Exploring how Salesforce's Einstein (Generative AI) works with Prompt Builder.

Summary

This video script showcases Salesforce's integration of generative AI, Einstein, to enhance CRM data utilization for personalized outreach emails. It demonstrates how to create and customize prompts for AI to generate contextually rich and personalized content. The script addresses concerns about trust and customization, emphasizing the use of CRM data to ground prompts and ensure privacy within the Salesforce ecosystem.

Takeaways

• 📝 Salesforce integrates generative AI, like Einstein, to assist sales reps in creating personalized outreach emails.
• 🤖 The AI's effectiveness relies on the use of 'prompts', which are instructions or questions given to the AI to generate content.
• 🛠️ Prompt Builder in Salesforce allows customization of these prompts to tailor the AI's output to specific business needs.
• 🔍 A generic prompt may result in an impressive but unusable output due to the lack of context about the user, company, products, and customers.
• 🔄 Grounding the model involves embedding contextual information into the prompt to improve the AI's output relevance and personalization.
• 💬 The example shows embedding the customer's name from the CRM into the prompt to make the AI's response more personalized.
• 📑 CRM data can be leveraged to enrich prompts, including product focus and customer-specific information, to achieve a more grounded and contextual result.
• 🔄 Flows can be embedded within prompts to make them dynamic and reactive to each customer's unique situation.
• 💰 Conditional context, such as a customer's tier level which might qualify them for a discount, can be integrated into prompts for tailored offers.
• 🔒 Salesforce ensures that all private data used for AI-generated content stays within the Salesforce trust boundary and is not used to enhance other results.
• 🔄 The process of creating a grounded prompt using CRM data and integrating AI-generated personalized results enriches the product experience for customers.

Q & A

What is the main purpose of using generative AI in Salesforce's contact records?

The main purpose is to enable sales reps to draft personalized outreach emails using the data in CRMs, enhancing the customization and personalization of communication with customers.

What is Einstein in the context of Salesforce?

Einstein is Salesforce's suite of AI-powered services that is built into the platform to assist with tasks such as email drafting, leveraging generative AI capabilities.

What is a 'prompt' in the context of generative AI?

A 'prompt' is a question or instruction given to an AI system, guiding it to generate a specific type of content or response, such as an outreach email.

Why might a generic prompt result in an unusable output?

A generic prompt may result in an unusable output because it lacks the necessary context about the user, their company, products, and customers, leading to a response that is too broad and not personalized.

How can the AI model be 'grounded' to improve the quality of the generated content?

The AI model can be grounded by embedding context about the user's goals and specific details from the CRM into the prompt, which helps the AI generate more relevant and personalized content.

What is the role of the CRM in creating a grounded prompt?

The CRM provides the necessary data about the user, their company, products, and customers, which is then embedded into the prompt to create a grounded context for the AI to generate personalized content.

Can the prompts be customized differently for different customers?

Yes, prompts can be embedded with dynamic elements or flows that adjust based on customer-specific data, allowing for personalized and contextually relevant responses.

What is the significance of the 'gold tier customer' in the script?

The 'gold tier customer' signifies a high-value customer segment that may qualify for special offers or discounts, such as a 10% discount mentioned in the script, which can be factored into the personalized prompts.

How does Salesforce ensure the privacy of data used in generating AI results?

Salesforce ensures privacy by maintaining all data within the Salesforce trust boundary and not storing it outside of it. The data is also not used to enhance any results other than those of the user's own.

What is the final outcome of integrating generative AI with CRM data in Salesforce?

The final outcome is the creation of rich, personalized results that are seamlessly integrated into the product experience, allowing for more effective and customized outreach to customers.

Outlines

00:00

📝 Leveraging CRM Data for AI-Powered Emails

The script introduces the integration of generative AI, such as Einstein, within Salesforce's CRM to enhance personalized outreach emails. It raises questions about the functionality, trustworthiness, and customization of AI-generated content. The demo showcases the process of creating a prompt within the Prompt Builder to customize AI responses. The importance of grounding the AI model with relevant CRM data is emphasized to achieve usable and personalized results. The script also addresses privacy concerns, ensuring that data used for AI prompts remains within Salesforce's secure environment and is not used for any other purposes.

Keywords

💡Contact Record

A contact record refers to a database entry that contains all relevant information about a customer or a client within a Customer Relationship Management (CRM) system. In the context of the video, it is the foundational data that sales reps use to create personalized outreach emails. The script mentions that every day, sales reps are leveraging this data, which underscores its importance in the sales process.

💡CRM (Customer Relationship Management)

CRM is a technology that enables businesses to manage their interactions with current and potential customers. It typically involves using software to organize, automate, and synchronize sales, marketing, customer service, and technical support. The video emphasizes the role of CRMs in housing the data that sales reps use to personalize their outreach, highlighting the system's integral part in modern sales strategies.

💡Personalized Outreach Emails

Personalized outreach emails are marketing communications that are tailored to the specific needs, interests, or characteristics of the recipient. The video script discusses how sales reps use data from the CRM to create these emails, which is a key strategy for engaging customers in a more meaningful way and increasing the effectiveness of sales efforts.

💡Einstein

In the context of this video, Einstein refers to Salesforce's AI-powered tool that integrates with their CRM platform. It is designed to enhance the user experience by providing intelligent insights and automation. The script mentions Einstein being built into the experience, allowing sales reps to utilize generative AI for drafting outreach emails, which is a significant advancement in CRM capabilities.

💡Generative AI

Generative AI is a type of artificial intelligence that can create new content, such as text, images, or music, based on existing data. The video demonstrates how generative AI, through Salesforce's Einstein, can be used to draft personalized outreach emails. This technology is transformative as it automates the creation of content that is tailored to individual customer profiles.

