Salesforce AI: Exploring Copilot and Beyond!

Feb 22, 2024 6:32:36 PM

Understanding the distinction between certain terms and concepts is crucial, as they can be confusing. GPT, or Generative Pre-trained Transformers, is the AI approach popularized by OpenAI. ChatGPT is the application that utilizes the GPT engine and gained popularity in 2023, becoming synonymous with generative artificial intelligence.

ChatGPT enables interaction with a language model by entering text and receiving responses. There are various language model engines, such as those by Google and Amazon, forming the foundation for tools like ChatGPT.

AI has been present for a long time

Generative artificial intelligence has a history of around 60 years, and recent advancements in computing have made it competitive. Salesforce has had the predictive capabilities of Einstein for many years.

There are two distinct ways to utilize these models:

One is through ChatGPT, a general platform for interacting and answering questions.
On the other hand, providers like Salesforce focus on specific use cases, such as customer service or generating user stories from process diagrams

The difference is clear:

ChatGPT serves as a general substitute for search engines, used to answer questions and engage in conversations.
In contrast, specific use cases like Salesforce generate tools to address concrete problems

Looking at the long-term perspective, artificial intelligence, especially generative AI, emerges as a set of essential tools to complement existing skills within companies and enhance roles and tasks. While there may be challenges, early adoption provides significant opportunities in the ever-evolving business landscape.

Return on Investment

Concerning the ROI of artificial intelligence, a notable improvement in productivity stands out, with the use of capabilities like ChatGPT in marketing leading to increased efficiency. On the development side, the utilization of solutions such as GitHub Copilot and AWS Code Whisperer significantly enhances productivity. A study involving over 1,500 developers gathered insights into the frequency of Copilot usage and its impact on their work. Effectively, generative text can assist in performing tasks that are currently done but in a smarter, faster, and automated manner. The results demonstrate that AI has a positive impact on developers' productivity (and happiness).

Faster code writing: Copilot can assist developers in completing coding tasks up to 50% faster.
Fewer errors: Copilot helps developers reduce the number of errors in their code.
More satisfying professional experience: Developers report feeling more satisfied and productive in their work.

On the talent front, many companies are upskilling their workforce, training them on how to use generative artificial intelligence in their actual workflow, yielding a return on investment by improving speed. One way to measure this is through conducting experiments, training a user group in ChatGPT, and measuring their results compared to a group not using it, verifying data such as speed, total number of tasks, etc. It not only enhances productivity but also quality. Even a 20% improvement in productivity already signifies a valuable return on investment.

Salesforce Sales AI and Service AI

Sales AI provides a comprehensive solution to enhance sales performance. It enables the automatic generation of personalized email based on CRM data, the creation of call summaries with actionable insights, and assistance throughout the sales cycle. Additionally, it offers automatic call transcriptions and guidance for next steps. It streamlines the transition from marketing leads to sales, automates prospect research, automatically syncs data, and provides predictive forecasting to optimize sales management. It eliminates manual data entry, improving accuracy in business forecasting.

Service AI provides advanced tools to enhance customer efficiency and satisfaction. It delivers AI-generated real-time responses for channels like SMS and Whatsapp, along with automatic summaries of conversations and knowledge articles to save agents time. AI-driven knowledge search makes finding answers quick, offering recommendations based on reliable data, case classification automation, and response suggestions to optimize service management. Moreover, it enables the creation of multichannel, multilingual bots, providing automation for routine tasks and enhancing the customer experience across various platforms. The API allows expansion to custom channels, offering tailored solutions for specific needs in marketing, sales, and human resources.

Einstein One

The Einstein One platform was announced at Dreamforce 2023. Before the event, there were numerous statements and brand reorganization with GPT products (Einstein GPT, Sales GPT, etc.). Now, Einstein One is the general term for an entire group of solutions.

Salesforce has chosen to omit the GPT term in the name because, while it will use the OpenAI model, it will also allow the use of other models. This is significant as it provides flexibility for potential future use of models like Amazon's Jurassic or Titan instead of OpenAI's model, a smart branding move.

The primary driver at this moment is chatbots. ChatGPT demonstrated its effectiveness in interaction. With Copilot, Salesforce is integrating a solution into the CRM to interact with that language model. It's an intelligent, automated, no-code integration using multiple language models. This solution includes Data Cloud, CRM applications, the trust layer of Einstein, and then artificial intelligence.

Salesforce is constructing a unique and cohesive set of layers in the product, from metadata at the bottom to Einstein One. This unification will streamline the work for those building solutions and for everyone using them. Salesforce has clearly identified what customers need: user-friendly, configurable artificial intelligence integrated into the CRM.

Copilot and Copilot Studio

What are the differences?

