"Bots are better without the conversation."

The title of Kik CEO Ted Livingston's recent Medium post, discussing the relative failure of early brand attempts at chatbots, says it all.

With so many marketers parroting on about customer experience, we may have temporarily lost sight of what users really value - speed and convenience. To borrow a phrase, 'don't make me think.'

Kik is a bot platform, which obviously makes Livingston best placed to make an authoritative assertion on the subject.

Here, he adds some more context to the idea of conversation as a red herring:

"I believe we’ll look back on the early emphasis on 'conversational commerce' as a mistake.

"Part of the misfire with the conversational aspect of bots has to do with the fact that natural language processing and artificial intelligence are not yet accomplished at managing human-like conversations."

So, it's not only users not wanting to think, but bots that are unable to (at least with the requisite accuracy).

Poncho the Weathercat is often cited as a typical chatbot that, whilst it often works, can make for frustrating conversations.

The WeChat template 

In his blog post, Livingston enumerates the advantages of messaging platforms and bots.

They include: 

  • Less friction - one interface to master, a single download and sign-up.
  • Discovery - more sharing of bots, through their inclusion in conversations between friends.
  • Consolidation - fewer apps.
  • Messaging as front door - services within a messenger, not vice versa. 

These are all characteristics of WeChat.

QR codes aid quick discovery. WeChat public accounts (essentially messenger bots) have proliferated rapidly (there are over 10m) and 40% of WeChat users read content from public accounts daily.

Mobile payment is also growing, as a service within Wechat.

Peer-to-peer payment is one of the most used features - WePay popularity has surpassed Alipay (46% use WePay compared to 31% Alipay) in part because of the interface and this social element.

It should be noted that WeChat public accounts, much like the later iterations of Facebook Messenger bots, include menus that further reduce friction when interacting in the messenger channel.

QR codes allow access to public accounts.

wechat qr

So, where do chatbots work well?

WalkTheChat provides a nice summary of challenges that chatbots are best suited for:

Narrow scope: Until artificial intelligence improves dramatically, responses are ultimately tied to decision trees, so a bot must occupy a niche, in order to provide accurate answers.

Wide range of inputs: If input range is narrow (e.g. ecommerce, where form fields are satisfactory), chatbots will only serve to make the experience more drawn out. Chatbots work best when the questions users could ask are manyfold.

Constantly evolving output: User inputs can be analysed in order to constantly improve output. This makes chatbots very useful in customer service situations, if given time to develop. The Alibaba chatbot allows a mixture of free input and multiple choice (often used to clarify input).

Conclusion

More sophisticated AI is creeping into chatbots, but until then, marketers should heed Ted Livingston's comments.

Experimenting with chatbots is all about failing fast and being well-positioned to take advantage when the technology takes off.

However, marketers shouldn't lose sight of what the user wants - experiences that are improved via messenger.

The novelty factor is not enough to sustain usage of a service that isn't quicker or easier than existing solutions.

So, what's the perfect use case for your business?

Ben Davis

Published 23 August, 2016 by Ben Davis @ Econsultancy

Ben Davis is Editor at Econsultancy. He lives in Manchester, England. You can contact him at ben.davis@econsultancy.com, follow at @herrhuld or connect via LinkedIn.

1229 more posts from this author

You might be interested in

Comments (5)

Pete Austin

Pete Austin, Founder and Author at Fresh Relevance

Totally agree.

Back in the 1970s, I wrote one of the earliest chatbots, to drive text adventure games. You could type things like "go north", "unlock door", "give a biscuit to the dog", "tell my dog to bite the troll" and so on. You could even string statements together .

Frankly, chatbots have got very little better over the intervening 40 years. They still rely on a limited vocabulary of what are basically keywords and still make stupid mistakes.

This is compounded by difficulties with accents. The funniest mistake I've encountered was Google mis-hearing my query in Aquae Sulis, "Find Roman Baths" as "Find Woman Breasts" (tells you all you need to know about the typical internet users it learned from).

It's much better to have a simple and reliable system, than one with flashes of genius but that makes more mistakes, so keeping your chatbot simple is essential.

almost 2 years ago

Ben Davis

Ben Davis, Editor at EconsultancyStaff

@Pete

Very funny :)

almost 2 years ago

Avatar-blank-50x50

Nick Phipps, Marketing Manager at SL5 7HY

We took a look at this a while back Ben - http://www.rawnet.com/blog/rise-of-the-bots

I agree with you that the evolution of Chatbots and the extension of smart AI is critical. I worry that like QR codes, the horse may have bolted too soon and the reputation may be damaged before it was ready to launch and there is a mass turn-off.

almost 2 years ago

Ashley Friedlein

Ashley Friedlein, Founder, Econsultancy & President, Centaur Marketing at EconsultancyStaff

I spoke about bots/conversational marketing at Google Firestarters recently (http://www.onlydeadfish.co.uk/only_dead_fish/2016/06/google-firestarters-20-artificial-intelligence-or-intelligence-augmentation.html).

I came up with a suggested framework for bot maturity:
Level 1 = functional request/response
Level 2 = scheduled tasks, assistive
Level 3 = predictive, smart

I recommended brands avoid trying to do anythign too branded (humour, attitude, feeling etc) for now and focus on service/productivity bots.

Funnily enough I think we're seeing command line style commands as what works best at the moment.

almost 2 years ago

Avatar-blank-50x50

Ashish Jain, ... at Chatbot enthusiast

Hmmm...interesting view point. I, however, was thinking of this a little bit differently. Since chatbots tomorrow will represent the initial face of machine learning to humans, we have to be responsible on how it evolves. You can read our viewpoints in our blog collection or also try building your own bot for free at www.engati.com

8 months ago

Comment
No-profile-pic
Save or Cancel
Daily_pulse_signup_wide

Enjoying this article?

Get more just like this, delivered to your inbox.

Keep up to date with the latest analysis, inspiration and learning from the Econsultancy blog with our free Digital Pulse newsletter. You will receive a hand-picked digest of the latest and greatest articles, as well as snippets of new market data, best practice guides and trends research.