Salesforce announces strong AI integration across its applications at DreamForce

At the recently concluded DreamForce Annual conference, Salesforce announces several mega AI initiatives that integrate across its entire spectrum of applications and beyond

The recently concluded DreamForce conference by Salesforce is being touted as the largest AI event of the year, with 100+ AI experts and visionary speakers across 1,500 sessions and attended in-person by over 40,000 attendees. It was also the Most Sustainable DreamForce ever, with 74% of the structures being re-cycled or re-used, 5500 bio-degradable signages, 10M gallons of water conserved, and 100% compostable meal packaging. All sessions were also net zero. 

AI was the main theme this time and while the software giant focused on how Salesforce has integrated AI into all of its services, the company’s founder and CEO, Marc Benioff called it responsible innovation paired by ethical use. “It’s all about using Data + AI and adding a layer of Trust on top of it”, he said.

Helping the Society 

Before delving into what’s new at Salesforce, Marc also stressed on the importance of helping the society beyond doing business, for which Salesforce has put 1% of its equity, profits, and people’s time into charity. This equates to about 700 million USD in donations, 8.3 million hours of volunteer time, and 54,000 nonprofits running on Salesforce platform. The company is also helping San Francisco’s public schools and hospitals and has raised over 100M USD for them.

Salesforce Becomes Third Largest Software Company in the World

Marc started by announcing that Salesforce has become the THIRD LARGEST software company in the world at USD 34.8 Billion in FY24 and SECOND LARGEST in Japan, where they have surpassed Oracle. 

Most Corporates Looking to Deploy AI, but Trust is the Biggest Concern

While Salesforce’s focus on AI is over a decade old with its Einstein AI platform, Marc divided the AI evolution into four waves. The first was predictive AI, which we’ve already been through. Second is GenAI, which we’ve been into for around 9 months. This would be followed by Autonomous and Agents based AI, and eventually we’ll get to General Intelligence AI. When we think of the last AI wave, scenes from movies like Terminator and Minority Reports run wild in the mind. “While there are both good and bad parts in those movies about AI, it will be our values that will guide us in the right direction right”, said Marc. Those values will help us achieve higher levels of customer success, productivity, and growth. However, like many new technologies, there’s a trust gap in AI while working with so many LLMs. They’re always hungry for data and keep getting smarter with it using advanced algorithms. “Unfortunately, most of this data is stolen and moving ahead, these LLM companies would like to have the corporate customer’s data, which is not something most corporates would agree to”, said Marc. Second, though these LLMs are smart, they may give convincing answers that aren’t exactly true and could even turn toxic very quickly. “What’s needed is to build trust into the applications that are built based on these LLMs as trust is a priority for every business”, he added. 

Announcing Einstein 1 Platform

Speaking of trust, Marc announced the Einstein 1 AI platform, a trusted AI platform that’s deeply integrated into all Salesforce applications (Sales, Marketing, Service, and Commerce). It’s not running on a separate data cloud or server so there’s no need for separate teams to operate it. Einstein is currently Generative but will soon be autonomous and eventually become General AI based. Besides integrating with the core applications, Salesforce has also connected it to Slack, Tableau, and even external applications from Google, Microsoft, etc.

All meta data on the Einstein 1 platform is integrated and shared across all SF applications. This covers the data model, sharing model, validation rules, record types, page layouts, sharing rules, permissions and profiles, code, and flows to name a few. This makes it fast and easy for customers to build their own applications once and write anywhere. All upgrades to the software are also automatic and seamless. 

This integration is part of a mission at Salesforce called More Core where the idea is to move more apps into the core. It will enable Salesforce’s Trailblazers (a community that uses Salesforce to build careers, companies, and communities on Salesforce platform) to do much more on the platform, automate a lot of tasks, etc. Moreover, it’s open so all Salesforce partners can benefit from it. Plus, it’s all low code/no code, enabling the trailblazers to do all automation non-programmatically. 

Speaking of trailblazers, between 2022 to 2028, Salesforce is expected to have 20 million trailblazers globally, over 2000 community groups, 11.6 million new salesforce ecosystem jobs by 2028, and $2 Tn+ net gain in business revenue globally. 

$500M Venture Fund Announced for Einstein Applications

Another major announcement at DreamForce was the $500 million venture fund, which is a part of Salesforce ventures. SF Ventures has deployed $5B with 28% IRR, making it one of the most successful VCs in Silicon Valley. Salesforce uses the venture capital to fund entrepreneurs and innovators. 

The new $500M fund is targeted toward Einstein applications so that more trailblazers can start their own companies, build on the platform and sell to customers. 

In AI We Trust! 

