• THE AI START-UP SOLVING BUSINESS PROBLEMS

    Date:6 November 2018 Author: Brendon Petersen Tags:,

    Artificial intelligence (AI) is revolutionising the way business leaders address age-old problems. Not solely the for the world’s leading technology companies anymore, AI solutions that are accessible and cost-effective can be implemented to sustainably improve customer satisfaction, pinpoint where best to reach potential customers and accurately forecast inventory requirements.

    AI is a buzzword in the tech industry at the moment, how do you describe it to the average man on the street?
    AI is the ability for machines to interact with humans the way another human would in a specific context. An example of AI in action is an ATM doing the work of a human teller by providing you with cash. In two decades, probably 90% of professions we know today will be taken care of by machines in this way. We, humans, will focus our time on new, more creative endeavours.

    “Our quick, accurate and cost-effective solution enables businesses of all sizes to set themselves up with the competitive advantage of using AI to optimise their data, grow their business and boost their bottom line.”

    This is according to Vian Chinner, a South African innovator, data scientist and CEO of Xineoh – the Canadian machine learning company using its cutting-edge consumer behaviour prediction algorithm to turn existing business data into usable insights.

    Chinner explains that, currently, the amount of data generated by businesses doubles every 18-months. “The human resources that companies have in place to interpret this data, however, stays the same meaning that the potential of the data is not being realised. At the same time, the capacity of AI innovations has far surpassed that of human ability.”

    According to global professional services firm, Accenture, the average business utilising AI will increase its revenue by 38% by 2022. Putting this theory into practice, popular entertainment subscription service, Netflix, has estimated that its AI algorithms save the business $1 billion each year.

    With this in mind, Chinner emphasises that it is as important for businesses to utilise AI in 2018 as it was to adopt electricity after the second industrial revolution.

    “While more and more industries are realising what an important role AI can play in helping them achieve their goals, the options available to the majority of businesses has been limited to-date. It is estimated that there are between 5,000 and 22,000 AI engineers in the entire world. Access to this AI talent is further limited by large technology companies which have monopolised 90% of the available expertise.”

    “Ruling out the big-budget option of hiring an experienced AI team in-house, alternative bespoke solutions currently available are often also too costly and time-consuming to be feasible long-term,” he adds.

    In terms of more affordable or free platforms, Chinner explains that they are not as accurate as they may seem.

    “Most existing algorithms have a strong bias towards recommending the most popular items. For example, if one were to predict what is in the average consumer’s monthly shopping cart, including popular items like bread and toilet paper would likely result in an overtly accurate prediction. The limited scalability of these solutions also often results in their failure when it comes to delivering on processing demand – providing thousands of results when the big data of an organisation may require processing in the millions.”

    How is Xineoh using machine learning and AI with African and South African companies?
    Xineoh helps companies predict consumer behaviour with AI. South African companies have the same needs as their international cohorts. Data collection is exploding while the human capacity to interpret this data is very limited.
    The human mind can hold an average of five objects in its working memory. So, if a human analyst tries to infer patterns, he/she will be able to consider only five examples at any given time. AI, on the other hand, can solve the same pattern recognition problem across billions of cases at the same time. This is how AI can achieve such extraordinary results – like predicting a person’s big five personality traits better than their significant other can, by merely analyzing 80 of their Facebook likes.
    Creating a cutting-edge, bespoke predictive model requires a very talented data science team. The amount of people in the world who can build this type of solution is estimated at between 5,000 and 22,000. Of these, 80-90% work at the dozen largest Silicon Valley companies. South Africa is estimated to have between 0 and 100, and in practice, it is likely much closer to 0.
    Large South African companies, therefore, have the same fundamental problem as international companies, namely the tremendous need for AI while facing a severe lack of the data science skills required.
    Xineoh provides a solution to part of this problem by providing a platform that can solve this consumer behaviour problem with a standardised solution that can be deployed to an enterprise very rapidly, in most cases only a few weeks. This is in comparison with a bespoke solution which has a 12-18 month development cycle.

    So, how can businesses realistically tap into the benefits of AI and stay competitive in their industry as the fourth industrial revolution plays out?

    Proven effective in the financial, retail, media and entertainment, as well as e-commerce industries, Xineoh delivers leading results.

    “Our predictions outperform others in terms of implementation turnaround, accuracy and popularity inclusion.”

    “We can provide businesses with relevant customer behaviour pattern insights within just two weeks of receiving their transactional data. This is in contrast with the six to 18-month timeframe of other AI solutions on the market. Our algorithm also offers scalability for businesses working with big data and presents an exceptionally low-level of popularity bias.”

    What are some of the benefits of machine learning and AI for the average man on the street?

    AI is touching the lives of people in both big and small ways. No day goes by without AI enriching your life. Here are two brief examples. In a small way, Airbus has started fitting airplanes with airplane dividers which have been developed with AI. These dividers have a higher tensile strength than traditional dividers but are 50% lighter. This will lead to fuel savings that will ultimately lower flight prices if all other pricing factors remain constant.
    In a big way, a Boston based company called Neon Therapeutics just published the most successful phase 1 melanoma treatment trial ever. They treated 12 advanced stage 3 and 4 patients and achieved a 100% complete remission rate in 24 months – this is a patient population who has a 5% predicted survival rate. They did this by analysing the tumour DNA and using machine learning to predict which antigens (neo-antigens) will be present on the tumour cell surface. They then created a peptide vaccine which teaches the immune system to look for these neo-antigens.
    Implementations such as these will likely revolutionise every industry. Medical treatment, in particular, will probably come from vending machines within the next few decades, just like you currently do most of your banking via an ATM.

     

    Referring back to the shopping cart example regarding popularity inclusion, Chinner explains that Xineoh’s solutions don’t just note the most popular patterns but delve in deeper, providing detailed and accurate insights on the behavioural patterns of a business’ different customer segments. For example, noting which consumers would be more likely to purchase whitening toothpaste over regular and which consumers would purchase luxury branded cereal in the R50 to R80 price range.

    “By matching people with products and patterns, inventory with opportunities, and prices with spending propensity, Xineoh helps businesses improve customer satisfaction and accurately predicts their inventory needs, where and how to market their offering for the greatest target market reach, and what part of their process is most responsible for customer turnover.”

    Which South African companies – if any – is Xineoh working with? 

    We work with various companies in the retail and media space. We are also about to conclude an agreement, which will result in us building by far the most ambitious machine learning project undertaken in the South African media industry to date.

     

    For more information, please visit: https://www.xineoh.com/.

    You may also like:



    Latest Issue :

    November 2018