“The reaction to ELIZA showed me the enormously exaggerated attributions an audience is capable of making to a technology it does not understand” — MIT professor emiritus Joseph Weizenbaum on the first chatbot created in 1964
Workplace chatbots promise to lighten employees' repetitive tasks and tackle information overload, as Andrew Pope and Chris McGrath have recently argued. I have advocated that the digital workplace should more closely resemble a workplace concierge and would be delighted if chatbots could deliver this vision. We also are seeing signs of a high level of interest in the market, as many intranet in-a-box vendors reported plans to add bot frameworks to their offerings.
However, while the technology is available and improving at a rapid rate, I'm concerned with whether organizations can actually implement them in practice. With the current hype levels, I anticipate a backlash in 2 to 3 years where we will hear the complaint, “our company created a chatbot but it never has the answers and nobody maintains it.”
HootSuite CMO Penny Wilson recently wrote about the growing disillusionment with consumer chatbots, so is the digital workplace leaping in just as the rest of the world is having second thoughts?
Let's not get seduced once more into thinking technology will solve our problems.
Why All the Chatbot Hoopla?
Chatbots are nothing new, so why all the hoopla now? While their ability to parse natural language queries has improved, there's little to no AI going on in generating an answer once the bot interprets the query.
What has changed in the consumer world is the dominance of smartphone use. This makes a simpler interface approach (via chatbots) much more appealing. We shouldn’t over-interpret consumer patterns though. Most knowledge work still uses large screens, but floor and field worker scenarios and office workers away from their desks can benefit from the same simplified interface.
While some may hail chatbots as the next big thing, they still need good old information management skills to implement them. Chatbot frameworks make the task code-free, but just like search, they need good content and ongoing maintenance to work well.
Repeating Past Mistakes with Chatbots
One of chatbot's primary appeals is its promise as a universal interface. Rather than saying “here is a menu of things I can do,” the chat dialogue allows unlimited requests.
Again, we return to the smartphone. Cascading menus work poorly with touch, and complex forms are harder to fill in on smartphones. Even on the desktop, integrating with people-to-people chat environments such as Slack, Workplace by Facebook and Microsoft Teams is convenient because employees can accomplish simple tasks like booking a meeting room without having to jump out of the conversation that may have triggered it.
Chatbots suffer from a paradox though: the more universal they get, the harder it is for them to perform well because they must disambiguate more options. I fear companies will try to simplify by releasing an HR chatbot, and then a Facilities Management one, followed by an IT one. Before you know it, we’ll be back to the bad old days of intranet information architecture, where everything was structured by department.
Our Fragmented Digital Workplaces
Chatbots face a second challenge in the fragmentation most of our digital workplaces currently suffer from. It’s not the chatbot’s fault. Most workplace applications for HR, expenses and travel have terrible user experiences of their own, let alone playing nicely with other interfaces. But just as intranet managers get blamed for the painfully bad expense systems they linked to, the chatbot interface will take the grief. Some vendors are making progress here, such as Beezy with Nintex, but there's a lot of room to grow.
In the consumer world we're not there yet either. Alexa requires you to know the name of a skill before you can hook into other services. I have to say “Alexa ask Hive to turn the thermostat to 21 degrees." Why doesn't Alexa know Hive is the right service for temperature issues?
Get Your Search Right First
For many queries, your workplace bot is just going to act as an alternative search interface. So ask yourself: would the same query in our enterprise search box give reliable results? If not, start there.
Bot frameworks such as Microsoft’s do well at spelling correction and parsing but require a substantial amount of training in the specifics of your company. Just as Siri struggles when you ask it to call a friend with an unusual name, chatbots will need significant manual intervention for customer names or product details unless correctly spelled (and mobile interfaces are more typo-prone).
Search results must work within the limitations of the chatbot interface. It’s no good if your mobile messaging bot says “I found this” and links to a hulking PowerPoint file. What’s needed are more like card-based search results, and that may need a rethink of your content.
More positively, chatbot search queries tend to be richer, so give us more context to improve results. And chatbots can make people search more: If someone asks the team a question on Slack, the chatbot can intervene and supply an answer using search.
Affordance and Chatbots
Right now, chatbots are like MS-DOS on steroids. Your command better be accurate or you won’t get an answer, and the interface won't give you any clues about what you can ask. Enter the concept of affordance, popularized by the graphical user interface, which suggests laying all options out clearly. Chatbots work fine when you know what you want (“claim hotel expenses”) but not so much when you forget what options you have and are stuck repeatedly asking “what employee benefits are there?”
Vacation Tracker in Slack is an example of one that gets it right. It’s more of an app integrated into the Slack interface than a natural language interface. Just type ‘vacation [command],’ and a more traditional set of dialog boxes appears to make your selection.
Without clear affordance, people tend to over-generalize AI competence. In an excellent piece on mistaken AI predictions, robotics entrepreneur Rodney Brookes points out that, “We all use cues about how people perform some particular task to estimate how well they might perform some different task.”
So when a chatbot does one thing well we tend to assume it can do a wide range of things equally well, but that isn’t the case. And this will inevitably lead to frustration.
We Still Want a Chatbot
Chatbots still make sense for specific niches: to support mobile-oriented tasks or where the scope of what is supported is sufficiently clear and constrained that businesses have a fighting chance of maintaining it.
Chris McGrath is realistic when he predicts it will take 5 years for most companies to have a chatbot of some sort and 10 years before they can handle any question. I’m less convinced when he says the technology of today is ready for it.
I also doubt companies are in a position to maintain the chatbots unassisted. In general, consumer chatbots don't yet work well, and like website search, they have a whole lot more resources thrown into them than their enterprise equivalents. Few companies have demonstrated they can adequately and sustainably resource enterprise search, so why would workplace chatbots be any different.
Similarly, the promise of chatbots is in gluing services together. That integration hasn’t happened yet. We risk diluting our efforts here by trying to add a chabot on top, before we get the fundamentals right.