Dreamforce 2023: The Power of AI and Data Cloud
Now free to try, Salesforce Data Cloud can help you unlock your AI strategy.
By Jim Clarke and Amas Tenumah
Dreamforce 2023 was one of the world’s largest AI events of the year. The Dreamforce campus had AI references everywhere, from deliberate design choices to fun Easter eggs.
For us, the highlight of Dreamforce 2023 was Slalom showcasing the first live, customer-facing generative AI (GenAI) chatbot built on Salesforce. Conference-goers even had the opportunity to chat with Vivi, Virgin Voyages’ new visionary chatbot, in a digital human form. She answered questions effortlessly (aside from the one-off, existential question, “What is the meaning of life?”).
There were many other notable announcements at Dreamforce regarding GenAI. Salesforce has grouped AI into a couple of themes we’ve seen systemically throughout all clouds: prompting and automations.
For Trailblazers, we highly recommend watching this keynote that covers nearly all AI-related cloud and product releases, including:
- Einstein 1 Prompt Builder
- Copilot for Sales
- Copilot for Service
- Service Cloud Case Management
- Marketing Cloud Segmentation Copilot
- Marketing Cloud Subject Line Generator
- Marketing Cloud Campaign Brief Copilot
- Commerce Product Description Copilot
- Commerce Page Designer Copilot
- MuleSoft Code Builder
- Slack AI
- Tableau Assistant and Tableau Prep
There’s no doubt that Salesforce leaders committed to going “all in” on AI. With other vendors a few steps ahead prior to Dreamforce, it’s clear the roadmap prioritization of AI was pushed to the very front.
So let’s dive into our key AI-related takeaways from Dreamforce 2023:
1. For AI to be successful, you need reliable data.
Nearly every session emphasized how critical reliable and accessible data is to AI. For any of these product announcements to work well, appropriate data is necessary to be able to do everything the product promises. However, the reality is that data is everywhere, and it’s messy.
With data being key to AI, Salesforce’s largest announcement coming out of Dreamforce was the Salesforce general availability (GA) $0 SKU for Data Cloud for customers with Enterprise Edition licenses and above. The power of this new wave of GenAI technology is the ability to process structured and unstructured data. To that end, Salesforce is making aggressive moves toward becoming the source for your GenAI data, and they’re doing this through Data Cloud.
In our opinion, Salesforce going big on Data Cloud is centered on two key strategies: big partnerships with data orgs (AWS, Snowflake, Google Cloud) and a very aggressive pricing strategy for Data Cloud (more or less free to try).
2. Accessibility of first-party data is key to navigating this new frontier.
For example, if a nontechnical person prompted GenAI to build a segmentation of all their financial services accounts that exist in North America and bought Salesforce in the last year, this technology could provide them with the four lines of code needed to generate that segment. Sounds easy enough, right? Yes, if we have the access to the right data in the platform.
With so many different types of data points — including purchase data, site traffic data, loyalty data, in-store shopping data, and social data — all of this data likely sits in disparate silos. It’s not uncommon for those datasets to be stored natively (in Snowflake, Google Cloud, or AWS, for example), but the effort to extract, transform, and load (ETL) this data into MarTech tools is done through custom integrations, adding to the cost of data storage in Salesforce. With Data Cloud, zero-copy data from these repositories could be exposed to the Salesforce Platform and be reachable by business teams and supporting GenAI tools.
For example, let’s say a fictional business called ACME houses all their purchase data in Snowflake, but wants to reference this data in Salesforce Marketing Cloud (SFMC) to create customized customer experiences. Traditionally, it would need a team to build an integration from that purchase data repository in Snowflake to SFMC, and the data would exist in those two locations. With zero copy, ACME can now “copy and paste” the data in SFMC without duplicating the data and taking up space in the system.
For AI to work, you need a lot of rich data, and zero-copy data is the best way to clone data without a ton of effort. Not to mention that there are considerable cost savings in not having to maintain two databases. Allowing marketers to use common language to seek insights and build segmentation increases efficiency, enabling teams to build for richer personalization.
3. Slalom’s accelerator helps clients harness the power of next-generation AI.
Slalom’s GenAI accelerator leverages this framework on the Salesforce platform while maintaining a flexible design so clients can explore solving many use cases. With the forementioned Vivi solution we built for Virgin Voyages, we focused on a virtual assistant leveraging Microsoft’s LLM — Azure AI — and their existing service cloud org, inclusive of their knowledge base (KB) for agents.
The power of using the GenAI framework and Virgin’s KB was that Vivi could contextually understand multiple questions at a time. We’ve found that people like to ask chatbots a lot of questions all at once. Frequently, these multi-question inquiries end up confusing the bot. With our accelerator, the bot can now research each question with “approved” KB articles and respond in a way humans can understand. By staying on script and contextually responding in a natural way to the questions raised in the chat, Vivi provides a better, more human experience for the customer.
Furthermore, the data extensibility of these solutions built on Salesforce means that we can integrate other data sources easily. Wanting to use Amazon’s LLM, validating case or purchase data, or looking at hotel inventory isn’t a problem.
Another powerful feature of our framework and its design is its ability to write back to these data sources. Let’s say a chatbot referenced a knowledge article but didn’t have sufficient information to answer a customer’s question. The customer would be transferred to a human agent who could provide support and fill the gap. With our framework, the transcript from that conversation would be analyzed by AI, automatically updating the outdated knowledge articles to fine-tune the experience for future customers.
Where do we go from here?
AI isn’t going to take our jobs, but keeping up with its advancements is crucial to keeping up with competitors.
We see huge potential with all things AI-related. Below are a few ways you can start embracing the potential of AI.
- Leverage prompts and features with low risk to improve textbook tasks.
- Invest in data and migrate it to Salesforce Data Cloud to support a few critical functions of your business.
- Invest in forming AI advocates within each business unit so that you’re bringing your teams along.
- Pilot a high-reward product to build momentum and provide a North Star within your company.
Start your GenAI journey today
Our Salesforce GenAI accelerator empowers businesses to create AI-driven bots capable of delivering personalized, dynamic, and efficient customer interactions in a low- to no-code builder experience across the Salesforce Platform and launch within two months. We also have turnkey Data Cloud to (AWS/Google Cloud/Snowflake) accelerators to help you get started. Interested in learning more? Reach out today to schedule a demo.
Slalom is a next-generation professional services company creating value at the intersection of business, technology, and humanity. Learn more or contact us today to schedule a demo.
