How should business leaders think about implementing GenAI - a short guide

As organisations are thinking about building internal and integrating external GenAI tools, there exists a lot of hype, but as you may have realised by now - leveraging the power of GenAI is expensive and requires a business use case. If you have ever sat for a product demo they will show you a slick product, and its easy to give into the buzz but looking at it through a first-principles approach can help you make an informed decision. Here is a simple framework to start with.

1. Identifying the right use case - GenAI can do a lot, but it also can't.
2. Buy vs Build?
3. Setting expectations for ROI - where do you see impact and buy how much?

✔ Identifying the right use-case

The first step toward any technology implementation is starting with the business case. This becomes the foundation of any subsequent decision. Below is a 3-step process to identify the right GenAI use case, starting with

1. Identifying a customer pain point.
2. Mapping relevant customer data - do you have the data to solve the problem? This could be internal product sales data, supply chain data, customer journey etc. This data must be cleaned and ready to be consumed.
3. Matching with GenAI capabilties - Once you know the customer problem statement and have looked at data availability, the next step is to match it with the proven solutions GenAI today can offer.

Sample Case:

1️⃣ Business problem - answer questions about product features and specifications. Eg. Is this service part compatible with the product I own?

2️⃣ Data - Product specifications and description documentation. The LLMs will be trained on this data set.

3️⃣ GenAI capabilities - Chatbot feature through which the customers can interact.

Note on Data

Embedding data across the customer journey is critical to building solutions using AI. Think of data as reusable building blocks for various applications. Once the data collection is in place, the next step is to build the right data architecture or the plumbing. Without the right data architecture, teams will take time to consume the data. And finally having the right set of data governance practices is important to maintain oversight and meet compliance requirements.

🛠 Buy vs Build?

With a solution identified the next step is to think about the implementation. While there are many factors to consider including but not limited to budget, organisation culture, capability etc, from a technology standpoint, you should consider the following 4 resources:

1. Large amounts of data - do you own this data? If not can you buy it? Without owning the data it is not possible to build an in-house genAI capability.
2. Compute and storage capacity - GenAI requires a lot of compute and data storage. Unless the development is taking place on the cloud, scaling compute and storage capacity can be a challenge. Even with cloud computing, the cost for maintaining the excess server capacity should be evaluated.
3. Talent and skills - Do you have the right AI/ML skills within you tech teams?
4. Implementation time - What is the urgency of the solution? How critical the need for the solution to be implemented?

Many technology providers such as Microsoft, ServiceNow, SAP, Salesforce, Adobe, Github etc. are investing in GenAI capabilities and offering solutions. This presents a plethora of options to evaluate and compare GenAI solutions. Unless the use case is highly spiralized and unique, its best to buy rather build.

💲 Setting expectations for ROI - where do you see impact and buy how much?

There is significant opportunity to be realised with GenAI, however most of the gains will accrue in the long term. According to Goldman Sachs, GenAI could boost annual productivity by 1.5%, driving $7 trillion in added economic value over the next decade. There are success stories where companies have realised benefits of GenAI such as Mercado Libre which found that augmenting their 9,000 human developers with GitHub CoPilot resulted in a 50% reduction in time spent coding. These productivity gains can translate to $$ savings.

However this is where most companies run into challenges. Estimating and proving the business value of GenAI.

This is why defining the metrics for ROI is critical. Today GenAI can move the metrics across

• Improving customer experience - NPS, CSAT, Response time etc.
• Lowering unit cost - COGS, Employee Productivity etc.

If the right expectations are set early on, it can set the implementation on the path of success.

I couldn’t be more excited for the value GenAI can drive in the enterprise – as long as we keep our eyes on the prize: the customer needs.

🤔 Want to discuss more on GenAI? reach out to me on LN — Amatya Agarwal