Introduction
Businesses across the world are all being impacted by AI and using it in some form in their business. They are also beginning to realise the opportunities it presents in terms of business transformation and creating or improving competitive advantage.
It is worth noting though, that small businesses do not have the resources to deploy a lot of AI due to cost and skill requirements. Despite this, there are any number of AI solutions and applications that can be accessed by organisations of all sizes to improve productivity and sustainability.
In the image below from the 2022 McKinsey Global Survey on AI (Chui et al. 2022) [^1], the use of AI is broken down by business function. Interestingly, the absence of Administration is noted, and yet this is one area where AI applications abound.
Figure 1: Commonly adopted AI use cases
Engaging and serving customers
When discussing this subject, the term chatbots should immediately spring to mind. These have become endemic across organisations in a drive to improve efficiency. The question of efficiency is a debatable one of course. No doubt they are more efficient for the organisation, but do they create greater efficiency for the customer? That answer is probably dependent on the circumstances or use case.
Chatbots are improving in their abilities to process data and that this will be a continuing trend. Organisations are looking to use chatbots in order to make their communications as efficient as possible with their customers. This is not only for customer support, but also for making recommendations for purchase.
Chatbots and intelligence
The level of intelligence a chatbot has is determined by a number of factors including the intensity of machine learning, the quality and amount of data used to train the bot and also natural language processing.
Understanding the context
Understanding the context is possibly the prime requirement of a chatbot. If it does not grasp the context, then it cannot possibly direct an enquiry efficiently (Reddy 2018) [^2]. A chatbot would be trained on contextualising sentiments such as abuse, sarcasm and anger, yet all the while reverting to its “mission” to respond in a calm and helpful manner.
Consistency
A chatbot should be consistent in its tone of voice and as human like as possible to emulate interaction with a human. It has even been suggested that a chatbot could have a sense of humour, but would need to be understated as not everyone appreciates humour and it varies across cultures (Fayon 2017) [^3].
Continuous learning
Chatbots are developed using machine learning. Consquently, the bot will continue to learn throughout its deployment and directing you to the information you are enquiring about. Liken this to a robot vacuum cleaner learning its way around the house.
Retail and e-commerce
There is no doubt that AI has completely revolutionised the way e-commerce works. AI now works to:
- Record user activity on the site
- Make recommendations of products based on previous purchases and views
- Seek testimonials for purchases to boost product reputation
- Send reminder emails that check out was not completed
- Recommendation of similar products
- Provide colour switching on many products
- Offer discounts if user indicates they are leaving a page
AI driven algorithms create personal data profiles on you based on everything you do online. Not only does this data get used in marketing, it can actually predict and suggest products based on your history with the vendor. Such a personalised approach drives a much stronger relationship between the vendor and the customer, privacy considerations aside.
It is interesting to note that Gupta and Kim (2007) [^4] state that “repeat customers are five times more profitable that new customers”. However, they also suggest that in excess of 50% of customers do not go on the make a third purchase. In support of this argument they cite Reichheld and Schefter (2000) [^5], but again, the date of the article cannot take the advances in AI over the last 20 years into account.
A clear picture seems to forming however, that the critical aspect of the success or otherwise of e-commerce is customer data in order to drive the experience. It is noted that in the above video though, when referring to Amazon, that they pass on the sale to the customer, but not the data (The future of shopping: what’s in store? 2021) [^6]. The actual vendor then has no control over the data or the tools to build customer retention and future sales, only Amazon. The video also provides the example of Nike who have opened their own stores so they can capture and capitalise on the data they collect.
Improving workplace communications
The bane of any knowledge worker’s life is email. It can often seem as if your entire day is taken up with reading and responding to them. According to Kooti et al. (2015) [^7], as email volume increases, recipients may reduce their level of responsiveness, but also use shorter replies. Oddly enough though, they also claim that peoples’ responsiveness does not change and even that they might be quicker with their response times.
AI will make a considerable mark on communications in the workplace. Rather than elaborate in text, the video below from Microsoft is offered as evidence of the coming changes. Note that the software, Microsoft 365, demonstrated in this video was deployed for testing in March 2023. Although not in general use as yet, it is an eye-opening indication of what is coming.
Assisting with healthcare
It has already been mentioned how AI was able to assist during Covid by speeding up diagnostic technology. Healthcare is an area that has enormous potential for utilising AI in many different ways. Daley (2023) [^8] cites several examples of AI being used in cancer diagnosis, particularly for early detection, and then individualised treatement plans. The predictive power of AI has been used by MIT with a model that can detect the risk of people contracting lung cancer (Ouyang 2023) [^9]. The video below explains how it works in greater detail.
Then there is a model developed by a neurosurgeon, Dr Daniel Orringer. The video below demonstrates its use in the case of an individual considered to have Glioblastoma. This is the most common form of brain cancer with a life expectancy of eight months. Under normal circumstances, the process is that the patient has surgery to remove a sample of the brain for testing. The testing usually takes three to four weeks, but using the AI model, the time to result is three minutes.
Improving cybersecurity
Earlier, it was mentioned about the ability of AI to detect patterns and this is a valuable opportunity in the world of cybersecurity. Transactions in an organisation will have a pattern indicating their normality. An incoming attack would not resemble the same structure and can be flagged for investigation due to suspicious activity. The video below from IBM explains several aspects of cybersecurity.