💡Prompt

In the context of AI, a prompt is a question or set of instructions given to an AI system to generate a response or perform a task. The video explains that behind every generative AI result is a prompt, which is customized in the script's demo to include specific details about the customer and the sales context, thereby personalizing the AI's output.

💡Prompt Builder

Prompt Builder is a tool within the Salesforce platform that allows users to create and customize prompts for the AI system. The script describes using the Prompt Builder to embed context into the AI's instructions, which is crucial for generating more relevant and personalized results.

💡Grounded Prompt

A grounded prompt is one that is enriched with specific context or data that helps the AI generate more accurate and relevant responses. The video script illustrates how grounding the prompt with CRM data, such as the customer's name and product focus, leads to more personalized and effective outreach emails.

💡Open AI

Open AI is an AI research laboratory that develops and applies AI technologies. In the script, Open AI is mentioned as the AI provider selected for the Salesforce platform. The video demonstrates how Salesforce integrates with Open AI to leverage its capabilities for generating personalized content.

💡Salesforce Trust Boundary

The Salesforce Trust Boundary is a concept that refers to the security and privacy measures within the Salesforce platform, ensuring that customer data remains protected. The script reassures viewers that due to partnerships with AI providers, data used to create AI results is never stored outside this boundary, emphasizing the platform's commitment to data privacy and security.

💡Dynamic and Reactive

The terms 'dynamic' and 'reactive' refer to the ability of a system to adjust and respond in real-time to changes or inputs. In the script, the concept is applied to the AI's prompts, which can vary depending on the customer's details, ensuring that the AI's responses are always contextually relevant and personalized.

Highlights

Sales reps use CRM data in Salesforce to create personalized Outreach emails.

Einstein, Salesforce's AI, is integrated to assist in drafting Outreach emails.

Questions arise about trust, customization, and language adaptation of AI-generated content.

The concept of 'prompts' is introduced as the mechanism for AI interaction.

A prompt is a question or instruction given to AI, in this case, Open AI.

Generic prompts result in generic, unusable responses due to lack of context.

To improve results, the model must be 'grounded' with specific context.

CRM data, such as customer names, can be embedded in prompts to add context.

Embedding CRM data leads to more personalized and usable AI-generated results.

CRM data beyond basic information can be used to further personalize prompts.

Dynamic prompts can adapt to different customers using conditional logic.

Conditional context, such as customer tier, can be factored into the prompts.

AI-generated emails are personalized and reflect the language of the business and customers.

AI integration with CRM data enriches the personalization of the product experience.

Data privacy is maintained as none of the CRM data is stored outside Salesforce's trust boundary.

Salesforce's partnership with AI providers ensures data is not used to enhance other results.

The demo showcases the successful integration of AI with CRM data for personalized outreach.

Transcripts

so what we're looking at here is a

contact record inside Salesforce and it

contains the kind of data that all of us

have in our crms today and every day

sales reps are using this data to create

personalized Outreach emails but notice

that Einstein is now built into this

experience which means those same sales

reps can now use generative AI to draft

those Outreach emails but I think a lot

of us see a demo like this and we have

some questions like how does this work

or can I trust it or how do I customize

this to speak the language of my

customers and my business and those are

the questions we're going to tackle in

this demo because what we need to

understand is behind every generative AI

result like this is a prompt and here

inside prompt Builder we can create and

customize those prompts so how does it

work well to explore how these

Technologies work let's recreate what we

just saw we're going to generate an

email

for sales and then give our prompt a

name and what you'll see is when I hit

create we'll land inside an environment

that helps us create a great prompt so

what is a prompt well a prompt is just a

question or instruction we give to Ai

and in this case it's a question or

instruction we're giving to open AI

which is the AI provider I've selected

for this prompt so where do we start

well there is absolutely nothing

stopping me from just typing something

really generic like write me an Outreach

email and sending that directly to open

AI but what you'll see is this gets me a

result that is both impressive and also

completely unusable and why is that well

it's generic because open AI is missing

a lot of context about me my company my

products and my customer so how do we

fix this we fix it by grounding the

model which is really just a fancy way

of saying that we need to embed some

context about what we're trying to

accomplish into the prompt for example I

could do something small like the

customer's name is and then embed the

customer's name from the CRM and what

you'll see is that we are now starting

to create a grounded prompt where I'm

not just saying write me an Outreach

email I'm also saying the customer's

name is and when we send this to openai

you'll see that we are starting starting

to getting more personalized result but

we have so much more data than just this

available to us in the CRM so what I'm

going to do is paste a new prompt where

I'm embedding context about me my

product focus and my customer we might

even imagine that we want these prompts

to be different depending on the

customer and to accomplish that I can

embed a flow inside my prompt which

enables it to be dynamic and reactive to

each customer and now we truly get a

grounded prompt rich with context about

me my customer and conditional context

in this case that the customer is a gold

tier customer which qualifies them for a

10 discount and now when we send this

prompt to open AI what you'll see is we

get the kind of Rich magical

personalized results that we saw in our

original demo this is how these

generative AI AI Technologies work we're

creating a grounded prompt using the

data you already have in your CRM we're

sending those grounded prompts to our AI

providers and then integrating those

Rich personalized results directly into

our product experience but hold on a

second what are we doing to protect all

the private data being used to create

these results well because of the

Partnerships that we have with our AI

providers none of this data is ever

stored outside the Salesforce trust

boundary and it's absolutely never used

to enhance any other results other than

your own which means we can call this

done because we have recreated the demo

we originally saw which means when we

return to that contact record We Now

understand that behind this click right

here is a prompt grounded with CRM data

which enables us to get generative AI

results that speak the language of our

customers and our business