To AI, or not to AI, when is the question…

If you’re going to, do it right and don’t just use it to speed up your bad processes

The end is Nigh!

AI conjures extremes, from visions of Skynet, to humans never having to raise a finger and do anything ever again. The reality (for at least the near future) is not as extreme, as most organisations are generally not mature enough with basic data and analytics, let alone more advanced approaches like AI.

"AI will be either best or worst thing for humanity" Stephen Hawking

The average maturity of Agriculture, Electricity, Transportation, Financial Services and Insurance industry clients we've assessed sit in the reactive space as show below. Meaning they generally there is an awareness and reactive response to data issues, but its typically siloed and not effectively managed.

This low level of immaturity significantly increases the risk of AI projects failing due to poor data foundations.

When data and AI is done right, organisations will differentiate themselves from their competitors, improve customer experience, accelerate business goals and long-term objectives. 

Not done properly, organisations could face significant reputational, regulatory, operational, financial and strategic ramifications or fines.

Unfortunately, organisations that are dabbling in AI, tend to use it to speed up bad processes or work around legacy systems. Is this really the best use of AI?

Out of the PoC phase, into the fire

Gartner and we have seen that organisations are still struggling to progress past Proof of Concepts (PoC) for AI. As digital transformation accelerates, AI and algorithms will used in more complex and dominant ways. However, the current challenges and pitfalls that plague existing black boxes and models are aggravated and obfuscated by AI’s increased complexity and lack of understanding. Some of these are:

1.      Trusting your black box and knowing exactly how it comes to an answer – ethical and fair

2.      Knowing your black box inputs, algorithms and outputs are correct, complete, accurate and unbiased

3.      Dealing with bad data, errors and exceptions

4.      Ensuring transparency within the organisation on how it works

5.      Maintaining privacy of your customers and security of their data

How I learned to love AI

Get your strategy, people, process, systems and data foundations in sufficient order, then you can get down to brass tacks.

Some things to think about:

1.      Provide AI and algorithm transparency to customers and internally. Consider education and openness vs secrecy and obfuscation. Most of the time if you can explain it simply enough, you don’t understand it yourself – at that’s a problem in itself

2.      Understand your why. What the customer and business problem being solved. Consider aligning your AI strategy to your company, Digital/Data and IT strategies

3.      People are key. Focus on culture and will not skill. Consider the change management required, people with the right Growth mindset and upskill them in business and technical skills

4.      If a process is bad remove it. Your customers and employees will thank you for it. Consider customer and project journeys to identify these bad processes

5.      Build secure, privacy and regulation by design. Consider security, privacy and regulatory requirements upfront. Translate complex concepts into easier to understand principles. While Ethical data governance is not newsworthy, a data breach sure is

As the AI winter begins to thaw. Technologies like Quantum computing will drastically accelerate your move to AI spring. However, if we don’t start work on your foundations now, this will just push you back into a blizzard, while your competitors are basking in the sun.

If you want to know more, would like to see how your organisation can prepare for AI and improve your information Maturity scale please feel free to reach out.