The Practical Application of Artificial Intelligence

“Artificial Intelligence will have a more profound effect on humanity than fire, electricity and the internet” – Sunder Pichai, CEO of Alphabet

Just like many others, I am intrigued by the transformation that Artificial Intelligence(AI) is going to bring on how we think, feel and execute. While the philosophies surrounding the good, bad and the evils of AI can be deliberated for days. For organisations in the midst of this revolution, it leaves a looming question of participation. The FoMo (fear of missing out) is justified. At the least organisations need to understand relevance of this evolving technology to their day-to-day business and alignment with broader organizational objectives. Are organisations able to leverage the technology and understand how to use AI most effectively to build a profitable business that delivers value to shareholders and the community at large?

My path to the world of data analytics hasn’t been a linear one. But the opportunities I have had in setting up centres of excellence in data analytics, research, sales and marketing teams and working in consulting capacity with clients and internal teams across different cultures and regions in different business models has given me some insight.

Every organization is different in that they have their goals, their own set of challenges and ambitions.? In my experience there are a few fundamental use cases for AI to deliver immediate value to most organisations. The themes below don’t even begin to scratch the surface of what is possible with AI. Organisations, large or small; startup vs fully grown and established have to evaluate the purpose and materiality of AI deployment. Lets start with the most obvious use case, automation.

Some Use cases

Expediting transactional processes and automating simple and repetitive tasks. ?Data from Fortune Business Insights show that retailers in India spent $5.50 billion in 2022 on AI-led automation and projects and this spending is projected to jump by 34% compound-annual-growth-rate (CAGR) to reach $55.53 billion by 2030. Automating the cashier’s role or the storekeeper’s role in inventory management will allow them to spend more time reconciling profits and addressing customer queries or better still, acquiring new customers. Maintaining a well-stocked inventory is critical to retailers, supply chain optimization through automation tools can significantly help improve efficiency, speed, accuracy and give staff an opportunity to upskill themselves on more value adding tasks that need decision-making and client interaction. Not to forget keep staff costs down, improve margins and accelerate time to market. This use case is applicable to most industries out there.

Administrative AI is a term most commonly found in healthcare, where experts agree that AI could automate many repetitive tasks such as recording patient notes, patient billing etc. With growing healthcare costs across the world AI is being used to build leaner value chains.

The most obvious use case in this category is the processing of large manual files month on month which engages a repetitive process. This could be with HR – compensation and benefit teams, finance and every other team that engages in sizeable data analysis.

Product centricity and improving customer experience.? AI Large Language Models(LLMs) can help sifting through large sets of data such as customer browsing, purchasing data and consumption insights that can help build customer personas and create relevant personalized/ hyper-personalized products and services for clients. Market research, customer sentiment analysis, customer communication and engagement through chat bots are some of the most common applications of AI for improving customer acquisition and enhancing customer experience. From a simple AI pop-up window on your personal banking application to the AI-augmented perpetual experience with Alexa – there are different ways of engaging with new and existing customers.

Use of algorithms for dynamic pricing, for product design where the AI algorithm can learn from previous designs and embed real-world learning and customer experience to improve product specifications and product experience are some other fantastic ways of using the power of this technology. ?In terms of marketing and media, Gartner reveals that 30% of outbound marketing messages from large organisations will be synthetically generated by 2025.** (Here's how retailers and e-commerce companies use Generative AI (indiaretailing.com)

Early warning system and prevention. A rather less talked about than should application of the power of AI is its potential to serve as an early warning system that makes prediction of future outcomes on the basis of past data patterns. In Healthcare, we often talk about the power of prediction using Machine Learning(ML) and AI to predict diseases and health risks for preventative care. Other use cases are AI-powered prediction of climate-related health risks and prediction of natural disasters. Relevant insurance covers can be built on the basis of data sets such as temperature trends, meteorological events, population density, and flood- and droughtprone zones.

In cybersecurity AI can be used to analyse large datasets to identify patterns that can pre-empt threats of phishing and fraud through abnormal user-behaviour, anomaly detection and through forensic analysis of incidents. Then be it transaction monitoring in financial services to detect potential fraud to monitoring of social media feeds for likely threats to public safety – the analysis of large data sets allows AI to build patterns and make predictions in real-time that can help mitigate risks before they escalate into significant issues.

Which of the above are relevant to your business ecosystem?

I recommend organisations put in place some enablers for successful AI-based organizational adoption and transformation before jumping on the AI bandwagon.

Enablers

Here is my top three!

Getting foundational data right. AI algorithms are trained on data. A good starting point for organisations is to spend time and resources in devising and implementing a data strategy that focuses on the entire lifecycle of how data is ingested into source systems, managed, cleaned and made available for analysis that an AI algo can help deliver. In my dealing with clients across industries I have heard many versions of “we have no data for analytics and AI” to “we have so much data, we don’t know where to start”. Deploying AI is a strategic and cost incurring initiative. Its always sensible to take stock of your organisational data ecosystem. Data could be in legacy systems within underlying databases or in unstructured excels and powerpoint files sitting on staff laptops. These can be consolidated and made available through appropriate Master Data Management(MDM) frameworks. The quality and accuracy of data needs to be driven by operational processes and control mechanism to gather data needed to train the AI algorithms. ?

The insurance and healthcare industry in not just the emerging markets but even the developed markets have borne the brunt of poor data capture and management at source not allowing their incumbents to take advantage of rapidly evolving technologies. Improved data connectivity between partner systems will allow the entire value chain to thrive.

Digital literacy and upskilling. In a number of industries there is very little readiness for the transformation that AI will bring about, not just in the world around us but at the workplace. The Future of Jobs report 2023 by the World Economic Forum identifies roles to do with manual dexterity and precision, roles such as basic programming and those involving reading, writing and mathematics to decline in importance over the next 10 years. AI-based tools are expected to be plentily available to replace those roles in the workplace, in part if not fully.

It is important to note however that there will be an increased demand in higher cognitive skills like advanced computing, critical thinking, advanced literacy and writing skills etc. In wake of this, waiting for the ‘storm to pass’ may not be the ideal approach; building future readiness is going to be key. Improving digital literacy of the workforce and improving their comfort with digital systems through training is one way of preparing for the transformation. Having the right experts to help set appropriate goals and navigate the organization through the AI journey is another.

The responsible application of AI. While AI certainly has the power to alter the future, with power come greater responsibilities. Fair use of the right data feeding into algorithms, removal of bias and protection of user privacy are some key principles of using AI responsibly. ?Organisations could set up an AI governance council that makes policies related to the responsible use of data, providing a framework that puts control measures in place and runs post implementation audits. Such committees could track progress of AI initiatives, monitor and evaluate AI systems, ensure AI deployment doesn’t lose momentum and generates the necessary ROI. ??

So then, how and when would you like to join the ride?

P.S. – I did consider getting generative AI to write this article for me but then I realized it just won't sound like Brahmjyot.