Decision-making driven by the right amount of data and experimentation leads to a greater chance of choosing the right option and making it successful. With the onset of AI, this process has become fast and accessible. We cannot, however, ignore the role of creativity and intuition in the decision-making process. In the times when AI is becoming one of the key technological developments of the decade, how does it incorporate creativity and how does it make the job of decision-making easier for organizations is an interesting question to ponder upon.
Quadrant team consisting of Apoorva Dawalbhakta, Associate Research Director and Sr. content specialist Shinjini Sarkar interacted on this subject with an AI monetization expert Somil Gupta, who is also the Founder & CEO, Algorithmic Scale and AI Influencer of the year.
The conversation started with Shinjini Sarkar, moderator of the discussion, mentioning the recent buzz around the word AI as a service with examples of direct monetization like Einstein, Salesforce’s AI service, and IBM’s Watson. Alternatively, indirect monetization like in Netflix and amazon ensure that there is a helping hand for people and lure customers towards subscription. Given this picture, she asked Somil what AI monetization means, what is its significance in today’s global scenario, and whether there is more to it than what people are commonly aware of.
Somil responded by agreeing that AI is more than just model as a service or a set of recommendations. In fact, monetization is about value realization from AI-generated insights. It impacts basic business economics. Looking back 20 years ago in the era of classical marketing, we used to have surveys consisting of 35-40 questions asked to the public. These data points were used to create some large mass segments in which the companies used to position themselves. Hence, entire business systems today such as ERP and supply chain are designed to serve these few large segments. However, with big data, eCommerce, and AI, we now have thousands of data points giving insights into the behaviour of the users. It helps discover thousands of micro-segments of customers. Companies can classify their customers into these segments, determine their purchasing power, and personalize their offerings. They can also use automation to orchestrate the value chain to serve.
Realizing the Value of AI Monetisation
The real value of AI monetization is the business capability to discover and serve these emerging segments, especially in the times when so many niche players are emerging to serve specific segments. As per Somil, AI monetization should head towards identifying upcoming segments and value pools, developing the ability to serve them using flexible solutions, and realizing value from that. That is how small and medium enterprises with emerging entrepreneurs and solopreneurs can get the best out of AI monetization.
Data Democratization
Building upon this idea of untapped potential lying in mid and small markets, Apoorva mentioned the buzz around data democratization. On one hand, the tech giants are galloping towards exponential technological innovations, while on the other hand, the questions of ethicality and feasibility of data democratization and open-source neutral AI services are understandably growing.
Commenting on this dilemma, Somil pointed out that the focus on data and AI has mostly been on developing the technology, its adoption, and use cases being minuscule in nature. A few people utilizing these tools are power users with specific purposes in their organizations. In contrast to that, one must make the technology and tools accessible to others when it is applied on a larger scale in the entire ecosystem to achieve end-to-end optimization. Hence, while the tech giants are doing well in their operations, they need to realize that they need to open up to the ecosystem so that everyone can use it. As mentioned earlier, the fact that there are millions of micro-segments implies that one player can’t serve them all. Large businesses must empower others (smaller players) to build niche solutions for serving their specific segments. Given this logic, Somil suggested that AI will become a platform play for the larger players and a market play for numerous small market players.
Universal Inclusivity – AI for all
Moving ahead towards the universal inclusivity/AI for all, Apoorva inquired how the AI and related services have fared in these initiatives. In response to this, Somil aptly pointed out that there is a universal component of AI and automation in all the digital products we use today. Whether they are grammar check platforms, emails filtering spam emails, or anti-virus software predicting malware, AI and automation are by default working in the background and people are consuming the benefits of AI. However, when it comes to AI-as-a-service empowering people to build solutions, there is a long way to go. There are two key reasons for this.
Firstly, the complexity of AI platforms makes it difficult for an average person to understand. We have not been able to simplify these models. More inclusion will come only with simplification. And secondly, there is a lack of training to use AI. AI is more about asking better questions than about answering them. Our education systems and rule-based paradigms prevalent in the corporate world do not leave space for asking questions.
Hence, the focus should be on these two factors – simplifying the solutions and empowering people by training them in a new way of thinking i.e., formulating better hypotheses so that AI can share the cognitive load effectively. That is how we should be moving ahead to achieve universal inclusivity.
Creative Decision-making
Next, referring to the data explosion in today’s age with humongous data being generated through every possible source and divided into disparate data sets, Apoorva pointed out that it is a big challenge for analytics and AI to make sense of such immense volumes of data points. He asked Somil how to lead creative decision-making and on a more fundamental level, what creative decision-making means in the context of AI.
