Diving Into Digital Transformation With Deloitte Consulting’s Managing Director of Applied AI

Data Basics, Scaling AI Joy Looney

As the competitor and consumer spaces shift to adapt to emerging innovations, companies have developed a thirst for machine learning (ML) investment, hoping to apply ML as a prime AI enabler. Organizations are jumping to fill their digital canvases, but many are struggling to define exactly what that looks like and approach the journey with clarity. 

In this EGG On Air episode, Ayan Bhattacharya, Managing Director of Applied AI at Deloitte Consulting, provides us with a unique perspective on the process of fueling enterprise digital transformation across various industries. He also addresses the main challenges and benefits of transformation and sheds light on the variation of organizations’ AI maturities. Bhattacharya helps us understand the ultimate vision and steps involved in the journey to transformation including forethought, careful application, and long-term perspective.

→ Watch The Full Episode Now 

Embarking on the Transformation Journey and Understanding the Scope of AI 

Some benefits of AI in the enterprise are fairly obvious, such as the ability to reduce time and investment into heavy research efforts; other benefits are not quite as pronounced, such as the ability to apply consumable and interactive data interfaces into business applications. Bhattacharya is a proponent of technology that looks at the wider ecosystem but comes back full circle to the people it can serve. He reminds us of the surprises in the innovation curve, from how phones have transformed into a kind of assistive technology to how ML is being applied in both government and the private sector. Structured data is infiltrating a breadth of organizations, from legal rights to digital healthcare services, but finding a way to harvest that data and then transform it to become something usable means embarking on a journey. 

The ability to achieve a data-integrated environment for your business means having the ability to identify the organizational needs upfront in order to appropriately invest in the right data talent. Correct resources must be identified and systems left to steep long enough for scientists to properly understand business challenges and adequately interface with the data long enough to extract the value that organizations are counting on.

Bhattacharya reminds us that, as with any journey, the transformation journey should begin with a well-plotted course. Taking the time to have conversations with chief data officers about broader implications on a cost basis is a key part of the process. Bhattacharya recommends asking questions like: “Can this architecture withstand a surge in transactions?” and “Are proper redundancy options laid out if the model is not available?”

The data landscape is a moving canvas; Bhattacharya compares this environment to a sandbox. As organic efforts meet partnerships with emerging technology players in the space, innovation forces companies to evolve. During this tumultuous but opportune time, companies can act as if in a sandbox, playing around with AI tools to see what interacts in the best way with their changing technology ecosystems. 

Variance in AI Maturity and Other Factors to Consider

Organizations’ maturity as it relates to data, digital prowess, and AI/ML is highly varied, so AI maturity is a malleable measurement dependent on how exactly an organization is choosing to utilize AI. In order to define a starting place for assessing maturity, it helps to have clarification on what exactly defines maturity. Bhattacharya views AI maturity as an organization’s ability to construct solutions using AI and drive the adoption of solutions across the organization. 

This coveted pervasiveness extends throughout the entire organization including customer interactions, employees’ engagement with AI, and relations with partners. The concept of AI maturity looks at core AI capabilities (ML, natural language processing, vision AI, etc.) and, depending on the particular realm of those capabilities that an organization is charting a path for, the maturity scale varies. Regardless of the realms or level of maturity, each organization will need to address roadblocks and pressures along the journey.

Episode - Home Page - Ayan-High-Quality

Thriving in the Era of Pervasive AI 

Deloitte’s “Data in Motion” concept is described as looking at AI from the lens of how data is being enriched and used along the way, extracting the most value from that data through insights for things such as commercialization. The ultimate vision is to “utilize data to make our lives easier” in all areas of an organization and, in the long term even drive change on a global scale. In order to thrive as an organization, we can’t ignore that, relatively speaking, everything is consumed as a service. “Data in Motion” accounts for the fact that, in these environments, there is a cost associated with capturing the data, staging the data, and executing the whole data pipeline. While boundless exploration is highly encouraged at the beginning of the AI journey, Bhattacharya asserts that there needs to be clearly defined governance and easily accessible defense mechanisms put into place for both desired transparency as well as data privacy. These precautions are necessary in order to avoid compromises in security.

Distributing Resources and Effectively Scaling  

Bhattacharya reveals that, in his experience, most clients can be placed into one of two categories: early adopters or fast followers. What he has observed is that, no matter the category, there is weakness in the connection between data science approaches and integration in practice. Realizing that modifications are at times necessary in the deployment of models and learning from others with successful implementations will help organizations to scale with long-term goals in mind. Proper model deployment in order to achieve a fully productive and collaborative model that reaches downstream business implications.  

The digital transformation journey involves incorporating automation, operation enhancement, and new offerings into your organization; distributing your resources in the most efficient way possible across the three should be your goal in order to effectively scale. Whether ML applications are pulling historical data or new datasets, either way, the long-term challenge is articulating a design for data ingestion that makes sense with your particular business dynamics and then effectively using resources at your disposal. Third-party data insights, as well as insights from enriched, internal data, can be employed in the digital transformation process.  

 Much of the appeal of ML has always been the exponential value curve and, regardless of external turbulence, the companies that continue to support innovation, in spite of challenges, are the ones that we observe scaling effectively, getting value back from their investments, and prospering further down the timeline. 

You May Also Like

Fine-Tuning a Model (In Plain English!)

Read More

Talking AI Democratization With Dr. Anastassia Lauterbach

Read More

6 Top-of-Mind Topics About AI & Trust in 2024

Read More

3 Concrete Ways to Drive AI ROI

Read More