Finding the right use cases

Overview

Identifying the right use cases is one of the most critical challenges when adopting AI-driven analytics.

This guide presents a structured, impact-oriented framework to help you uncover high-value opportunities that align with your organisation's goals and are feasible with your current data.

The approach is based on our experience consulting for Fortune 500 companies and has been designed to support cross-functional teams across operations, marketing, analytics, product, and executive leadership.

By the end of this guide, you’ll be able to:

• Align analytics initiatives to your strategic objectives
• Map your core processes and uncover bottlenecks
• Match AI capabilities to real-world decision workflows
• Score and prioritise opportunities based on impact and feasibility

The framework is based on the following steps:

1. Define business objectives
2. Identify key processes
3. Review AI capabilities
4. Match capabilities to subprocesses
5. Score opportunities

Define business objectives

Effective analytics starts with a clear strategic goal. Every use case should tie directly back to an organisational objective. This is how you justify investment and measure impact.

Start by identifying 1–2 primary objectives. These may include increasing shareholder value, improving customer retention, reducing operational costs, or enhancing product experience.

Each opportunity will then be judged by a simple question derived from your objective.

Note: Example 1

Objective: Increase shareholder value → “How much will this improve shareholder value?”

Note: Example 2

Objective: Improve marketing efficiency → “How will this improve the ROAS of digital campaigns?”

Identify key processes

Next, identify the key processes that support your objectives. These are typically end-to-end workflows—such as marketing funnels, sales pipelines, customer support flows, or internal production chains.

Focus your effort based on the scope of the objective:

• Organisation-level objectives: Map the entire value chain, from customer acquisition to delivery and support.
• Product or service innovation: Focus on how users interact with your offering—via Jobs-To-Be-Done or service delivery steps.

Note: Example 1
A manufacturing company aiming to improve shareholder value might map:

• Marketing → Sales → Delivery → Support
• Internal flow: Production → Quality Control → Supply Chain

Note: Example 2
A software company aiming to improve its product might map:

• User onboarding → Core workflows → Feedback loops

Once identified, translate the key processes into process maps ensuring you include:

• Inputs and outputs
• Stakeholders
• Process steps and decisions
• Tools and data sources

Conduct this collaboratively with domain experts, data professionals, and managers. Ask each group to prepare a process map from their perspective, then synthesise them into a single view. Here is a high-level example for a fictional insurance brokerage:

Note: Example: Insurance Brokerage

Process	Decision	Inputs	Output	Stakeholders
Marketing	Segment customers	CRM and purchase data	Target segments	Marketing team
Marketing	Select messages and channels	Segment and campaign data	Channel recommendations	Marketing team
Sales	Score leads	Lead data and past performance	Lead scores	Sales team
Sales	Engage and convert	Lead scores and account context	Outreach priorities	Sales team
Renewals	Identify customers due for renewal	Customer and renewal data	Renewal risk list	Renewals team
Renewals	Proactively engage customers	Renewal risk, account history, and playbooks	Engagement priorities	Renewals team

For each decision, capture the current process steps, the systems involved, and the evidence people use to decide whether the outcome was good.

Review AI capabilities

Before identifying solutions, ensure everyone is aware of the current capabilities of AI-driven analytics.

Capabilities often fall into the following categories:

• Event Prediction
• Classification
• Granular Forecasting
• Segmentation
• Search
• Writing

Refer to the Capabilities page for a full overview.

Match capabilities to subprocesses

For each process, choose a few key decisions within the process and simulate how a human would solve them.

Ask: What information do they use? What steps do they take? How do they assess quality? This helps uncover the logic and data dependencies behind decision-making—and prepares you to match AI capabilities.

Once you've identified the key decisions, match them to the appropriate AI capabilities.

For the insurance example, the processes might map to the following capabilities:

Note: Example: Insurance Brokerage

1. Marketing
   • Segment customers = Data cleaning + segmentation
   • Select messages/channels = Data cleaning + classification
2. Sales
   • Score leads = Data cleaning + event prediction
   • Engage and convert = Data cleaning + event prediction + writing
3. Renewals
   • Identify customers due for renewal = Data cleaning + event prediction
   • Proactively engage = Data cleaning + event prediction + writing

Score opportunities

Now assess the use cases you have identified for both impact and feasibility.

Impact

• How strongly does the opportunity support the original business objective?

Feasibility

• Is it achievable with current AI capabilities?
• Is the required data available?
• Can it be delivered within timeline and budget constraints?
• Will stakeholders support the initiative?

Tip: Scoring isn’t about precision—it’s about consistent comparison. Use this process to structure conversations and prioritise action, not to perfect a mathematical model.

Next Steps

• Run a pilot workshop with one team or department
• Select one high-feasibility, high-impact use case for implementation
• Use outcomes to build buy-in and refine the framework for broader rollout

For related background, review the Capabilities and Prompting guide pages.