Many companies are considering how to make artificial intelligence (AI) real for their organizations. Breakthroughs with the use of artificial intelligence are constantly in the news. Healthcare companies are using information driven from AI initiatives to initiate or accelerate research and development efforts on testing, treatment, and infection control. Organizations in financial services use the findings to streamline and optimize processes like criteria for credit decisions and real-time accurate predictions and detailed forecasts for a particular market or environment.
AI drives strategic decision making and significant value creation for the marketplace, and an adoption of the new paradigm of management.1
The 3 A’s of AI
The common mistake organizations make, is to think of AI as just another technology or just another technology innovation. AI is so much more. It is an innovative focus on the future rather than the past or the present. It, literally, allows the manager to plan for the future by inventing the solutions to take advantage of it. Three critical capabilities need to come together to make this possible. There are what the 3 A’s: analytics, algorithms, and automation, not necessarily in that order.
- Analytics - Analytics facilitates the exploration, exploitation and extraction of insights about past and current actions, occurrences and events. Analytics also allows us to extend these insights about the past and the present into predictions of the future. This casts past and present trends in a future context to understand what the future holds for those actions, occurrences, and events. The processes needed to define the applicable future state scenarios require extensions of the data models with semantic, symbolic, and dimensional attributes or features2, hence the need to transform analytics into algorithms.
- Algorithms - Transforming analytics into algorithms allows dynamic modification of the parameters of the process. Algorithms also test the results to learn what transformations of input produce plausible and explainable outcomes. Indeed, machine learning and deep learning algorithmically discover transformations of input features to a target decision dimension of the domain of interest, whether it is the probability of efficiency of a drug, probability of adverse pharmacovigilance events, or binary outcome for a business decision. In many cases, it is necessary to make algorithms work together to expedite the intake of the data, processing of the data, and execution of the actions uncovered without a man-in-the-middle. This capability is provided by automation, the final A of the triple-A acronym.
- Automation - Automation does not just make it possible for autonomous execution of algorithms but is used to enable B-STEAM processes (business, science, technology, engineering, arts, and mathematical processes). Basic automation enables the automation of data processes, intermediate automation is good for rules and simple processes with few steps without decision points, advanced automation does well with complex processes with fixed programmable decision points, robotic automation is perfect for rapidly repeating actions with fixed programmable decision points, and intelligent automation is great for very complex processes with complex dynamic decision points. Automation generates alerts, notices, and commands that can be channeled to operationalize complex management planning and business decision making, to operate a robot, enable autonomous / driverless cars, make a home smart, and so on and so forth. Within the business context, these are considered AI-enabled capabilities, though it may not always pass the Turing test of intelligence.
The path to AI enablement
Now, we know the critical capabilities needed for AI enablement, it is as simple as defining a program that progressively matures the 3 As of AI within the organization. Doing so requires four critical management activities: establishing a shared agenda, creating a capability enablement strategy, creating an execution roadmap to realize the strategy, and initiating a set of demonstration projects to showcase the enterprise viability of the agenda to corporate leadership, be they public or private. This sequence of actions sets the path for the AI enablement program with an AI capability roadmap. Although this is a common sequence for building capability roadmaps, the AI capability enablement roadmap puts specific demands and considerations as follows:
- Define agenda beyond the traditional planning scope and horizon: The agenda defined for the AI capability enablement program should not focus on the 3-5-year strategic plan that organizations typically put in place to support their growth, product/service development, talent management, and technology investments. The agenda should also include an assessment of innovations within the operational space of the organization, especially, those that could change the market dynamics in their favor as well as against them. This agenda should be unconstrained by the current situation of the organization but looks at the leaders in the space and envisions solutions that create the runway to leapfrog the leaders in the space. The outcome of this exercise is a list of items for the achievement of the agenda
- Establish a capability strategy that matches the innovation agenda: AI capability enablement strategy connects the current situation in the organization to ideas of a future state to transform the organization in the direction of the established enterprise agenda. The current state assessment raises pain points or constraints, as well as the necessities to address these constraints. For the AI enablement capability strategy, the 3 A’s discussed above are very useful in establishing the state of the people, processes and technology, the level of maturity and gap to be addressed.
- Design the roadmap that offers clarity in the transformation of the people, process, and processors (technology) connected to the agenda: AI capability roadmaps are difficult to create because several of the items from the agenda and the strategy may not be considered actionable. Typical gaps like the employee skill gaps, ad-hoc intake process for analytics, manual processes, monolithic transactional systems, absence of decision benchmarks and thresholds, etc. require creative design thinking to appropriately account for them in the roadmap. The roadmap needs to tell the story of the enablement of enterprise resources (people, process and processors) including pictures of the intermediate states towards the defined agenda.
- Identify demonstrations to showcase the agenda to leadership for strategic and financial support: It is rare for the entire organization to come to the realization at the same time for AI capability enablement. Typically, parts of the business that are struggling to achieve target performance would most likely spearhead the drive to AI. So far, it has been marketing in an attempt to take advantage of programmatic online and demand marketing opportunities, product / service development while planning for the next logical product / service, etc. Some organizations have formed teams to focus on the 3 A’s. Demonstrations that showcase the integration of these capabilities to deliver the expectations of AI make getting the strategic support and the funding discussions much easier.
What does it mean for your organization?
AI is a core capability, therefore, the investments made hold huge promise for return on investment (ROI) of over 300% and almost infinite revenue from artificial intelligence (RAI)3. Making sure that the AI journey is deliberate and progressive prevents the organization from spending on solutions to toy-problems lacking value or contribution to the future of the enterprise. Establishing AI enablement in terms of 3 A’s should ensure the journey is tied to the actionable and valuable transformation of the 3 P’s of enterprise execution abilities: people, processes, and processors (technology). A well thought out AI capability roadmap is the most important first step in this journey of many years.
- Towards a Dianoetic Paradigm of Management with Analytics, Automation, and Artificial Intelligence. Refractive Thinker Vol XVI: Generations: Strategies for Managing Generations in the Workforce. Amazon.com.
- Analytic Extensions to the Data Model for Management Analytics and Decision Support in the Big Data Environment. Semanticscholar.com.
- From ROI to RAI (revenue from Artificial Intelligence). Forbes.com.