Building the Business Case for AI & ML in Your Business

Introduction

Artificial Intelligence (AI) and Machine Learning (ML) are no longer just buzzwords; they are becoming the driving force behind the digital transformation for businesses across the world. As technology continues to evolve, organisations are turning to AI and ML to increase efficiency, reduce costs, and drive growth. However, adopting AI and ML requires a significant investment of time, money, and resources, which can be a daunting prospect for many companies.?

In this blog, I will explore how organisations can make a compelling business case for AI and ML adoption, including the benefits, challenges, and potential ROI of implementing these technologies. I will also provide some practical tips and strategies to help you get started on your AI and ML journey.

Things to Consider When Getting Started With Your Business Case

When it comes to adopting new technologies like AI and ML, businesses often face a trade-off between time, cost, and quality. On one hand, companies want to implement AI and ML solutions as quickly and efficiently as possible to stay competitive and meet customer demands. On the other hand, they need to ensure that the adoption process doesn't compromise the quality of the end product or service. Or bring ethical challenges to the surface.?

Additionally, cost is always a concern when it comes to implementing new technology, as companies need to justify the investment in terms of the value it will deliver over time. Things to recognise that your organisation will have to absolutely invest in when it comes to building a new AI & ML fuelled capability include;?

1. New technology and foundational architecture needs (E.g., Data Platform, Data Catalogue, MLOps & Data Science Toolkits)
2. New or additional engineering skills (E.g., Data Scientists, MLOps Engineers)
3. Internal marketing, awareness, education and training for colleagues, partners, suppliers and potentially customers.?
4. Organisations need to assess the availability and accessibility of the data required to train and develop AI and ML models. This may result in extensive data clean-up or readiness activities.?
5. Investing in the creation of a new ethics framework and AI/ML engineering standards.

It’s also important to have a view on your investment needs so as to understand both CAPEX and OPEX requirements usually over a 2 or 5 year payback period. This will enable your organisation to model at which month/year your investment becomes net positive in terms of money out the door, weighted against the value that your AI/ML investments are generating for your business.

These investment volumes are often dictated by the type of use cases your business has in mind for AI & ML. Broadly speaking, AI & ML can be used to generate value across the following scenarios:

• Cost Savings: AI and ML can help organisations reduce costs by automating manual processes, optimising resource utilisation, and reducing waste.

• Increased Efficiency: Organisations can improve efficiency by automating repetitive tasks, optimising workflows, and reducing cycle times with ML & AI.

• Improved Quality: The quality products or services can be improved by detecting defects, reducing errors, and improving accuracy. Reducing the need for rework and duplicative cost.?

• Personalised Customer Experience: AI and ML can help organisations provide a personalised customer experience by analysing customer data, predicting customer preferences, and tailoring recommendations and offerings to individual customers. Enabling organisations to cross-sell their products and services.?

• Better Decision-Making: Organisations can leverage ML & AI to make better decisions by providing insights and recommendations based on data analysis and predictive modelling.

• Competitive Advantage: AI and ML can enable organisations to gain a competitive advantage by improving efficiency, quality, and customer experience, allowing them to differentiate themselves from competitors.

• New Revenue Streams: The use of AI and ML can generate new revenue streams by analysing customer data, identifying new market opportunities, and developing new products and services that enable organisations to launch new products in future markets and growth region’s.?

However, the data points required to substantiate the investment to support your organisation's adoption of AI & ML are typically dictated by the use case you want to tackle.?

Different Horses for Different Courses?

In order to build a justified business case to support your adoption of AI & ML it’s important to firstly understand your use cases and specifically whether they are centred on enabling your organisation to make money, save money or reduce risk. Equally, it could be a mixture of the three!?

In each of these scenarios the data points you need to support your investment case will be slightly different. Namely;?

• To make money, you’ll be interested in things like current customer portfolio & demographics, time to deliver new services and features, size of market opportunity, transaction volumes, propensity to purchase, mean time to purchase customer satisfaction data and wider economical considerations.

• To save money, you’ll be interested in things like existing operational costs such as labour/people, facilities, asset data, technology investments, frequency of events, mean time to completion and additional OPEX overheads.?

• To reduce risk, your organisation should be concerned with data points like the offsetting of regulatory penalties, the impact of reputational damage, market competition & revenue loss.??

Equally, your use cases could be tied to directly improving customer satisfaction targets or supporting the realisation of ESG reporting obligations. Irrespective of the use case, specific data points will need to be captured to establish a forecast return on investment (ROI) and total cost of ownership (TCO).

Let's break these down in a bit more detail.

What Data Points Should You Look To Baseline?

Typically, when building a business case for AI and ML adoption, organisations should consider capturing operational metrics to demonstrate the impact of AI and ML on areas like procedural efficiencies, cost savings or money making objectives. Here are some examples of the operational metrics that organisations should consider capturing:

• Production Rates: The number of units produced over a specified period can demonstrate the impact of AI and ML on production efficiency if the right interventions are made across a supply chain. This should consider factors such as volume, time and maximum throughput with available resources.??

• Cycle Times: The time taken to complete a particular process or operation can indicate the efficiency of the process and the potential for improvement through AI and ML. This can also be tied to frequency or occurrences and lost revenue/customer conversions as a consequence of operational overruns.?

