AI and business – are you ready for this?

Artificial intelligence (AI) has the power to change the way in which we work, play and interact with everyone and everything around us.

One of the distinct advantages of AI is the ability to ingest unstructured data, be it text, imagery, video or more. With advances in natural language understanding and processing, and progressions in machine learning with regard to the recognition of objects in images and video, data can be analysed in an automated fashion with no human involvement whatsoever.

And this data comes from everywhere: sensors used to gather shopper information, posts to social media sites, digital pictures and videos, purchase transactions, and cellphone GPS signals, to name but a few.

In fact, with access to these exabytes of public and private data, and through AI engines that can process text, video, voice and images (coupled with natural language understanding and generation, sentiment and tone analysis), cognitive processing and analysis can revolutionise the way in which we consume, create and cohabit with current and future technologies – effectively blurring the line between human and machine processing capabilities.

But, what is the benefit to business?

AI can play a significant role in building profit, driving down operating costs of any business while also increasing revenue. Enterprise AI solutions and platforms can streamline operations, drastically reducing the time, cost and effort for all but the most complex of tasks, while at the same time providing rich data and analysis around efficiencies, bottlenecks, processes and resources.

Used in business, any repetitive task that requires a human to consistently and mundanely repeat it multiple times is a great candidate for the use of AI and robotic process automation (RPA) solutions. And this won’t just be at the factory level – we’ll also see positions within fields such as accounting, law and even medicine changing in the future. Human labour will be reallocated and re-purposed into more complex data modelling, scientific analysis and sales-focused tasks, once again increasing revenue while driving down costs.

And we’re already seeing practical examples of AI all around us. As a case in point, one of the more prevalent implementations today would be text-based automated bots. These automated bots can also be voice-enabled, like Google Home and Amazon’s Alexa. Having said that, there are companies that have accomplished much in the field of AI, such as Snips.ai, SoundHound and Rasa. These are companies that do not merely orchestrate cloud-based services, choosing rather to develop their own platform and address a specific niche market.

Taking a practical approach towards AI

However, not everybody is building AI into their business strategies, mostly because it all seems so “out there” – too many talking about AI, not enough showing businesses how to implement it.

So what can you do to find a workable approach to implementing AI?

The first step should be to look at existing manual, human-based services and functions, and secure a clear understanding of which of these can be automated.

Next, follow a clear and defined process of mapping out the required automation steps in detail. Once that is done, choose an AI platform and vendor that will help you map the required automation steps to a list of AI micro-services, formalising integration points and data structures, testing it all, and moving into production.

A practical working formula, as prescribed by Cobus Greyling, technical product manager: Design & Innovation at Ocular Technologies, has been to:

• Start small – using bite-size chunks, work towards a final solution using small, measurable and flexible components as part of the bigger picture.

• Be clear on your objectives – an inherent risk with building AI solutions is that they tend to miss the mark completely, due to the end-result or goal not being clearly defined and documented, and that makes it difficult to gauge the success or failure of the project.

• Execute – many solutions fail due to the components not being fleshed out and tested, so work with agile principles and spend the time executing.

• Measure – make sure your solution or application is built around configurable, measurable metrics and build those directly into your development cycle. If you can’t measure it, you can’t manage it.

• Grow – if your solution works and works well, it can only grow at scale and you need to be ready for it, so make sure your platforms and infrastructure allow for this.

The final piece of advice is that you should have fun! AI inroads can be immensely satisfying and extremely gratifying when everything goes according to plan.

Enjoy making your own mark on AI rules and solutions, and you should find it to be a richly rewarding experience.