Data. AI. Abracadabra.

Reading Time: 5 minutes

AI inspires loyalty and loathing alike, in good measure. Implementing AI at scale is challenging, for sure. But scaling teams, processes and tools in an integrated, cohesive manner offsets the setbacks and sets one up for a seamless transformation experience. How do businesses today view this tectonic shift led by AI? What are their considerations in charting the way forward? At a panel discussion convened by Vue.ai, at the REBUILD ’23 Jamaica retreat, enterprise leaders revealed what it is about the technology that keeps them awake and metrics to sleep on, while considering AI partners to collaborate with.

AI inspires loyalty and loathing alike, in good measure. Implementing AI at scale is challenging, for sure. But scaling teams, processes and tools in an integrated, cohesive manner offsets the setbacks and sets one up for a seamless transformation experience. How do businesses today view this tectonic shift led by AI? What are their considerations in charting the way forward? At a panel discussion convened by Vue.ai, at the REBUILD ’23 Jamaica retreat, enterprise leaders revealed what it is about the technology that keeps them awake and metrics to sleep on, while considering AI partners to collaborate with.

First you wrestle, eventually you embrace

Most change management plans work this way and AI transformation is no different. The panelists discussed the on-ground logistics of periodically surveying their world around for signals to act on, in the specific context of AI adoption. In both product- and services-oriented companies, Product Managers act as sentinels, sensitizing the management on the evolving tech scenarios and on the changes coming round the corner. Taking cognizance of the rapidly evolving landscapes, managements too have added ears to the ground to evaluate the availability of new-age opportunities and increase the top line. Companies such as Citigroup closely monitor the trends in music and app downloads for insights on user preferences. And certainly, a crucial part of change management is to have honest communication amongst the stakeholders, regarding the implications of AI and the new opportunities it could usher in.

Data dilemmas in companyland

AI can lay the groundwork; but human intervention is still necessary to eliminate bias. That’s the basis of laws, such as the one promulgated in April 2023 by the state of New York, that requires human audits of AI-generated work, said Amit Arora. Amidst the rising reliance on AI algorithms to achieve greater business efficiencies, prioritizing their nuanced deployment is paramount for organizations, declared Amit. Sectors such as healthcare and insurance carry a very low margin for error, operating within highly regulated environments and thus requiring greater oversight to ensure safe and reliable decision outputs for end users. On the other hand, industries such as retail and telecom have more flexibility in AI implementation.

For the effectiveness of the audits, Carol Grunberg proposed audits by external members – conducted by individuals unfamiliar with the industry’s inner workings – to help uncover biases and ensure fair AI practices.

Machine-mandated dress codes and other AI implementation woes

When the Board asks about the AI roadmap, the C-Suite feels nudged to answer in the affirmative. Lack of clarity in expectations amongst the leadership teams leads to communication gaps, hindering the application of right methodologies towards AI, lamented Amit Arora. Environmental, Social and Governance (ESG) issues related to the AI rollout are also increasingly dear to organizations, he pointed out. Convincing the management of one use case at a time by demonstrating the financial benefits of adopting AI was Dr.Karthik’s prescription to increasing the technology’s footprint across various functions. Should the AI technology be developed in-house or in collaboration with a partner? Or should it be outsourced completely? Where would the IP ownership rest? (see related #POVue) Is there an opportunity to white-label the developed technology for other organizations to deploy and gain incremental revenues? A few of the points to ponder over as panelists discussed the challenges organizations face in adopting AI and the factors that hinder its large-scale deployment. 

Being a critical aspect of both product- and services-led organizations, customer experience is also a domain that lends itself well to leverage automation for innovation. Acknowledging the opportunities AI presents to engage, educate and excel in customer delight, Washington voiced the need to strengthen AI in the two key areas of data bias and ethics. 

Enterprises and AI – Is there love in between?

Citing proven use cases in tele-ophthalmology to diagnose Diabetic Retinopathy, in tele-radiology to draw automated interpretations and insights from X-Ray, CT and MRI images, in the tele-diagnosis and differentiation of COVID vs non-COVID pneumonia and of cardiac attack vs gastritis, Dr.Karthik expanded on the utility of AI in healthcare. Personalization of insurance premia based on individualized risk assessments and accelerating drug discovery were other areas where AI’s impact in healthcare has been demonstrated.

Juxtapose this against a global group such as Citi that faces a different set of challenges – internal and external. Aggregation of customer records from across a multitude of touchpoints, developing unified user journeys, compiling a golden source of user profiles and sharing the information across the entire organization, spread over wide geographies is a challenge that AI could help solve, revealed Carol Grunberg. The mixed nature of client data, including its quality and structured-vs-non-structured-ness is a challenge for AI to measure up to. Demonstrating the ability to learn from data and adapting quickly is critical to advancing disruption and generating value across the chain, emphasized Washington Drumond, explaining his organization’s priorities. 

Although AI can auto-generate medical prescriptions at scale, a human doctor’s intervention is required by law for validation. Dr. Karthik Anantharaman stressed the importance of AI getting up to speed in order to facilitate at-scale deployment while minimizing human interventions in healthcare. The see-saw balance between disruption/innovation unleashed by AI on the one hand and data governance mandates on the other is a play to be well rehearsed, if companies are to effectively use AI to improve efficiencies, per Washington Drumond.

The session was moderated by Shelly who explained the utility of AI in a spectrum of use cases right from automobiles (Tesla) to clinical analysis (Cure.ai). Valuable takeaways on the whole from the discussion, as business leaders geared up to drop anchor in these unchartered waters.