The winner is not the best model, it is the best data

AI and Machine Learning (for the purpose of this article, I will refer to as AI) is complicated and only for trendy start-ups, tech companies, or any business with a surfeit of large bean bags, right? Actually NO.

The fact is, that many well-established businesses can benefit significantly from AI, essentially because the key component to the best AI solutions is data and lots of it. Clearly, start-ups and relatively new companies won’t have lots of data as they haven’t been around long enough to gather it or have had the time and resources to design their systems to be capturing the right data. Whereas established businesses, that have survived long enough to become mature, have likely been continually gathering and working with their data.

There is likely to be expertise in various functional areas of your business that have real-world experience to interpret the data and, with the right help and technology, start to unleash the power of an AI platform to accelerate their evolution and leap ahead of any new market entrants. Of course, it’s not all rosy, as you may have to deal with the inertia of ‘we’ve always done it that way’, which assumes that there is only one way – this is why demonstrating fast results can turn around any reticence and bring along the naysayers more quickly than you might assume.

Your data, combined with proven and established technologies, can be turned into real business insights that can provide indications of new growth opportunities that you probably intuitively thought might be there, but can now be supported by facts and evidence. It can expose opportunities you had no idea existed and will certainly identify issues and opportunities that humans might miss – just due to the vast amount of data that AI is able to process and glean unseen patterns from.

The issue for businesses, blessed with years of experience and data to go with it, is to be able to use their data assets effectively. Data quality, location and accessibility is the challenge. However, using these tried and tested methods and combining these with AI itself to assist the process, you can identify what’s good, what isn’t from what may seem like an over-whelming mish-mash of formats, data-types and duplication.

Honing the AI to walk-through your data and identify the key elements will be more efficient than humans and allows you to quickly build up the repository of data needed to set the AI engine loose to start providing insights into your data. Often, these insights will have been either impossible to get to up until now, or just hidden by the volume of ‘noise’ that humans would have to wade through. This noise, even in itself, is a big deterrent for a lot of companies, who find it better to put the time and resources into their offerings rather than clear out the accumulating proverbial hoarder’s pile in the attic. But therein lies treasure.

One issue, I’m sure that you’ll have found, is lots of technology providers telling you that they have the best tech, but that is but a fraction of the challenge. What most businesses need is experienced, battle-scarred expertise to help cut through the hype and identify the real challenges within your business to implementing an effective AI solution.

I would suggest you start with companies that can demonstrate this real experience of implementation. Find a company who can tell you war stories, not sell you the world. Not just building Technical Proof of Concepts and Prototypes, but real applications of the technology, working to support businesses that exist and can demonstrate a real ROI e.g. how much it costs to analyse, build, test and deploy, versus the return as a result in hard numbers of increased sales, returns from new business opportunities or reductions in operational costs/increased efficiencies.

The first step in the process, as we have found from experience, is to really get a clear idea of why you are doing this in the first place, formulate an AI Strategy, the most efficient way to do this is probably by utilising some external expertise, again from someone who has experience and a structured process – resulting in a solid business case, based around a easily achievable goal that will deliver benefit on day one.

This may be as simple as answering a business question that has been particularly difficult to answer to date, or even to use the technology to create that lake of clean data that you can then allow your business users to explore with confidence. This will require a detailed understanding your data, where it is located, what structure it’s in and the quality of it.

Get an idea of where you and your business currently stand by completing the free to access EC Risk Scorecard and get an overview of our Data Discovery Programme to understand our approach to this critical initial phase and what it will deliver to set you up properly to succeed.

From there, you can take various directions. At EC, we help you to take small, manageable steps towards the delivery of your AI project, with quantifiable deliverables at each end-point to demonstrate their incremental value, and resulting in the building of a business case to support the initiative. This is part of the EC Data Discovery engagement.

Find out more about the AI projects that Elemental Concept has worked on, across various sectors and organizations of varying maturity. The key thing for them all is that they are likely to be much like you if you are reading this: Data-driven businesses, sitting on a gold mine of data.

If you would like to know more, at a detailed level, reach out to our Head of Data Science, Ghislain Landry Tsafack, who leads our AI strategies. Ghislain has drawn on years of research, insight and on the ground experience to create a go to guide for Ethical Frameworks for sustainable AI and Data Science and is perfectly placed to advise on how to responsibly optimise your data for the best returns for your business.