Handing over too much control to AI

Have you witnessed the flood of me too opportunists who have jumped onto the AI agent and chatbot bandwagon, boasting of producing solutions which will save businesses time and money, and a subsequent reduction in the number of employees they need?

Well there is some truth behind some of these claims, but there is a strong bias for many of these AI agent and chatbot developers in favour of growing their network and building communities of followers.

Is that even a problem?

Well no, there's nothing wrong with networking and building a growing community, however...

Let's assume that you are a business owner who has paid a lot of money and placed a lot of trust in someone to produce you a robust and reliable tool to handle many important functions, including directly communicating with existing and potential customers on behalf of your business.

How happy would you feel for that solution to then be given away for free to anyone who wants it?

Allow me to explain....

Many hopefuls continue to jump onto this bandwagon of creating (or cloning) methods of getting AI to do the work of many skilled and experiences people at a fraction of the cost, with instant results and exaggerated claims of the huge savings in time and money.

But instead of focusing on building and maintaining robust tools and system integrations which can reliably serve a business long into the future, many AI chatbot and agent developers use social media posts to emphasize the speed (often in minutes!) that they have been able to create a new tool or flow

They'll claim that this new tool was sold for big bucks, but with the objective of rapidly growing their network on LinkedIn, they can be seen to offer their new miracle template for FREE to anyone who:

(a) responds in the comments with I WANT IT and
(b) connects with them.

That business model was never something which featured in my 29 years of doing what I do... (more about that later)

I would actually be more concerned with IP (intellectual property), NDAs (non-disclosure agreements) and general business ethics and etiquette, instead of publicly offering my latest work for free so that a community of followers can copy it.

Thankfully not everyone is falling for it, but the lack of testing by many chatbot and AI agent developers is also concerning.

I have quite a long web career with a extensive background in manually coding, developing and managing websites and large eCommerce systems. At time of writing this, I'm in my 29th year in this industry and it therefore came as a kick in the **** to read claims that myself and anyone like me would soon be history due to AI.

Whilst I haven’t been exclusively developing code throughout those 29 years, I have played several related roles, from design, development, business analysis, project management, etc. I have worked on multi-$million global projects, and clocked up many air miles in travelling between here in the UK to various parts of the US and Europe.

So what? you might ask

I say this to highlight the areas of responsibility I had... my expertise, my understanding of the business requirements of international partners, franchisees, markets, etc., and also of the varied compliance rules of different countries around the globe.

From a technical perspective I understood the dependencies within our systems, the threats and opportunities of repurposing, replacing or enhancing certain front-end or back-end components or systems, etc., and much more.

I also know, through years of experience, how important it is to test thoroughly.

Throughout my career, the importance of capturing and communicating business requirements has been vital, and it takes time...

Soliciting requirements

Requirements gathering requires not only analysis, but it also needs real people talking to real people about what they want and what they need (needs and wants are not always the same!), as well as the value, impact and cost of implementing their requirements.

Managing stakeholders

Quite often myself and those stakeholders would need to negotiate on priorities, budget, and MVP (Minimal Viable Product, which can be an important concern when a large project or range of projects cannot be launched in one go.)

Testing - where too many corners are cut

I also never underestimated the importance of testing, including functionality testing, user acceptance testing, regression testing, and user testing (which is not the same as user acceptance testing)

These different types of testing approaches should be carried out for all relevant scenarios, and following testing methods where inputs and expected results can be measured, assessed, approved, rejected, etc., with any bugs being fixed and retested.

Unfortunately testing is often overlooked or regarded by many as an unnecessary cost in both time and money.

With some of the chatbots I had built... and also thanks to my approach to thorough testing and ongoing testing... I realised with horror that I had got stung by the way that changes to the platform I was using to build chatbots had suddenly started causing some of my bots and agents to respond in completely inappropriate ways, including changing prices and product specifications, etc.

Businesses need to remember that a customer facing AI chatbot or agent is communicating on behalf of that business. It’s impossible to anticipation every conversation and user intent, but the reality is that the potential to frustrate or mislead customers is pretty high.

The danger is that business owners might not realise what users (potential customers) are being told, and users might not realise that they have been misinformed until the consequences of those communications lead to problems, ranging from disappointment or inconvenience to MUCH worse!

So as much as I admire and appreciate the efficiencies that AI tools can bring to businesses, there needs to be some level of control that shouldn’t be handed over to AI.