Let´s talk about BOTs

Messaging apps are taking over the app store´s rankings with incredible retention and engagement rates. So Bots/Conversational UI are getting all the attention, and for many good reasons.

We spend more time texting and talking in our smartphones than others activities:


On top of this, it’s pretty clear we’re reaching a kind of app fatigue, and we have been rethinking the idea of “there is an app for that!”. People really don’t always want an app for different tasks, each app uses storage space, and each of them has a unique learning curve for users to go through. Even if users love some of the apps we have, we don’t really need an app for every little thing.

Why do we see apps that do everything?

A clear example of this is the chinese app Wechat, a superapp that takes care of everything. You can book a restaurant, order food, pay or transfer money to a friend, and many many more things without leaving the app. As you can see, this is the reason why Facebook is investing so much money in Whatsapp and developing Facebook Messenger with bot integration. Below you can see a great video from the New York Times explaining the phenomenon of Wechat and placing some really important questions for the future of superapps.

Google isn’t staying far behind with Ok Google and Google Home. Amazon as well with Alexa, connecting UI conversationals and machine learning, AI, natural language processing and voice recognition with Internet of Things. But I will leave this for a different post.

When thinking about integrated chatbots, we know the entry barriers are lower, we don’t need users to download an app, or go through a registration process. Bots are a great opportunity to capitalize an unsaturated channel.Developing a bot takes less time, and can be more efficient when experiencing Internet issues because they´re mainly made of text. In the end, the truth is we prefer conversation. We all have that friend that asks questions instead of googling them, but they ask them anyway because they want an interaction. Also, we know that younger users don’t even use email anymore because they prefer instantaneity.

Sometimes users are more open to chat with bots because they don´t fell judge, so they even share more sensitive information than they would with a real human.

At Lateral View we are creative solution driven. When we develop apps we do it because it’s the best solution to a user problem. We are always thinking and searching for alternatives to have a wide and innovative sprectrum of ideas. We are aware that there’s a change towards conversational UI, rich media inside chats and superapps.

At Lab55 (our Innovation Laboratory) we have been developing a bot for brands to reach their clients in an easy frictionless way: a personal shopper. And although our product is still evolving and still growing we have learned some facts about building bots that we want to share:

Solving a real problem

Remind yourself not to get caught in the trends and forget the human centered approach. Solve a real problem and iterate. We need iteration and user testing more than ever! And always applying usability rules. The positive thing about Bots is that it’s is easier and faster for your product to grow and learn to keep improving it.

We humans love to project human traits onto everything.

And that´s because we are empathic beings. The bot is going to have a personality even if you don’t want it to because of the way it answers to your target audience. Semantics will never read as neutral so it’s better for your bot to represent the brand personality or the idea you have behind it.

Make it frictionless

Simplify data entry, gather information from the user´s account and remember preferences.

Avoid replication trees

If you have several categories to offer to your user, tray no to list them one after the other.

Don’t turn your bot into a labyrinth

Make sure the flow is a logical one, test it in the context of the conversation. To avoid a never ending labyrinth where users can never reach the solution. Specially if your Bot is a customer service one. Remember to add a human being at the end of it if the user can’t find the solution inside the bot.

Context is everything (always)

Remember the context and previous conversations. Help users with shortcuts if they end up where they don’t want to. For example: “I’d like to change the product I want to purchase”

And talking about context, choose your channel

As an example a user can probably chat through Slack at work and use Messenger for chatting with friends. So be aware if the channel is the right context for your solution.

Don´t spam

Use notification smartly and remember that it´s really easy for user to block you or report you as spam. Use notification to bring the user back to the bot but allow them to set them off if they want to.

Don´t forget about groups

Integrate bot with groups conversations if it suit your solution.

Natural Language

When possible, allow the user to type answers in their language instead of listing all the options, or innovate in the way they input info, for example an image search.

And this leads to my last advice and conclusion:

Not everything is conversational

We have already seen great ideas but we are clearly on the early generations of these technology. We are still in an incipient growth between the retrieval based models and generative models. And language has so many layers (irony, semantics) that we face a very difficult challenge. But maybe, the key is that we don’t actually need bots to talk just like us, but to solve specific problems of our everyday lives.

Let’s take notice and remember the WeChat example.


Writing absolutly everything it will probably take us more steps than a visual UI. What we are trully seening is a merge between integrated chatbots in messaing apps, that adapts to the users needs in their context offering an interactive UI elements combined with text soaked in the chat screen.

Let’s Talk