Time flies like an arrow and fruit flies like a banana. We now have arrived at the final instalment for bot week.

Caught you off guard? No? I apologise. This writer loves horrible semantic based jokes. Which is more than what chatbots can comprehend.

This week, we’ve looked at the good, the bad and the ugly. We’ve seen how bots could be useful to us, but also the source of much frustration. Especially since they are still a long way away from being a true AI.

So, what is the future? Where could this go? Or will it be another one that bites the dust?

First off, addressing the fact that AI is years from being complete, we won’t be seeing any Sonny like assistants in the next 10 years. Although I would love to be proven wrong, the rate at which technology is advancing means we still have a lot of loose ends to straighten in the AI segments.

Ultimately, chatbots will be the most useful when they are able to reasonably identify and comprehend human interaction, learning from human syntax. Going one step further, we may even make progress in the voice recognition arena and we can talk to chatbots using voice commands.

Until that happens, chatbots are still fun to use, amusing at best and occasionally helpful. It certainly speeds up much of the process of doing tasks such as scheduling meetings and perhaps keeping you updated with the weather – provided we know what questions to type. Right now chatbots are still babies learning alphabets, and unless more programming is done, the scope of use is seriously limited.

That being said, it still proves to be an interesting way to look at future communication. Chatbots allow the illusion of personal communication without the burden of having an actual human interaction. Sounds very anti-social, but hear me out: there are more customers than there are staff for any given company. That’s how the logic works. To provide a full end to end service that can be custom catered is difficult. With chatbots, customers can choose their preferred services accordingly, and companies can rely on most commonly sought after services and products to programme the system to be able to perform certain sets of tasks. Should the commands fall out of the preset parameters, a human can take over. This saves time, money, and generates useful data that the machine can pick up and learn from.

Chat bots eliminates the need for applications as well. So if you happen to know someone who prefers the keypad brick from the prehistoric age, it still works – provided it’s not hosted in an application. (I will probably continue to call my mum’s phone a dinosaur phone, for amusement sake.) Technology is meant to be inclusive, and sometimes it’s probably better to take a step back so everyone can move forward.

That said, developers will have to take an alternative look at bots in general. At the time of this writing, I have been fiddling about with several bots on the messenger platform. Discounting the minor frustrations at bots being unable to understand certain commands if typed differently, there are certain ways in which information is presented in messaging apps that could be rather counterproductive.

Take a news bot for example. There was a news bot that allows you to read full news articles on the messaging app – paragraph by paragraph. While the others provide links to the full article, with no summaries. These aren’t the “fault” of bots – but rather how developers choose to present information in a clear concise way. Messaging applications and chats are meant to be short and to the point. With current chatbots, they are still long, clunky and text heavy. Businesses have to rethink how they can best use these platforms to integrate their services, and not force feed data in hopes that customers will be looking for a similar user experience while navigating a website or an application.

In terms of data storage, it isn’t that much different – but the type of data you collect could prove to be a different gold mine for analytics in general. Besides, using a messaging platform means data built for that platform can be transferred easily, instead of being siloed in different applications.

Would chatbots take over applications? I don’t think so. They might exist side by side, but at the moment, applications still win in convenience, layout, and number of taps needed to perform a single task. Although, there is the advantage that messaging apps don’t tend to take as much space on your phone compared to running 3 separate applications, when you can use 3 different chatbots (and more!) in one. That is until there are more bots than a human can handle – then we will have the same problem of “which bots or apps do I kill?”

For chatbots to move forward, it will have to appeal to people using it. And right now, it seems like it’s more testing out how the technology works and figuring out what they can do, instead of a productive tool that could change consumer behaviour for the better. Data shouldn’t be blindly collected and sit idly in a datacentre – if user feedbacks are any indication, it should be taken into account. Use these platforms, understand your users, and change your front end by analysing your back end.

By all means, keep it in Beta and work out what is the best approach for your business before releasing it as a commercial product. But please, don’t just jump on the bandwagon and hope for the best.

This is the last of a 5 part series. Parts 1, 2, 3, 4 can be found by clicking on the respective numbers.