The shifting priority of AI to excellence-led monetization

As much as enterprises are increasing investments in AI, there is a real and persistent struggle to demonstrate business value from AI and realize returns on investment

The shifting priority of AI to excellence-led monetization - CIO&Leader

As much as enterprises are increasing investments in AI, there is a real and persistent struggle to demonstrate business value from AI and realize returns on investment. Centers of Excellence (CoE) or innovation centers for next-generation technologies have played a significant role in creating thought leadership and bringing about best practices, but there’s more to be done in a planned and concerted manner to leverage and monetize AI to its fullest and as such, CoEs have to show more action and outcome bias and become Centers of Excellence and Monetization (CoEM), as we would like to term it.

A CoEM is a natural extension and enhancement of the CoE. It consists of all the fundamental building blocks of a CoE and takes it a notch further to bring about the monetization bias. A CoEM is designed and built on the same four fundamental pillars as the CoE, and those being:

  1. Strategy
  2. People
  3. Processes; and
  4. Technology

Strategy

Businesses today are looking to use AI to find answers to their complex problem statements while progressing in their digital transformation journey. A data-driven strategy and decision-making framework become an important aspect in this journey before we can actually get talking about monetizing AI. The strategy function of the CoEM is one of the most important pillars. It defines the vision and direction of what the enterprise can practically and realistically achieve so that morale and confidence in AI initiatives remain positive.

The action and monetization bias comes from the ability of the CoEM to identify and prioritize high-impact use cases where investment needs to be made. This is easier said than done. It often happens that organizations bite off more than they can chew when it comes to finalizing a charter for their AI journey. Choosing the right use cases to work on is a balancing act. As much as the use case(s) needs to be able to positively impact either top line, bottom line, or customer experience, it is also pertinent that organizations pick ones that are realistic to execute. Initial successes, however small, significantly boost the confidence and morale of the teams and leadership in equal measure. Ensure that all small and incremental successes are published and celebrated. As much as one would like to do everything on their own, the key to AI success is in creating an ecosystem of partners and academia who can jointly add value to the initiatives.

People

How should the AI CoEM look like? Should all AI competencies be centralized resources or decentralized in multiple departments? What is the role and responsibility of each identified stakeholder in the ecosystem? What are their interdependencies? How can an enterprise keep them motivated and continuously curious? How can we put in place a mechanism to add new stakeholders with the right skills and temperament to ensure sync with the current teams and objectives? These are all critical questions about the people strategy aspect of a CoEM.

Once the answers to these higher-level questions are built into the design, ground-level challenges also needed to be addressed, and an important one being - Identifying the right data owners and stewards. This is a highly underrated aspect and raises its ugly head whenever clean, and tightly governed data is needed to push through initiatives. Organizations struggle to make this available since data is stored in multiple places, and due to a lack of proper governance, enabling democratized access is impossible. These cannot be solved just by tooling but needs change management in the people process and mindset.

Process

The next important building block of a CoEM is the process element. More often than not, organizations attempt to fix many problems by throwing more technology and automation at it, and they fail. This is because the underlying process itself is inherently broken. Therefore, if you already have an AI CoE in some shape and form but not yielding the results and outcome that you were expecting, it might be worthwhile to have a serious relook at existing processes, and if needed, ask fundamental questions on why the organization does things a certain way. When it comes to CoEM, processes lay down a systematic way in which stakeholders who are directly or indirectly associated with the initiative can collaborate better, create more re-usable assets and ensure that all the knowledge, assets, and best practices are stored securely and made accessible on-demand to authenticated and authorized users. The monetization bias comes from the repository of re-usable assets, feature stores, and libraries which allow the culture of ‘fail fast, succeed fast’.

Technology

The technology landscape of AI, like many other technology domains, is an ever-evolving one. The high rate of obsolescence of approaches, methodologies, and concepts pose a constant challenge to an organization’s ability to be ahead of the curve and putting the investments in the right basket. The technology strategy within the CoEM should include, as part of its design, the framework for tools and technology evaluation, best practices for architecture standardization, and managing operations of AI solutions in production.

The beauty of AI is that the solutions powered by it allow a level of flexibility and self-optimization. Additionally, the technology also provide enriching outcomes by adding more data dimensions and features. The choice of tools, technology, and framework ensure that the AI solutions remain living, breathing entities and adapt to the evolution of the business.

The authors are Sandeep Sudarshan who heads the Business & Solutions Consulting Group in Subex and Arundeep Sivaraj who is a Director in Subex’s Business & Solutions Consulting Group


Add new comment