According to Columbus (2019) [^10], the most common use of AI in cybersecurity is for the use of network security. He goes on to note comments from Brian Foster that “the most successful cyberattacks are executed by highly professional criminal networks that leverage AI and ML to exploit vulnerabilities”. So, it is not just a tool business and other organisations are using for cybersecurity, but the criminals as well. The increasing use of AI puts even more pressure on legitimate networks to keep pace with developments and stay a step ahead of the criminals wherever possible. Recent publicised attacks here in Australia have highlighted the exposure that major organisations have in their networks and the potential for public backlash where private data is compromised.
The video below looks at AI within the cybersecurity area and how it can be used to defend against attack. It should be noted however, whenever reading or viewing information on AI and cybersecurity that it is exceptionally date sensitive. Change in the area is rapid so currency of articles and videos is important.
Logistics and supply chain
The development and deployment of self-driving cars and trucks will have a huge effect on the supply chain when it eventually occurs. This development is inevitable according to Peters (2018) [^11], although she does acknowledge that this may be some time off due to the risk to human life.
In other areas though, we see huge warehouses being established by Amazon and Alibaba. The picking is all done by robots that are controlled through an AI network. The mechanistic manner in which they operate delivers considerable improvements over human picking. The video below claims the robots take 15 minutes to pick an item compared to a human taking 45 – 60 minutes. The AI knows where every package is and where it needs to go once picked. The video below demonstrates how these robots are working in an Amazon factory.
The streamlining of manufacturing
There is considerable similarity between the industrial revolution and the age of AI we are currently seeing. Business will not adopt systems that are more expensive and AI is offering the ability to decrease costs, extend operating hours and improve quality. According to Dorfman (2018) [^12], robots offer precision assembly, even at a micro level and also make workplaces safer for humans by not exposing them to dangerous work that can be completed by a robot.
The video raises some interesting questions around social justice. What is in store for our world if millions of workers in developing countries are replaced by AI. There is mention of a T Shirt manufacturing business starting up in Arkansas claiming that production costs will be on a par with those in developing countries. We can only wonder what the reaction of these people in developing countries will be if AI drives them into deeper poverty.
Conclusion
This module has explored how AI can be used in business. It has also raised some interesting questions about the externalities that AI may create. There is probably no doubt about the potential and power of AI, but it is the effect on the livelihoods of human beings that is concerning. It is reasonable to argue that in the developed world that workers replaced by AI will either find new jobs created by AI as claimed. Alternately, there is a case for a universal income to be adopted. Will developing countries have the ability to transition and maintain some degree of dignity for its citizens?
References
[^1]: Chui, M, Hall, B, Mayhew, H, Singla, A, & Sukharevsky, A 2022, The state of AI in 2022—and a half decade in review | McKinsey, viewed 10 July 2023, https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2022-and-a-half-decade-in-review.
[^2]: Reddy, R 2018, ‘How To Make A Chatbot Intelligent?’, BotCore November 9, viewed 15 July 2023, https://botcore.ai/blog/how-to-make-a-chatbot-intelligent/.
[^3]: Fayon, N 2017, How To Give Your Chatbot Soul, Medium, viewed 15 July 2023, https://chatbotsmagazine.com/how-to-give-your-chatbot-soul-3fdda6f3c369.
[^4]: Gupta, S & Kim, H-W 2007, ‘The Moderating Effect of Transaction Experience on the Decision Calculus in On-Line Repurchase’, International Journal of Electronic Commerce September 1, Vol. 12, No. 1, pp. 127–158, Routledge.
[^5]: Reichheld, FF & Schefter, P 2000, E-Loyalty: Your Secret Weapon on the Web, Harvard Business Review, viewed 15 July 2023, https://hbr.org/2000/07/e-loyalty-your-secret-weapon-on-the-web.
[^6]: The future of shopping: what’s in store? 2021, viewed 15 July 2023, https://www.youtube.com/watch?v=ad-GuV6YIMI.
[^7]: Kooti, F, Aiello, LM, Grbovic, M, Lerman, K, & Mantrach, A 2015, ‘Evolution of Conversations in the Age of Email Overload’, Proceedings of the 24th International Conference on World Wide Web May 18, pp. 603–613, International World Wide Web Conferences Steering Committee, Florence Italy.
[^8]: Daley, S 2023, AI in Healthcare & Medical AI Examples to Know | Built In, viewed 17 July 2023, https://builtin.com/artificial-intelligence/artificial-intelligence-healthcare.
[^9]: Ouyang, A 2023, MIT researchers develop an AI model that can detect future lung cancer risk, MIT News | Massachusetts Institute of Technology, viewed 17 July 2023, https://news.mit.edu/2023/ai-model-can-detect-future-lung-cancer-0120.
[^10]: Columbus, L 2019, 10 Predictions How AI Will Improve Cybersecurity In 2020, Forbes, viewed 17 July 2023, https://www.forbes.com/sites/louiscolumbus/2019/11/24/10-predictions-how-ai-will-improve-cybersecurity-in-2020/.
[^11]: Peters, C 2018, ‘5 Ways AI Will Transform the Logistics Industry’, AltexSoft November 2, viewed 17 July 2023, https://www.altexsoft.com/blog/business/5-ways-ai-will-transform-the-logistics-industry/.
[^12]: Dorfman, P 2018, 3 Advances Changing the Future of Artificial Intelligence in Manufacturing, viewed 17 July 2023, https://redshift.autodesk.com/articles/future-of-artificial-intelligence.
