There is no doubt in my mind that the messaging revolution is here, and that the rise of bots will re-write the tech space. There are many reasons why bots are so exciting, not least because they are easier to build, they can create a more powerful engagement with customers and it’s a wide open market right now.
The market is being driven by consumer preferences, to consume services over messaging, and businesses need to consider their engagement strategies for the future.
It’s no surprise that the very first wave of bots were overall fairly basic and almost disappointing. There is a fairly large disconnect between the hype and reality at the moment, but this will close over time. Just like it closed with the app explosion. Early apps also had their issues and the user experience was average at best, now companies like Uber are basing their entire business model on the fantastic user experience from using an app. That said, there are a lot of common mistakes being made by companies large and small, which you should try to avoid.
Bots are like every other tool/ system/ process in your business. They can be exceptionally clever and useful and an invaluable resource that can reduce costs, admin time and help with prioritisation. But utilising them to their best effect means that you need to scope them carefully to avoid failure.
So before you rush into a bot strategy, here are our top 6 things you need to consider:
1. Why do you want a bot? What does success look like?
Really consider where a bot would fit in your business strategy? Where would it make the most sense for you to implement? During the sales process? Before that? After that? Or maybe even across the entire cycle?
What problem are you actually solving, or are you building a bot for the sake of building one (tip, don’t bother).
2. Who is your audience?
Whatever your use case, you need to carefully consider the audience demographic you are targeting. Where are you currently feeling the pain points and how would a bot alleviate those or do they risk making it worse if poorly implemented? Where are they already hanging out? Which messaging channel should you use?
3. How are you going to build it? And then how do you implement it?
So by now you have established 3 things:
- You want to build a bot
- You’ve decided what you want it to do
- Where it would be most effective in the business.
You need to take stock of the skills and technologies you have available in house vs requiring external vendors. Which existing business tools and workflows do you need to plug into and how?
How are you most likely going to get user acceptance? How will you know it has been successful? This all depends on the users you have and what you want to achieve.
4. Revolutions take time — don’t run before you can walk
Just like Google when they reinvented how we search for information, when Facebook, WhatsApp and WeChat started the messaging revolution or when Apple reinvented the smart phone, the bot revolution will take time for widespread adoption, and lots of mistakes will be made on the way.
In regards to individual bots, the classic startup theory applies, focus on one specific area and execute it well through iterations. Build out initial use case’s which can be executed well, and rapidly iterate based on customer feedback.
5. Unknown Use Cases
One of the challenges of building something niche or limited to begin with though, is what happens when a user wants to do something the bot can’t understand, there are various ways to deal with this situation:
a) Look at ways to prevent this happening by guiding and/or structuring the bot responses to be helpful
b) Cope with failure scenarios where the bot can prompt to attempt to correct/redirect the conversation appropriately
c) Don’t be afraid of using Human Escalation behind a bot, it can be a very powerful tool.
Personally, we think it’s important that the bots have built in human escalation when the bot providers want to benefit from it (which is why Converse.AI supports exactly this!) so that it can instantly route the issue to the most relevant person to help out.
By using human assistance to help out where required, Companies can launch, train and interate bot powered services much more quickly, while still delivering genuine benefits to customers. On the other hand, having a user who’s conversation with a bot has failed, and has no easy way to escalate to a human, leads to a very bad user experience indeed (and we’ve seen plenty!)
Coming back to point 1, Success isn’t just what it is for you, it’s also what would a customer see as success. This is just as important in the process.
6. Consider your brand and your bot personality.
Bots will become a key way for customers to interact with your brand, but that also creates a challenge. Staff members get trained about branding and being ‘on message’, but you will need to ensure the bot is trained in a similar way. They are just as much brand representatives as you are. BUT, that is not to say that they need to act like a human. They are bots, they can act like bots, in fact its better that’s its obvious that they are bots!
Undoubtedly, bots will change the way we as end users interact with businesses, and I think that everyone should be considering how to adopt them successfully sooner rather than later as, much like the app explosion, it will become the expected norm by consumers, but there is a long way to go. So get started now, make sensible decisions and don’t expect too much from your bot early on.
Visit our panel session on 19th October with Facebook, Twilio and Sage for more on how to build a relevant bot or even try it out for yourself at our bot building workshop!
Tony Lucas is the co-founder and CEO. He is passionate about building businesses that solve problems using technology. Converse.AI is Tony’s third start up business having sold his previous companies: XCalibre in 2009 and Flexiant in 2016.
Converse.AI — enabling customer success through conversation.