An AI Experiment

#Longpostalert ?

Inspired by Marc’s speech on Salesforce AI Day,I ran a little experiment that I would love for you to be a part of.

I watched the entire session and jotted down notes based off of my understanding.Then I fed those notes to ChatGPT in a prompt asking it to create a LinkedIn article using those.If you didn’t watch the keynote,this is a good summary for you,and if you did,it will be interesting to know your thoughts about how accurate you think this is and whether Prompt Engineering is the most valuable area to upskill in , going forward.

The ChatGPT Article, as is-

In the ever-evolving landscape of sales, marketing, and commerce, the integration of augmented and generative capabilities powered by AI has brought significant advancements. However, ensuring trust, equity, and responsibility in AI applications is paramount. This article explores the potential of trusted and responsible AI, highlighting the impact on equity, profitability, and employee well-being. We delve into the journey of AI, the challenges faced, and how Salesforce is pioneering the next generation of generative AI with trust at its core.

The Evolution of AI in Sales, Marketing, and Commerce:

Since the advent of AI's first generation, represented by Einstein in 2014, remarkable progress has been made. Einstein has been instrumental in generating trillions of predictions, enabling lead scoring, and empowering researchers, data scientists, and engineers worldwide. New techniques, such as auto feature engineering, auto model solutions, LLMS, and DecaNLP, have pushed the boundaries of AI capabilities.

The Importance of Trust in Generative AI:

Trust is the primary foundation of generative AI. Earlier approaches would vacuum data from the internet, often lacking trustworthiness. However, with LLMS becoming more productive, the new generation of generative AI focuses on trust at its core. Trust is cultivated through a sharing model that amalgamates data and presents it through algorithms. However, challenges such as security, privacy, scale, and ethical considerations have emerged, demanding innovative solutions.

Building Trust and Responsibility:

Salesforce recognizes the importance of trust and responsibility in AI. With an AI ethics team, the company is committed to protecting human rights and ensuring data privacy. Salesforce firmly believes that customer data is not their product. By providing accurate, safe, transparent, and empowering AI solutions, they strive to make ethical business decisions. Furthermore, Salesforce is committed to sustainability by minimizing their carbon and water footprint.

Expanding Networks and Data Cloud:

Salesforce's networks have expanded exponentially, handling trillions of transactions per week with Einstein. The Data Cloud, the fastest-growing cloud platform, has become the receptacle for every customer's data. By creating the Data Cloud and extending the core platform to other organizations, Salesforce aims to provide comprehensive customer insights and drive data-driven decision-making.

Leveraging Generative AI with Trust:

Salesforce's Customer 360, augmented with generative AI capabilities, unlocks new possibilities. However, trust remains a crucial consideration. The sharing and security model addresses concerns related to privacy, data control, bias, and toxicity. The first-generation trust layer, developed through Einstein, enabled predictive AI in CRM. Now, the challenge is to leverage generative AI's productivity without compromising data security.

Ensuring Data Safety and Usage:

To establish trust in generative AI, several steps are taken. It starts with a prompt or question posed to the LLMS. However, the generation may lack context. Training the LLMS or using grounding techniques with keywords helps improve contextual understanding. The data used in training is protected through a trust layer, separating corporate data and responsibly grounding prompts. Data masking ensures prompt deletion, preventing sensitive data from being learned by the LLMS.

Trusted Infrastructure and Data Lake House:

Salesforce's Hyperforce provides a trusted infrastructure layer, ensuring the security and reliability of AI operations. The Data Cloud, combined with the lake house concept, consolidates data and enables powerful insights. Intelligent assistants powered by AI streamline processes and enhance productivity. Data, considered the fuel for AI, is treated with utmost care to maintain trust and responsibility.

Salesforce's Impact on Sales, Marketing, and Commerce:

Einstein GPT, as the most trusted AI, empowers sales and marketing professionals to personalize landing pages and generate highly performant content.

**My Notes**

Augment service agents-sales,marketing,commerce capabilities augmented and generative capabilities?

Focus on trusted and responsible ai

Equity,profit,employees?

Business is the greatest platform for change

Trust primary part of generative ai?

Innovation

Einstein first gen beginning of ai journey

Since 2014

Einstein-trillion predictions

Leads scoring

World class -researchers data scientists engineers

New techniques -auto feature engineering?

Auto model solution?

LLMS

deca nlp?

Prompting new tech

Auto regressive techniques

Protein generation

Trust security scale privacy ethics problems-new ai

Earlier-they would vacuum all data available on the internet and then give the best case using their llm?

Not trustworthy?

LLMS to become more productive?

Does not work in regulated industries?

Next gen generative ai with trust at its core

Sharing model

Amalgamating the data and presenting through their algorithm?

No cell based security in the generative model

2016 1st trust model

Ml/deep learning?

Without looking at the data

Same technology and create get trust layer

Trust and responsibility

Ai ethics team

Protecting human rights

Your data is not our product

Accurate as business decisions are being made

Safe

Transparent

Empowering?

Sustainable right size carbon water footprint?

Networks got expanded

Trillion transactions a week using Einstein?

Data cloud fastest growing cloud-

Creating data cloud with teams and technologies outside Salesforce thus extended core platform to others

12 trillion annually

Receptacle for data

Data cloud beginning of every customer

Customer 360 plus augment with generative ai

Trust?

Sharing and security model

Privacy,hallucinations,data control,bias,toxicity-not sociological but technical terms in ai thru trust gap

First gen trust layer through Einstein?

Predictive but also generative ai crm

How do I trust generative ai and how can we leverage all the productivity in generative ai without giving away data

Data safe and how does it get used

Starts with prompt,ques you ask the llm

The generation misses context

Training llm,or you cud use grounding giving keywords

Then generation is immediately usable

Protect that data from getting lost in the llm -trust layer-separate corporate data and responsibly ground prompts w

Data masking zero retention

Data masking-prompt never gets stored it’s deleted none of the sensitive data is learnt by llm

Hyper force-trusted infrastructure layer

Data cloud-lake house-

Assistants?

Data is the fuel for ai

Einstein gpt can take content that has performed best and make a personalised landing page

Public data

Actions like generating highly performant emails

Quick / summary

Will this be the end of competitive sales

Einstein gpt most trusted ai

Most important operational challenges-service sales agents struggle

To find the right answer

Sellers can meet and connect

Drives revenue across segments?

Emotional connection with consumers

Commerce gpt

Product descriptions in brand voice and tone and localised in languages

Tableau pulse-keeping up with the business as it happens?

Recommendations with human in loop

Collaboration through slack

Salesforce services to skill up your employees

Responsible ai?

Built in. Low risk?