Chatbot vs. AI Bot - which is here to stay, which to invest in and why?
To Chat with a Bot or not to Chat with a Bot, is that the question?
Back in August I delivered an exceptionally simple and laymen overview of Bots, ChatBots and Artificial Intelligence (AI).
Missed it? catch up here.
At the time I wrote the article, my inner gut feeling was that while I technically and fundamentally grasp the landscape, components, ecosystem and all the connected parts that make up the awesome world of Bots, ChatBots, Artificial Intelligence (AI) and the Internet of Things (IoT) (and aside from being my latest hobby and fixation), the rest of the lay population is nowhere near being on the same page. However, and as sad as that may be, we are all interacting with Bots daily, even if we are completely oblivious to it.
Something in that article struck the right chord, connected, hit the nail on the head (so to speak) and bridging the gap for my readers, between having no information whatsoever (about this Bot stuff) the one day, yet a clear picture of where it’s at the next.
I’m delighted to report that the article was a huge success and has resonated through an interesting crowd, bringing me a pleasurable following, many a conversation and has demonstration that I am on the right road to entrenching myself in the space…I guess I am starting to think of myself as an evangelist for the technology which I have never thought of before. I am super excited because the positive adoption of my article and the human feedback I have received has endorsed that this technology is just at the beginning and I plan to bring my followers along for the ride, from ground level up and right through this next digital revolution.
So here’s the next instalment.
When I changed my email address earlier this year, I logged into my mobile phone providers web portal and updated the email address and auto email notification settings. The old email address was removed from the system (at least that is what I thought). Six months later I have never received any email notifications, no bills (lucky those are on a direct debit), no correspondence, nothing. Then a week ago, my wife exceeded her 12GB data limit (don’t ask me how) and received an SMS alert that we are to be charged additional fees for that. I then realised we had not received any notifications or warnings by email so I logged back into the portal but, alas, the old email was nowhere to be seen. This prompted me to connect with my mobile service provider.
Who wants to lookup a contact #, then dial, then push 1, push 3, push 5 then explain …. so BANG!, I fire up WebChat (my channel of choice) because this was such a simple thing to fix, right?…maybe not.
I was immediately greeted (friendly) by “Katorina” and asked how I could be served…I speedily typed one sentence explaining the problem, they asked for my credentials to securely identify me which I provided.
Whilst waiting, staring at “dot dot dot”(as “Karotina” typed), I explained I was in a big rush and needed this resolved urgently as it was such a simple fix..then a PAUSE and the reply by Katorina: “Cool….”. I wait, and wait, but nothing….another response comes on screen saying: “Thanks for your patients while I investigate your request”. I reply again that I am in a rush and not quite sure how much there is to investigate, they reply: “Cool…”. I type: “This is getting frustrating how long should I expect this may take?” and again “Cool…”. I’m now feeling that botty feeling coming in (which secretly makes me quite excited because immediately I see the opportunity of engaging with this Bot but don’t want the Bot to know that of course) so I type: “Why do you keep saying Cool…, I’d like a response please, have you resolved the issue” at which point a generic response appears: “Please accept my apologies as I will need to transfer you to another customer service agent…” and with that “Martin” (and now clearly a human) responds: “G’day, how may I help you…” well that only lead me to unleash genuine frustration. I now have to re-explain the whole story and of course, to add fuel to the fire, I now throw in that I know I was interacting with a Bot yet this technician was oblivious to the fact that I was on the chat for 10 minutes already. Although he focused on immediately resolving the issues, it was as if he was unaware of the previous bot engagement I had been through and as if he was the original responder to my initial chat request.
There is so much to take away from this experience (about Chat Bots) which I am bursting to share.
Firstly, it has reinforced the need for me to clarify the specific differences in consumer facing bots that already exist and after only a short space of time since exposure to adoption:
Chatbots, generally rule based botsI believe these will surface (and are in fact already surfacing) as a the most common bots in consumer facing businesses, in the short-term.
These Bots are generally easiest to build with less infrastructure than their more complicated sister/brother - AI Bots. Think of these as built almost for a single purpose for example to Chat: and to provide very linear and single dimensional support for example: customer service. They have a specific (usually cheaper) technology and architecture requirement and are quicker and easier to build, deploy and implement but therefore cheaper, less scalable and less robust with lots of potential down side for the long term for example with reliability, analytics, learning, less of an investment, more of a cost and liability (as I see it). Platforms are starting to emerge delivering precise PaaS solutions specifically around ChatBots for Example; Kore and Pullstring amongst many others.
AI BotsAI Bots are similar to Chatbots but one fundamental difference (which seems quite a simple add-on, on the surface) drastically affecting the tech requirements demanded to support a Chatbot through to more advanced levels, as well as ongoing support, is the artificial intelligence component, AKA the NLP (Natural Language Processing capability ), the brain, the engine, the learning centre.
A good example I like to use to best demonstrate the fundamental difference between a Chatbot and AI Bot is the development of intelligence (and literacy) in children and, adding the advanced layer of how children learn values and morals and decision making beyond the literacy aspects. I suppose these are also the characteristics that define a human vs a machine… IE, once they are literate, how do they learn to decide right from wrong.
