Generative AI Applications: The Opportunity to Seize - Bishal Dasgupta, MBA ‘24

Greetings, I am Bishal Dasgupta, a second-year MBA candidate at the Joseph M. Katz Graduate School of Business and a Digital Transformation Consultant. The focus of this article revolves around Generative AI Applications: The Opportunity to Seize.

This article delves into essential concepts, emphasizing the potential for value creation in this swiftly expanding technological sphere. It explores the opportunities available to entrepreneurs who can harness the power of Generative AI by developing applications tailored for various use cases.

This piece serves as the inaugural installment in a series of articles I intend to publish on my LinkedIn account. These writings will culminate in my comprehensive MBA Capstone Report/Article.

Generative AI belongs to a category of artificial intelligence used for tasks such as generation, summarization, and classification. Its capabilities encompass a wide array of functions, including data reorganization, classification, editing, summarization, question answering, and content creation, among others. Notably, numerical and optimization applications serve as the primary value drivers for other types of AI.

While most models generate content in a singular format, large language models possess the capability to process and analyze multimodal content, spanning text, audio, images, and videos.

Generative AI Value Chain and Opportunity for New Entrants:

Indeed, certain startups, including notable names like Cohere and Anthropic, have achieved remarkable success by constructing their own models.

It is crucial to highlight the immense potential inherent in fine-tuning existing models. Many organizations stand to benefit significantly from fine-tuning pre-existing models by providing them with pertinent data and adjusting their parameters. This approach has proven to be an outstanding alternative to the labor-intensive process of building models from scratch.

Generative AI, with its transformative capabilities, holds the promise of creating value beyond the scope of existing AI and analytics. It stands poised to enhance labor productivity, reduce costs, and unlock revenue-generating opportunities. The nature of the use cases and the value they unlock suggests that certain industries and functions are expected to derive comparatively greater benefits. The following is a curated selection of industries and functions poised to harness the most substantial potential:

As part of the MBA Capstone project, the objective is to meticulously outline the business and technical implementation strategy for the chosen use cases. However, presented below are some exemplary use cases categorized by function. It's important to note that this list is not exhaustive.

Transformation of Customer Operations:

• Customer Self-Service: Enabling interactions with a chatbot equipped with contextual information about customer queries.
• Customer-Agent Interaction: Empowering agents to access real-time customer information and predict query resolutions, enhancing query resolution efficiency.
• Agent Self-Improvement: Providing agents with updated client information, next best action/offer recommendations, scripts, and conversation summarization tools for continuous improvement.
  
  

Revolutionizing Marketing and Sales:

• Strategic Insights: Allowing marketers to identify trends and gain insights through quick cohort analysis and analysis of unstructured data from disparate sources.
• Content Creation: Facilitating the creation of multiple content versions using natural language prompts.
• Conversion Enhancement: Virtual bots mimicking human-like qualities, building trust and rapport with customers.
• Customer Retention: Increasing customer retention rates through personalized, relevant communications and targeted upselling strategies.
  
  

Innovations in Software Engineering and Tech Product Development:

• Data Management: Leveraging Generative AI to label, classify, and clean vast volumes of data efficiently.
• Architectural Designs: Creating diverse architecture designs and configurations tailored to specific project requirements.
• Code Assistance: Empowering engineers to utilize Generative AI for code writing and refactoring, streamlining the development process.
  
  

These use cases represent only a fraction of the potential applications of Generative AI technology, showcasing its versatility and transformative impact across various functions and industries.

In addition, there are also various applications of Generative AI across modalities:

The following section provides an extensive overview of critical considerations essential for implementing Generative AI, as seen through the lenses of both a Strategy Consultant and a Product Manager. In my forthcoming articles, I plan to delve deeply into each of these strategies. Here, I aim to introduce these elements, many of which are well-established concepts in the field.

1. Value-Centric Approach:
   • Prioritize value as the guiding principle in decision-making processes.
2. Strategic Ownership:
   • Take charge of the vision, timeline, goals, and priorities to ensure alignment with organizational objectives.
3. Use Case Design and Prioritization:
   • Develop a use case and prioritization matrix, focusing on Revenue, Cost, and Growth factors.
   • Select key use cases based on metrics that drive financial value.
4. Decision-Making Framework:
   • Evaluate the subscribe/buy/build decision, considering options such as utilizing publicly available open-source models and fine-tuning models with open-source data.
5. Change Management and Innovation Culture:
   • Foster an innovation culture by nurturing core competencies and exploring adjacent opportunities.
   • Upskill the existing workforce to adapt to the evolving technological landscape.
6. Ethical Considerations:
   • Implement moderation models and control temperature to manage the randomness of output and mitigate hallucination risks.
   • Define data tagging protocols, especially in contexts related to high-security compliance needs, deciding whether to bring the model to data or data to the model.
7. Data Strategy Evaluation:
   • Adopt existing services or build customized models on open source platforms.
   • Upgrade enterprise technology architecture to seamlessly integrate and manage Generative AI solutions.
8. Product-Platform Implementation:
   • Define a robust architecture for seamless implementation.
   • Establish a centralized cross-functional Generative AI team.
   • Embrace an agile operating model, encouraging early failures and rapid scaling.
   • Reimagine the technology function by improving software development, assisting Generative AI in reducing technical debt, and automating IT operations.
9. Strategic Tech Partnerships:
   • Collaborate with the right technology support partners, focusing on selecting suitable models and implementation partners with aligned goals and expertise.

In subsequent articles, I will elaborate on each of these strategies, providing in-depth insights into their implementation nuances and their impact on the successful integration of Generative AI within organizations.

CONNECT WITH BISHAL ON LINKEDIN