything faster, easier, and more slick, or it’s a way to connect flesh-and-blood human beings with flesh-and-blood human beings. The greatest value of IM is the way it connects two people; what is possible to translate over IM is exactly the qualities that people like in library workers. Joviality, friendliness, openness. In an IM conversation you can convey information quickly, but the real beauty of it is that you are talking to a real person. You can ask a couple of questions at once, the way most people are wont to do; you can get a name and feel that you have a ally in that big concrete building. Library staff can clarify a patron’s question on the spot, tell a person it’s really not a dumb question at all and make all the same reassuring noises we make in real life, and in the meantime learn a bit more about the search and the patron in order to provide a good answer. IM reference offers many of the same benefits that face-to-face reference offers, and this will become more and more obvious as library staff get more accustomed to the technology and more literate in the social sphere and culture of IM.

No one likes those phone trees. Everyone wants to talk to a real person, to get real assurance from someone who appears to know what they’re talking about. Automation in the library is a good thing, but there’s no need to go automating the reference librarians. The real blessing that web technology brings to a library is the capacity to connect library staff with their patrons more directly and more easily. Don’t hide them behind a bot.