The Bot Tax

The developer conference season is finally behind us: WWDC, I/O, BUILD and F8. Now it’s time to exhale and look back at the concepts tossed our way to see if any of them are actually worth pursuing. Specifically, Bots—one of the most oft-discussed trends. Facebook’s Messenger has Bots. Amazon introduced a “Skills” Bot into its Echo/Tap/Echo Dot products. Microsoft created a couple dozen APIs for its Bot Framework to integrate with Cortana and Skype. Google’s chat Bots integrate with Google Assistant and Google Home. And finally, Apple introduced their Bot via SiriKit (though across only 6 domains, with presumably more to come).
Now, if you’re building consumer-facing use cases, that’s a lot of platforms to have to develop for. Each with different languages, frameworks and UI/input elements. Worse yet, marketers will insist IT build these bots for every platform. Why? “Because our competitors are doing it.”
The technology is intriguing, but keeping up with so many platforms is one heck of a tax to place on Development and Testing. Don’t forget, they still have all those other apps to maintain along with a raft of other app ideas (all identified and prioritized in our App Roadmap).
Luckily, for those of us dealing with internal-facing apps, the way forward is a little clearer. Here are some suggestions to keep in mind should you decide to get started:

1. Ask yourself, “Is this is worth doing?”: Will a chat bot make our service faster, better or cheaper? Is there an opportunity to re-engineer the business process at hand? How often do you foresee the Bot being accessed by your target users?
2. Narrow your focus: Amazon Alexa and Facebook Messenger simply aren’t viable target platforms for enterprise apps. But even if you’re focused on consumers, Amazon’s only shipped 3-5M Echos, all within the US. And Microsoft isn’t really a mobile platform per se, but many enterprises seem to be adopting Surface Pro 2-in-1s. So take a look at your device population to determine if that’s a viable platform. This predictably leaves us with iOS and Android. But be sure your user scenario fits into one of Apple’s 6 SiriKit domains. If not, you may want to stick with Android only (or not move forward at all).
3. Bots don’t need to be sentient, but they do need to be context-aware: People don’t expect your bot to be like Samantha (from the movie “Her”), but the experience needs to be context-aware and personalized as well as being fun and engaging. The options you surface must reflect an understanding of the user’s situation and/or an interpretation based on the user’s inputs.
4. Be ready for a brand new style of interface: A year ago I was bullish on voice as a primary mode of interaction. I still am. Bots don’t require voice, necessarily. But in many cases it makes sense for voice to be the primary input method. The interaction is the same. Only now we have more means of leveraging non-visual controls. Not only is the UX different, you should also think about the bot’s voice or personality–particularly for consumer-facing bots—but for employee-facing bot interactions as well. Ensure the content, tone, and phrasing matches your brand. Thinking through what this looks like is not trivial. Don’t underestimate the effort involved.
5. Don’t build a blabberbot: Knowing when to invoke your Bot can be a tricky balancing act. The idea is to simplify and streamline, not annoy and disrupt. We all know the challenges of app adoption and the propensity of users to delete or stop using apps after just one bad experience. Users may be even less forgiving when it comes to Bots.

Final Thoughts
Bot technology is intriguing. They have the potential to create more ‘natural’ interactions and better apps. But most enterprises still have a long way to go in crafting a mobile strategy—let alone executing against it—and delivering real value to the enterprise. So keep your eye on Bots, run a POC or two, but don’t pass up opportunities to build apps that matter in the meantime. Consistently deliver value and the budget for ‘fun’ stuff will still be there. Remember, bots are like 2 seconds old. We’re still way early in the cycle. There’s nothing wrong with letting others suffer the hard lessons as we focus on our more pressing priorities. Want to discuss the rise of Bots (or any other emerging technology) with a Propelics strategist? Check out our Emerging Technologies Kickstart to get started.