From drug discovery to patient interactions, the number of AI use cases in pharma is large and spans the entire value chain. As a result, leaders in pharmaceutical companies find themselves flooded with suggested AI projects from different departments across the organization. A common scenario we have observed at Intellishore is that leaders encourage business units to individually pursue AI use cases across the board to ensure the organization is “doing something with AI.” Although this approach can be tempting to avoid falling behind industry peers, it often ends up resulting in issues further down the road, such as a lack of resources, poor adoption, or an inability to scale successful solutions across the organization. Therefore, having a structured approach to selecting the right use cases to pursue for your organization is, in our experience, crucial to your AI success.

As a leader in the industry, navigating this selection process can be challenging and poses several questions: How do you estimate the value of an untested AI use case? And how do you know whether the organization is ready to support and, eventually, utilize the use case in everyday tasks?

This article is the second in a series of four, where we provide our experience and perspective on how companies in the pharmaceutical industry can strategically implement AI solutions to maximize its impact and value. The focus of this article is to introduce our conceptual framework for selecting the right use cases for your organization and creating a guiding roadmap for implementation.

Click here to read article 1 in the series

Intellishore’s High-Level Approach for Selecting AI Use Cases

Starting with the Strategy

We have seen many AI initiatives fail to generate real value because they address minor operational hurdles rather than focusing on the high-level strategic goals of the company. For instance, casually implementing a GPT-based chatbot, as many companies do these days, will not necessarily solve any of your core business problems if not done purposefully. As a result, we always recommend that clients start with the value proposition of the company as well as the corporate strategy and goals therein. If organizational goals include reaching more patients and healthcare professionals (HCPs), growing an existing therapy area, or expanding to new therapy areas, your AI efforts should target these objectives specifically.

As a second step, we recommend aligning on a set of guiding AI principles before starting to prioritize among use cases. These principles usually include considerations of where to position on various spectra.

• Do we target short- or long-term impact?
• Do we want to automate processes (hands-off) or enable human decision-making (hands-on)?
• Do we adopt a wide or narrow initial focus—i.e., do we target initiatives across the entire value chain or start within, e.g., commercial?

The prioritization of use cases should reflect both the identified strategic objectives and guiding AI principles.

Defining the Evaluation Criteria

A well-known approach for evaluating organizational initiatives against each other is scoring them based on value and feasibility. However, evaluating value and feasibility objectively can be difficult. The first step entails defining a small set of criteria to estimate both value and feasibility. We consider it helpful to consider three high-level categories for both value and feasibility criteria.

Establishing Value Criteria

To select the right AI use cases, it is imperative to take a holistic approach and avoid focusing solely on initiatives addressing immediate operational challenges or opportunities. Instead, our experience suggests focusing on the company’s long-term vision and competitive edge. To evaluate an initiative holistically, decision-makers ought to consider strategic priorities, the potential for revenue increase, and the possibility of efficiency gains and cost savings. Therefore, we recommend the following value criteria:

• Strategic Fit: The strategic goals vary based on the specific company, but if it is a priority to, e.g., improve the HCP experience, AI initiatives that facilitate this hold a higher value. Example: Implementing AI to personalize communication with HCPs can enhance engagement and satisfaction, aligning with strategic goals of improved HCP experience.
• Top-line Performance: If initiatives through, e.g., enhanced decision-making, are likely to result in an increase in revenue, they hold a higher value.
  Example: AI-driven market analysis to identify new revenue opportunities in emerging markets can directly boost top-line performance.
• Costs & Efficiency: If an initiative enables business processes and results in a significant decrease in costs or time spent on specific tasks, it holds higher value.
  Example: AI-enabled automation of routine compliance tasks can reduce operational costs and improve efficiency.

Establishing Feasibility Criteria

Successfully implementing AI initiatives is a complex task relying on a mature foundation – both technically and organizationally. Therefore, we recommend considering the following factors when evaluating whether a use case is likely to be developed and adopted successfully:

• Data Availability: If the data necessary to support the AI initiative is readily available and is both well-structured and well-managed, the initiative has higher feasibility.
  Example: A drug discovery AI project will be more feasible if extensive and clean historical clinical trial data is available.
• Technical Capabilities: If the organization’s technical landscape and capabilities are advanced and compatible with the AI initiative, it has higher feasibility.
  Example: An AI-driven predictive analytics tool for clinical trials requires a robust IT infrastructure and skilled data scientists.
• Organizational Maturity: If the necessary competencies are available and the initiative does not require a significant change in ways of working, it has higher feasibility.
  Example: Implementing AI for optimizing supply chain logistics is more feasible in a company with mature digital and data governance practices – among users & developers.

Summary – Intellishore’s High-Level Value & Feasibility Criteria

Prioritizing Use Cases

Once tailored criteria have been defined, the identified AI initiatives need to be individually scored based on each criterion. With one of our clients, we recently established a central governance board tasked with retrieving business cases for AI use cases from across the organization. To ensure comparability, we created a use case intake template that can be used to describe an AI initiative as well as its related benefits and challenges in a structured and uniform way.

Based on the drafted business cases, it is the task of the central governing unit to score each initiative within each of the defined value- and feasibility criteria. The governing unit is to make this assessment as objective as possible across all use cases. This process elicits total scores for both value and feasibility, which informs the prioritization of initiatives into four different buckets, determining how the use case is treated moving forward, as illustrated below.

The prioritization of use cases helps organizations establish a clearly defined AI vision to work towards. However, to develop an actionable plan for how to reach the goal, it is imperative to identify gaps between the maturity of the current setup and the necessary maturity to support the AI vision as well as outline initiatives that will bridge these gaps. Identifying the right AI-enabling initiatives to prepare the organizational & technical foundation to support your AI vision is the focus of the next article in the series.