• Quality Metrics: The number of defects or errors in the production process or product can indicate the quality of the output and the potential for improvement through AI and ML. This could be applied to software development practices and equally to the use of digital twins to support product testing and simulation/scenario analysis.?

• Resource Utilisation: The usage of resources, such as staffing, equipment, and raw materials, can indicate the efficiency of the production process and the potential for optimisation through AI and ML.?

• Costs: Everyone’s favourite subject. Cost can be allocated across a raft of dimensions. Whether this be, people, resources, assets, facilities or technology enablers.?

• Downtime: The amount of time production is stopped due to equipment/software failure or other reasons can indicate the reliability of the production process/product and the potential for improvement through predictive maintenance with AI and ML. As well as the impact of the downtime. Both financial and reputational.?

• Inventory Levels: The levels of inventory can indicate the efficiency of the supply chain and the potential for optimisation through demand forecasting and inventory management with AI and ML.

• Customer Satisfaction: The level of customer satisfaction can indicate the quality of the output and the potential for improvement through AI and ML to personalise and improve customer experience.

• Market Share: Your organisations current market share for a specific product or service can demonstrate the impact of the investment on the business's competitiveness and growth potential.

• Customer Resolution Windows: The frequency of customer interactions, the type of query, the speed of resolution and the volume of first time resolutions for enquiry types.?

• Revenue & Market Opportunity: Views on the current trajectory of growth and profit forecasts for specific products and services are key. Highlighting the opportunity to either accelerate growth by streamlining the production and supply chain of goods with AI & ML driven solutions can be valuable in establishing a sound business case. Equally, highlighting operational efficiencies with a combination of the data points called out above can be quality compelling for C-Level executives.

Numbers Alone Won’t Win The Battle

While having a robust business case for AI and ML adoption is essential, it's not enough to convince stakeholders to invest in new technology alone. To truly gain buy-in from decision-makers, you must have a compelling event and an interesting story to support the business case.?

A compelling event is a trigger that creates a sense of urgency or necessity for change, such as a crisis, a change in market conditions, or a regulatory change. This should be woven into an interesting story that captures the imagination and emotion of the audience. Whilst communicating the value and potential of AI and ML adoption in a memorable way. Focussing on the impact it can have on colleagues, customers and how the wider community would see your organisation is a great starter for 10.?

By combining a compelling event with an interesting story, organisations can create a sense of excitement and momentum around their business case, which can help secure the resources and support necessary to implement AI and ML solutions successfully.?

Having this articulated through a number of mediums can be super helpful to winning hearts and minds. Short snappy video testimonials from customers and colleagues can highlight their challenges and experiences with your products and services. Supporting these with impact statements, detailed hypotheses, infographics and data driven charts and figures can ensure your message is capable of traversing multiple stakeholders across the business.?

In short, having a clearly articulated “Why?”, tied back to your organisations business objectives and strategic outcomes is critical.

Tailor Your Message Accordingly For the Audience?

Tailoring your story to the audience is essential when presenting a business case for AI and ML adoption because different stakeholders have different priorities, perspectives, and levels of technical expertise.

Always consider the relevance of your story. For instance, C-level executives may be more interested in the financial benefits of AI and ML adoption, while engineers may be more interested in the technical aspects of the technology and how it will make their roles more exciting and value focussed. Rather than built around the mundane and operational tasks in their day to day role.?

In addition, not everybody understands technical jargon and three letter acronyms. Therefore, your story should be communicated in a language and style that the audience can understand. Technical jargon and complex explanations may be appropriate for engineers but may confuse or overwhelm other stakeholders. Try to make things relatable to everyday situations that everyone is familiar with.?

One such example I have referred to was by leveraging the Amazon Store concept to explain how a data catalogue, data products and metadata accelerate the discovery and accessibility of key data sets across the business. Namely, the data catalogue provides me with an Amazon Store search bar. I use that to identify a book/dataset I am interested in. I can review samples of the book/data schema to ensure it meets my criteria. I can read reviews to increase my confidence that the book, or in this case, the data is what I am looking for, and so on.

Making your story relatable is key to winning hearts and minds to support your business case of AI & ML in being signed off. Finally, tailoring the business case to demonstrate empathy can be a key way of persuading budget holders to sign-off new or additional expenditure. This can build trust and rapport with stakeholders, increasing the likelihood of? securing their support for your AI/ML business case.

In Closing

In summary, building a compelling business case for AI and ML adoption requires careful consideration of several key factors, including business goals and objectives, data readiness, resource availability, ethical considerations, and the potential ROI. By assessing these factors and capturing relevant operational and business-based metrics, organisations can demonstrate the impact of AI and ML adoption on operational efficiencies, customer experience, and business growth.?

Additionally, to gain buy-in from stakeholders, it's essential to have a compelling event and an interesting story that resonates with the audience and communicates the value of AI and ML adoption effectively. By tailoring the story to the audience's needs, concerns, and interests, you will be able to increase the persuasive power of your business case and secure the resources and support necessary to implement AI and ML solutions successfully.?

In short, AI and ML adoption can generate significant value for organisations, but building a strong business case and communicating its potential effectively are crucial to realising the full benefits of this technology.