AI for business, all you need to know Part 1
In this article, AI in business, all you need to know Part 1 – I will be covering what is AI, why do we need it, the benefits and business case.
Organisations need AI for the same reason that they need websites – it can give an advantage over the competition in the short term, and if they don’t get it in the long term they will fall behind the rest of their industry as adoption becomes mainstream.
There are literally unlimited business cases for AI, but like any other technology innovation they must be tempered with a cost versus benefit investigation. Currently, there is a lot of interest in AI across business sectors, and leaders sometimes demand ‘we must have AI’, rather than focusing on the specific problems that they could cut their teeth on.
Is AI threatening jobs? I don’t think so. Poor business models, bad customer service and wrong leadership decisions are threatening jobs far more than AI. Organisations that take advantage of the possibilities to amplify human performance with AI are always going to be the long-term winners. There is a company called Stitch Fix that recommends and delivers an outfit personally chosen for each customer every month. AI algorithms trawl through thousands of clothing items and automatically produce short lists based on each customers’ unique preferences, but human stylists make the final choice for the best fit. Stitch Fix currently employs 2800 stylists – a great example of jobs that wouldn’t have existed without AI to augment them.
Let’s get down to answering some commonly asked questions about AI.
What is AI, and what isn’t?
AI, as we see it in business today, is properly known as ANI, or Narrow AI, where specific tasks are solved with software algorithms in a way that can seem human-like. These algorithms learn from large amounts of historical data, and eventually find patterns and trends which help them process new input, rather than being pre-programmed for every eventuality in advance. Examples include identifying spam emails (based on historical categorisation of spam by users), flagging customers most likely to buy from a competitor, and cars that drive themselves. We are still a long way from AGI, or General AI, where software can deal with any general human task requiring intelligence. Steve Wozniak, Apple’s co-founder, said famously that he would believe AGI has arrived when a robot can enter a stranger’s house and make a cup of coffee.
What isn’t AI is when software uses a traditional rules engine based on pre-existing conditions to make decisions. For example, in the UK, there is a famous phrase used for predicting weather, ‘Red sky at night, shepherd’s delight. Red sky in the morning, shepherd’s warning.’ Any software that used these inputs as a mechanism for predicting weather would be rules based, not AI. However, any software that could recognise a specific photo as a red sky, rather than a red tomato, is likely to have been developed using AI.
Why do we need AI?
The human race needs AI to help solve a number of big problems that are too expensive or too large-scale to solve using human ingenuity alone. In the medical field, AI can help tag diseases such as lung cancer from CT scans and X-Rays with far higher accuracy and capacity than human doctors. On the consumer level, Microsoft have a ‘seeing eye’ application that feeds blind people aural input on their surroundings, such as ‘two smiling children on a swing’, which could only be accomplished in non-AI life by having a human accompany every blind person and narrate the journey. In food production, AI can help farmers improve yields by deploying drones to detect weeds and plant growth issues, as well as processing soil sensor data to identify what impacts high production in certain soils. In energy, AI could integrate renewable energy sources into an existing grid, such as Eskom’s, in an optimal way that suits demand predictions better. Business needs AI, like most other technologies, so that it can outperform the competition.
What are the benefits of introducing AI-powered solutions to the business?
Business organisations need AI for the same reason that they need websites – it can give an advantage over the competition in the short term, and if they don’t get it in the long term they will fall behind the rest of their industry as adoption becomes mainstream. Judicious application of AI can provide the traditional benefits of business software, such as reducing costs and increasing profits, but it can also provide whole new business approaches. A classic example is Amazon’s AI recommendation engine, which suggests potential items based on your purchase history and those similar to you. Now every online supplier has a recommendation engine – but Amazon still gained massive competitive advantage from it initially.
What are the business cases for AI?