Chatbots are like putting an ipad in a 2–3yr old childs hands. It’s an intuitive experience and without human intervention whatsoever and the support of Siri and Apps alone, the child could actually learn to say words and after a while speak sentences and engage with other humans.
As an aside, maybe this concept of educating children will be the first generation of education overhauls we will experience which could replace how our current early learning education systems work?
BUT, to teach the child how to learn values, morals, decision making…this requires parental guidance, experience, making mistakes, life lessons. Injecting parental guidance and the added layer into children turns literacy into something far greater…it creates the intelligence demanded to make life decisions and think for oneself.
Similarly, Artificial Intelligence bots of course provide an overall function far greater than a mere Chatbot and require a brain and a Natural Language Processor (NLP) to convert and harness the outcomes of conversations, apply some intelligence and convert that knowledge through learning into powerful decision making capability, over the long term.
In addition, there are some supportive requirements in the background that make up and define the eco system needed to build, run, nurture and maintain a ‘living’ AI relationship beyond just a Chatbot.
In simple term those supportive elements are:
The above exposes that if you are not tech savvy, a developer or from the inside caverns of the trade as it applies and relates to Bots, ChatBots and AI developments, and you are wondering what it takes to build a bot or you are considering the very tip of the iceberg when it comes to having your own Bot, then I hope this article is being the sherpa to guide you down the track of the bigger picture surrounding the bot experience itself as an overview (simple) of the ecosystem and what you should be thinking about.
Although the above covers much of the broader infrastructure and potential conceptual hardware related aspects of having your own Bot (Chatbot or AI Bot), the more relevant commercial point to this article is the fundamental principle that Bots are not a set and forget technology nor do they merely solve a problem.
When it comes to the place a Bot holds in the business model and, if your thinking lies around building a bot to solve one problem (that the business has today) so that you can then focus on other areas of the business, you are on the wrong page altogether.
Bots need to evolve in order to realise their true value and in doing so they could and should produce commercial results increasing the bottom line. Rule based ChatBots are less likely to do this for the long term.
Using the mobile phone company example presented above to demonstrate; the company should have Bot related procedures in place such that the human customer service technician will report my feedback and experience so that the Bot and technical development team can update the Bot to improve the future customer experience, thereby allowing the company to utlise the human resources in smarter ways than simple ‘Bot work’.
Of course I am generalising by saying Bots are not a means to an end. Sure they could be. But, in that capacity they don’t realise the potential they could/should, and then they translate into a liability.
The purpose I am building Bots for is far greater. My challenge is to make the reader understand that when built to fulfill an ongoing solution, a solution capable of evolving and adapting with the customer’s needs and journey, right by their side and in their moment of need, that is when Bots apply the ultimate value as an asset to the business which lead to multiple commercial benefits.
My philosophy is that Bots should be built for the long term as an asset to ultimately (and regardless of where in the pipeline or business structure they plug in) to assist the customer journey, experience, behaviour and engagement at all points, thereby enhancing business. For that purpose and in order to help business improve forever more (which never ends and is never complete or over), they require love, care and attention. They have the ability to learn and automatically adapt themselves for improvement, support and maintenance of the customer experience, they just need guidance to nurture them to their true potential.
Other elements are at play once your Bot is built, implemented and installed.
When a Bot (but more so an AI Bot) and indeed the entire supporting ecosystem and infrastructure are correctly built and with the right purpose and objectives in mind, they deliver additional (bonus) by-products which human intelligence investments cannot provide as easily, efficiently, concisely and as cost effectively and these are just a few:
One can spend some money on a Chatbot and fulfill a purpose, which still carries with it all of the above benefits and by-products, but when one spend just a little more and takes steps into the realm of an AI bot, then one makes a long term investment in more than just the right technology but rather, a new business partner powering the profits of the future.
I acknowledge that Bots have a long way to go. The Mobile phone company I dealt with who have lead the charge by adopting early are feeling this pain. Even Facebook’s messenger boss admits it can’t escape the fact that Bots are still in early staging and require ongoing testing, measuring and improvement (you can reference and evidence his article here). But, the good news is that adopting bots early does not mean paying the hefty price tags of the past, for the latest technology revolutions. Bots today are extremely cost effective and you don’t have to leverage PaaS thereby effectively renting someone else’s.
When I frustratingly ranted to the human customer service agent on my mobile phone providers web chat, I sincerely hope he escalated my ‘Bot’ feedback to their Bot development team, to tweak the bots intelligence thereby guiding it to fulfill the human need it had clearly missed on my query and in order to improve for customer experience and saving the company time, money and inefficiency for future customers.
My next ‘brain fart’ which will make for a great follow on article to this one, is all about Brand voice and Brand identity and ensuring these are native to the Bots of the futures programming and to ensure company values shine through, even with Bot engagements…stay tuned.
If you have the need for a bot or just wish to chat about this stuff, I encourage you to contact me.
Article written by Glenn Miller.