Google Cloud recently published 601 real-world gen AI use cases from the world's leading organizations.

The publication is worth a read to get a sense of what businesses have been working on for the past couple of years.

There is a broader point: any B2B company should have a knowledge base listing customer use cases and impacts, and should use it to personalize all outbound sales communications.

I was recently invited to attend a company’s presentation about AI solutions for sales. The company’s follow-up email was customized with my name, but the rest of the message was generic, leading to the obvious question: shouldn’t they have used AI to personalize the email?

We can use Google Cloud’s blog post to try this out.

Let’s consider the following scenario: You are selling AI solutions and you’ve got a list of email contacts from a webinar. Each contact is associated with an industry sector (e.g., financial services, healthcare) and a function (e.g., sales, product, customer service).

You can use the list of customer case studies to generate personalized emails and tell each prospect about the specific ways that GenAI can help their business.

Read below for details.

The “GenAI for you” outreach assistant

The “GenAI for you” outreach assistant is able to take a list of email contacts, access a list of customer use cases and business impacts from a knowledge base, and send one customized email to each contact.

Each email looks like this:

Let’s look at the steps needed for a simple implementation of this approach.

Inputs

Email contact list

Typically, a registration form for gated content (e.g., webinar, report) generates the email contact list, which looks like this:

Knowledge base

The knowledge base is a list of GenAI use cases and business impacts.

For the purpose of our demo, we can use Google Cloud’s list of 601 GenAI use cases, which looks like this:

We can get the knowledge base online from Google Cloud’s post. For the purpose of the demo, we simply get it from a knowledge_base.md file.

Email generation

The list of relevant case studies can be generated using the following prompt for each prospect:

You are an sales assistant tasked with generating between 3 and 5 AI case studies for your customer. 

You are provided a knowledge base consisting of a list of real-world AI use cases, located between the <knowledge_base> tags below.
You know the industry and function of the customer:
* Industry: {industry}
* Work fuction: {function}

Select the case studies that are most relevant to your customer and exhibit the most striking business value, prioritizing the use cases that include quantitative impact.

You must respond with a list of case studies in bullet points in markdown format, with each case study as a bullet point, and the case study details as sub bullet points. The case study details must include, first, a **summary** sub-bullet point describing the concept of the use case and the name of the company, and second, a **business value** sub-bullet point describing the impacts achieved. Do not include any other text in the response.

When formatting Markdown, the bullet points must start with a star, and the sub-bullet points must be indented with four spaces followed by a star.

Here is the knowledge base:
<knowledge_base>
{knowledge_base}
</knowledge_base>

You can find the full demo and code at this link:

Visit this notebook.

In case the knowledge base is much larger, you may consider filtering it using RAG first, before the email generation step.

Takeaway messages

Sales outreach personalization is achievable at scale using GenAI tools and relevant knowledge bases. Of course, the quality of outputs depends heavily on the knowledge base and prompt engineering. By tracking email opens, clicks and responses, businesses can measure the performance of their agents and refine their approach.