There are literally unlimited business cases for AI, but like any other technology innovation they must be tempered with a cost versus benefit investigation. If there’s no tangible monetary or transformational benefit from using AI to solve a specific business problem for your business, then there isn’t a business case. There does seem to be a lot of interest in AI across business sectors currently, and leaders sometimes demand ‘we must have AI’, rather than focusing on what are the specific problems that they could cut their teeth on.
AI can help in all areas of digital transformation. For empowering employees, Microsoft provides email Office365 users their weekly dose of ‘MyAnalytics’, labelled as a fitness tracker for the workplace, where AI identifies engagement with other co-workers through Outlook usage and Team meetings and identifies possible new connections with other employees to improve efficiency. To improve customer engagement AI can be used to recommend additional products that are appropriate for an existing basket at either an online store, or at point-of-sale at a traditional shop. An optimising operations AI example would be the use of predictive maintenance, where any process that maintains equipment can be enhanced to better predict machine failure thus saving costly production line issues, as well as optimise capital by avoiding stockpiling unneeded spares. A typical product transformation approach would be where customers turn their product into a service, and use AI to ensure maximum uptime, possibly by using predictive maintenance as above – General Electric being the classic example, where they used AI and big IoT data to move from selling turbines to providing energy-as-a-service.
What are the top AI applications?
We use AI applications every day without even realising – for example, Waze rerouting your mid-school run, Siri understanding your child’s voice when he’s putting on an Italian accent, and your bank calling you when they catch a North Korean credit-card transaction. Even classic business apps that have been around for decades are being enhanced with AI-based functionality – see Excel’s new AI-enhanced ability to import a spreadsheet from a photo taken with your Android phone. Chatbots, cyber-security and marketing related solutions like targeted adverts are probably the most used AI applications in business today.
AI applications that have been around the longest include weather forecasting and code-breaking solutions. However, the applications that really seem to catch people’s attention now are self-driving vehicles, tumour-identifying algorithms and collaborative intelligence that allow robots to operate on the manufacturing shop floor next to human workers.
Nearly every enterprise today claims to be using AI in its products or services. How do you evaluate how good a company’s AI really is?
One way to evaluate the quality of AI involved in marketing is to put yourself in the position of a customer of that enterprise. If the AI enhancement improves your customer experience in some way, such as advertising a feature that you have always wanted, or offering a service that really suits your lifestyle, then it’s good AI. I can tell my bank of twenty years that its AI is not great – given my banking profile, I really don’t need a funeral cover policy, and I especially don’t need a targeted advert for it on my cell phone every week. However, some model somewhere, presumably intelligent, keeps picking me out as a potential purchaser of a funeral cover, despite my continued indifference – so the model is not learning from mistakes.
Which companies do you think are leveraging AI well, and why?
There are many companies that are leveraging AI well for optimising certain internal processes but those are difficult to determine from the outside, as these are often businesses efficiency differentiators. From my point of view as a consumer, Discovery Insure does a great job – they use AI models in conjunction with vehicle telematics to tell me how badly I’m driving, combined with smoothie-based gamification to drive adoption. Their AI has (almost) stopped me from using my cell phone at all whilst driving. Netstar also does a great job in a similar area, with a lot of emphasis on safety related incidents, such as crashes.
With regards to empowering employees, Altron Karabina works with Clevva.com, a company that is using AI to enhance call centre agents’ capability to cope with client requests. Their virtual assistant bots are improving first-call resolution and net promoter scores by offering the right level of contextual help to agents, with models that learn from historical call data, not by building complex rules engines manually.
With regard to product transformation, I’ve just come back from an Australian trip where I Airbnb’d and Uber’ed around everywhere – it’s difficult to avoid talking about those two companies when they’ve become verbs already. Cliched examples, I know. I’m waiting for AI to aid disruption in South Africa in a similar way for areas such as banking, insurance, assurance and traditional retail, with benefits for both staff and customers.
That’s it for AI in business, all you need to know Part 1. Part 2 will be published soon, so keep an eye out for it. Let me know if you have any questions in the meantime. Would love to chat!
Contact us for more information on leveraging AI for your business.