Somil replied to this question by firstly mentioning the Pareto 80-20 rule in which an organization operates. It means that the systems we have built are designed for the 20% of cases that happen 80% of the time. This implies that there are 80% of cases not served by the organization, meaning untapped opportunities either because of the lack of awareness or the inability to serve them.
According to him, the business and commercial framework to define a good decision should be set first before deploying AI for finding out the optimum solution. Hence, creative decision-making means first understanding what is not being served and then to think how the business can reorganize its capabilities to serve a new segment. It involves taking a step back from the system/IT oriented to capabilities-oriented thinking. He aptly gave an example of 3M’s weak glue that they branded as post-it notes, reorganizing that capability and orchestrating it to make it commercially successful. AI helps to identify how can a business best utilize its capabilities within the set risk-value-cost framework of the company to create profitability. Deciding the framework is the job of humans (which is fundamentally creative decision-making) and AI optimizes outcomes within this framework.
AI and Human Development
The conversation then moved on with Apoorva inquiring whether one needs to worry about the ill effects of investing too much into AI monetization in terms of human development. Somil responded to it by pointing out that the investment flows more into AI technology as opposed to AI monetization. The questions of simplification, ethics, and risks are difficult to manage in the technological frame of mind. A commercial business model can handle these questions as it is based on relationships and people in the ecosystem. Since AI monetization focuses on value creation through building accessible models and empowering others, there cannot be enough investment in the field. It looks at the impact of AI. He strongly suggested that the investments should flow from technology development to value realization through monetization. This will also drive higher adoption of the solutions.
Adding to it, Apoorva mentioned that there are certain values such as critical thinking and corporate cultural strengthening that humans bring to the table in the context of AI. Regarding that, he asked whether humans should work to become AI-proof/robot-proof to keep their livelihoods intact.
Responding to this question, Somil expressed that there were two facets to it. First, the AI humans deploy today is narrow AI in the sense that it is used to automate hundreds of manually handled tasks. These tasks are part of the bigger picture called jobs. Built for human-level cognition, the jobs are defined in quite an unstructured way. Secondly, once a particular task is automated, we move on to the next level of value creation. For example, the job of executive assistants evolved from note taking 30-40 years ago to manage the calendar and productivity of the executives upon the arrival of more sophisticated automation and productivity tools.
Having said that, Somil expressed his concern that we should fear people with experience, knowledge, and familiarity with AI tools. With their knowledge and tools, these people can use appropriate data to test and validate hypotheses. It gives them higher chances of becoming successful in their experiments. Hence, he strongly recommends investing in learning AI and to incorporate data and AI, an experimental mindset, and more questions into your job. It will give the people the liberty to automate their tasks and move up the productivity ladder by focusing on something better.
Cost of Uncertainty
Next, Apoorva mentioned one of the statements from Somil’s blog that A very simple approach to getting started with monetization is to start defining a return on investment with a simple formula (accumulated benefits – the cost of uncertainty)/total cost of resources. He inquired about confusion as to how organizations define their cost of uncertainty as it becomes quite vague and probabilistic in many situations.
He used the example of a predictive model with 95 percent accuracy to answer this question. The cost of uncertainty is defined by determining the accumulated benefits or additional revenue an organization accrues by making a right prediction using that model minus the cost or loss to the company by making a wrong prediction.To determine these benefits and costs, the organization needs a clear commercial framework.
The Road Ahead…
Moving on to the last question, Apoorva asked how creative and intuitive would data-driven decisions taken by AI services be in the future, especially considering the gamut of technological developments happening in the space. The purpose was to understand where this space heading is towards.
In response to this, Somil pointed out that when it comes to cognitive abilities, AI can be a fruitful companion. However, creativity and intuition are human values at the core. Since AI works within a given framework, the creativity and intuition of decisions taken by AI are going to be limited to that only. AI can find out new combinations of outcomes within the set framework, giving more creative solutions. However, it does not operate beyond the set rules and limits of the framework.
Hence, Somil believes that AI will have an upper hand within a fixed frame and humans will get an upper hand when the frame changes. This presents a leeway for constant collaboration between humans and AI as the frames will keep on varying with ever-changing conditions in the market and the world. Humans have been tackling most issues as trade-off problems, where we must sacrifice one output to achieve more of another option. For example, we to achieve technological development, we must sacrifice the environment in terms of pollution. AI will enable us to move away from this trade-off approach to more cohesive solutions, where we can achieve both outputs with the help of creative solutions. We can achieve sustainability and technological development together. With this new strength at hand, we will keep on updating the frame to figure out better-optimized solutions to many problems of climate change, wastage, sustainability, etc. AI will be our true companion in this journey.
Author
Vaishnavi Dave is a Content Writer at Quadrant Knowledge Solutions.