Copilot is the ready-to-use conversational AI assistant integrated into the user experience of each Salesforce application. Einstein Copilot enhances productivity by assisting users within their workflow, allowing them to ask questions in natural language and receive relevant, reliable answers based on secure and proprietary company data from Data Cloud. Additionally, Einstein Copilot takes proactive measures and provides additional options beyond user queries, such as offering a recommended action plan after a sales call, checking the status of a consumer's order, or adjusting the shipping date.

Copilot Studio is where the configuration of what this assistant can and cannot do takes place. It is a straightforward tool for companies to create an entirely new generation of AI-powered applications with prompts, skills, and custom AI models. It can be used to close sales deals faster, optimize customer service, automatically generate websites based on personalized browsing history, or convert natural language messages into code, among many other business tasks. It provides configuration capability for Einstein Copilot to be available for use in consumer-oriented channels, such as websites, to drive real-time chat or integrate with messaging platforms like Slack, WhatsApp, or SMS. The configurators are:

Skill Builder: Allows businesses to create custom AI-driven actions that perform specific tasks. For instance, it's possible to create a "Competitive Analysis" skill for meeting preparation. This skill will analyze current market data, sales figures, and send API calls to external databases, resulting in a comprehensive overview of the competition. As the meeting date approaches, the sales representative simply informs Einstein Copilot, which quickly selects the "Competitive Analysis" skill to provide crucial insights and talking points.
Model Builder: Businesses require flexibility to select their own AI models (BYOM) for their specific business needs. Model Builder will enable customers to choose one of Salesforce's proprietary LLMs or seamlessly integrate preferred predictive and generative AI models from partner companies, and train or fine-tune them with data in Data Cloud without moving or copying them. This equips Einstein Copilot with more accurate information and content tailored to the unique dynamics of employees or customers. Model Builder will support "bring your own model" integrations with Amazon Bedrock, Amazon SageMaker, Anthropic, Cohere, Databricks, Google Cloud's Vertex AI, and OpenAI.
Prompt Builder: Allows the creation, testing, and deployment of AI-generated prompts that match your company's brand and communication style, all without requiring extensive technical expertise. For example, a marketing specialist could request it to generate a personalized message and discount for a new product based on the customer's purchase history and location. Then, Einstein Copilot will automatically generate personalized messages aligned with individual customer preferences, referencing previous purchases and demographic information. Unlike other notification engineering tools that require integration and specific skills, Prompt Builder is a low-code experience that quickly delivers well-informed notifications in any CRM application or workflow, with the Einstein Trust Layer included.

Trust Layer

Einstein Trust Layer is a secure AI architecture natively integrated into the Salesforce platform. Designed for enterprise security standards, Einstein Trust Layer allows teams to harness generative AI without compromising customer data, while enabling businesses to use their trusted data to enhance generative AI responses.

Integrated and Grounded: Defaultly integrated into every Einstein Copilot, Trust Layer grounds and enriches generative prompts in reliable company data through integration with Salesforce Data Cloud.
Zero Data Retention and PII Protection: Companies can rest assured that their data will never be retained by external LLM providers, and customer Personally Identifiable Information (PII) masking offers enhanced data privacy.
Toxicity Awareness and Compliance-Ready AI Monitoring: A security-aware LLM will safeguard against toxicity and brand risks by "scoring" AI generations to ensure confidence in the safety of responses. Additionally, each AI interaction will be captured in a secure and monitored audit log, providing companies with visibility and control over how their data is utilized.

The Future

How much more can we expect AI to evolve in the coming months and years? The waves of AI were already showcased at Dreamforce:

We are currently in the generative wave.
The next wave will be autonomous bots.
The final wave will be AGI, general artificial intelligence.

We are currently in the early phase of the generative wave of artificial intelligence (AI), with companies implementing solutions and much work to be done. As this wave succeeds, it lays the groundwork for the next phase: autonomous bots. The role of a Salesforce administrator will evolve towards overseeing and managing AI, moving away from managing specific processes.

With advancements in reducing human intervention, the importance of trust stands out. While delegating functions to AI makes sense, involving humans before sending responses to customers is essential. Labeling AI responses is vital to inform people about their origin. In this transformation process, AI plays a significant role, generating significant implications for the skill set required within a company.

Conclusions

AI is not a passing trend but a transformative force in the business landscape. Embracing AI, understanding its nuances, and adopting the right AI tools are essential for ongoing success in the digital era. Salesforce has paved the way for a new era of artificial intelligence. Training, deep understanding, and readiness for future waves of AI are crucial to harnessing the full potential of these innovations. Trust, ethics, and humanity must remain guiding principles on the journey toward an AI-driven future.

Explore our Quick Starts for an immediate start with Sales Cloud and Service Cloud.

If you are ready to take the next step, contact us to discuss how we can help your business succeed with the Salesforce platform.

Sergio Gaffoglio Sergio is responsible for overseeing the Salesforce area, including commercial, technical, and strategic objectives. With over 25 years of experience in the digital industry, he provides leadership and plans technological and marketing projects.

YOU MIGHT ALSO BE INTERESTED IN THIS CONTENT