“GenAI applications like ChatGPT take you directly to the LLM to enter data. Salesforce assures privacy of customer data with the Einstein Trust Layer," - Parker Harris, CTO, Salesforce

The entire AI revolution will be built on trust, for which Salesforce feels it’s important to have new standards for ethical and humane use of AI. For this, several years ago, Salesforce created a Chief Trust Officer in the company well before anyone else. The company firmly believes that customer data is not their product, unless many companies especially on the consumer side who use customer data to make money. Salesforce strongly believes in promoting responsible AI globally, due to which they allow customers to have control over their data. This sort of transparency helps builds trust. 

On this, Parker Harris, Co-founder and CTO of Salesforce also stressed on the need to build trust in AI and how Salesforce is doing it. He referred to something that they’ve built called the Einstein Trust Layer. GenAI applications like ChatGPT take you directly to the LLM to enter data. Salesforce assures privacy of customer data with the Einstein Trust Layer. The company will do secure data retrieval and dynamic grounding so that the response is relevant. They don’t send PII (Personally Identifiable Information), Credit Card info, etc. to the LLM. All of it is masked when sent to the Secure LLM Gateway and unmasked when it returns back to the customer. The company has built over 10 such LLMs, which don’t keep customer information. Moreover, when the data comes back, the Einstein platform first checks it for toxity, ethical issues, etc., keeps an audit trail of all the ways that AI was used for customer’s reference to ensure they can track the value they’re getting from it. 

Enabling Enterprises to Leverage their Data Assets

Every enterprise has vast amounts of data that they’d like to be able to use, but they find it challenging to do so because it’s located across several islands of disconnected data (e.g., mainframes, multiple cloud providers, etc.). In fact, according to Salesforce, 71% of company applications are disconnected, and would like to bring them together in order to apply AI on it. This is where Einstein Data cloud can help by connecting all data and integrating it into all the SF apps. As it’s not a separate platform, customers can create their own flow, dashboards, etc. 

The company is already processing 2 Trillion+ records per month on this platform with data coming not only from the CRM, but other apps and platforms as well like AWS, Data Bricks, Google, etc. Salesforce has partnered with them and integrated their data lakes into the Einstein Data Cloud. This enables customers to get a single, holistic view of their customers. 

Sanjana Parulekar, VP Product Marketing at Salesforce demonstrated how easy it is for customers to connect their wealth of data from different systems with their critical business apps. As the platform works on the principle of low code/no code, customers can connect to any data in their organization, including external data using connectors. Once they connect with the data, they can harmonize it into one data model and easily create powerful workflows. Using Data Cloud and Tableau for instance, you can provide access to all the required data to your business users in real-time. With fully integrated data, customer interactions become richer. Moreover, Data Cloud & Tableau is free to get started as the company gives you 10K profiles to get started. 

Einstein Co-Pilots

Commenting on this, David Schmaier, the President and Chief Product Officer at Salesforce said that the AI revolution is a customer revolution, allowing you to use Einstein to connect with your customers in entirely new ways. Taking this one more level above, he announced another Salesforce Innovation—Einstein 1 Co-Pilot conversational AI based assistants for every Salesforce application. Schmaier said that these are “One of the greatest innovations in the history of the company”. It allows everyone to become an Einstein and get more productive with every single app.

Built on the meta data framework, the Einstein 1 Co-Pilots are also there for every cloud, be it sales, marketing, service, e-commerce, or other apps like Tableau, Canvas, MuleSoft, etc. 

“Data powers AI and AI will power your customer experience” he added.

Einstein 1 Co-pilot for Sales allows you to look at all your leads, automatically fix your meetings, automatically transcribe meetings, and even send you action items. It can generate quotes, proposals, and the contract language for your sales team. Similarly, Einstein 1 Co-Pilot for Service andautomatically give responses and answers to your questions. Your Customer Service Reps will automatically be up to date about your products and customers. They can generate answers on the channel of your choice—SMS, email, WhatsApp, and other channels that your customers are concerned about. Similarly, the Einstein 1 Co-Pilot for Marketing provides features like segment intelligence, which allows you to take customer journeys, segments, and marketing campaigns and make them more tailored. “It’s mass personalization at scale for your mktg department”, said Shmaier.

Similarly, the Einstein 1 Co-Pilot for Commerce can put up and generate your commerce site automatically, designed for the highest conversion rate possible. It will automatically create all product descriptions, landing pages and even the literature for your products. Similarly, there are co-pilots for Tableau, Canvas, MuleSoft, and even industry specific co-pilots. For instance, Tableau Co-pilot allows you to quickly go from raw data to actionable insights from a conversational interface. 

Lastly, he announced the Einstein Co-Pilot Studio for companies that want to customize Einstein Copilot to build and tailor AI assistants with relevant prompts, skills, and AI models to accomplish specific sales, service, marketing, commerce, and IT tasks. 

The article has been written by Anil Chopra, VP-Research and Consulting, CyberMedia Research. The author was hosted by Salesforce in San Francisco