Dreamforce 2023: “It’s not magic, it’s real.”

After Salesforce Dreamforce 2023, it’s time to share some fresh thoughts about the 2023 vintage.

Forward Looking Statements at Dreamforce 2023

If you joined any session (on time), you couldn’t possibly have missed the infamous Forward Looking Statements slide from Salesforce. Legal teams spent hours, days and weeks preparing this language — which nobody could actually read when presented, usually for just a split second.

The biggest takeaway from the “Forward Looking Statements”: do not make a purchase decision based on the future state of the product, but based on what is possible today.

Sounds simple, but this can be quite challenging when what is real cannot be identified by buyers who don’t have a deep level of familiarity.

This year, these statements have a special flavor after the turbulence recently raised in regards to Dreamforce 2022.

Dreamforce 2023: A Dream With Two Faces — Rabbit (Data Cloud) And Einstein (AI)

2022 was the year of real-time through a Genie that has since been put back into its lamp. Well, at least the name Genie didn’t survive for Dreamforce 2023.

When a soccer team isn’t performing, the “solution” is usually to change the coach. At Salesforce, when a product is struggling, the solution is to change its name.

What was recently Salesforce CDP or Genie is now Salesforce Data Cloud. And Data Cloud is clearly presented as the central platform to underpin all the Salesforce solution ecosystem in the future, very much the way Adobe has been presenting and pushing its Adobe Experience Platform (AEP) for the past few years.

On the other side of the spectrum is a more familiar face: Einstein. Previously presented as the framework bringing AI capabilities to Salesforce solutions, it’s now completely front and center.

Data Cloud — with organization’s data — and Einstein — with AI/Generative AI capabilities — are combined to form the newly announced Einstein 1 Platform.

You got it, it was all about Data and AI at Dreamforce 2023! However, when Ali Ghodsi, CEO at Databricks took the stage at Moscone Center to present AI/GenAI product, strategy and vision in June, it came with warnings. Ghodsi was clear: we are at the beginning of a long journey. Marc Benioff took a different stance: it’s here and it’s not magic, it’s real. One CEO is saying it will take time to realize, and the other is saying it’s ready to go — leaving us scratching our heads.

Data Cloud For Free

You may not have to actually pay for the Data Cloud, but what you do have to pay for is… socks. If there is a company who can give away enterprise software, at the center of its growth strategy, it seems to be Salesforce. Sure, no educated buyer will believe in free food for all. But it’s enticing, isn’t it?

One could wonder if Salesforce will make up its revenue from Mulesoft socks, available at $20 a pair! Monetizing socks — a hit since in-person conferences restarted — could be visionary. Unfortunately, you won’t be able to pair that with Gucci “cloud” sneakers, these were only for Marc.

Zero-ETL is not Zero-Copy

The Salesforce integration with Snowflake has been (re)announced: a zero-ETL connector between Salesforce Data Cloud and Snowflake — dropping the term zero-copy used last year.

There’s a key difference in the terminology. With a data sharing integration, referred to as “zero-ETL”, the integration efforts and costs are reduced, streamlining the data copy process between both systems.

However, for IT and data teams, data copies themselves are a real problem. Buyers should keep in mind that a truly composable CDP solution enables a zero-copy approach by integrating via a query pushdown — also known as Connected App in Snowflake’s terminology, such as what is provided by ActionIQ via its HybridCompute solution.

Back to Salesforce Data Cloud and Snowflake Data Cloud integration, I recommend this great article to decipher the technical details. Keep in mind that the connector is still in development, and the availability will also depend on the region of your Snowflake environment.

As great as Snowflake is, clients are also investing in alternative solutions such as Databricks and Google BigQuery. The integration with Databricks was publicly announced in the press last week. On the ground, I couldn’t find any Salesforce employee aware of this integration or announcement.

Non-Snowflake customers, please be patient or look for alternatives outside Salesforce Data Cloud.

Salesforce Data Cloud Deployments Are Unlike Any Other Salesforce Product Deployment

For those who joined Dreamforce 2023 to learn more about success stories in deploying Salesforce Data Cloud, the trip might have been wasted.

Most stories were very informative, but organizations speaking up were very much at the beginning of the journey. One financial services client spent year one to POC the solution and year two to implement it. Fingers crossed that the use cases deployed from there provide value.

The IT leader from that organization did warn people in the room: you need to carefully assess what data you want ingested in the Salesforce Data Cloud. You don’t want to overthink it, but at the same time you need to prepare for future use cases… In his case, he only expects a 10% subset of what has been loaded in Salesforce Data Cloud to be used.

Finally, Salesforce Marketing Cloud was described as “easy to deploy” while Salesforce Data Cloud “is hard.” Make sure that your experience with other Salesforce products doesn’t give you unrealistic expectations — apparently “forward looking” applies to deployment.

Artificial Intelligence Named Einstein

Credit to Laura J. Alber, CEO at William-Sonoma, who regrets the conversations around artificial intelligence and suggests dropping the term “artificial.” She is right: organizations should strive to produce authentic experiences for customers and not artificial ones.

Salesforce gave a big stage to talk about AI and generative AI at Dreamforce 2023. It’s been fascinating to see the reactions of software companies when it comes to generative AI. While many companies are investing and trying to find the right approach, Salesforce seems to have figured it all out already with many generative AI capabilities announced available in pilot this month (September 2023) and GA as early as next month (October 2023).

Among the use cases presented, all the usual suspects were there such as content creation for marketing (email subject line, the email itself, images and more) — very similar to what was presented by Adobe during the recent Summit conference.

Einstein Co-pilot goes a step further to support services teams, providing suggested messages to send to a client during a conversation. The end goal: accelerate the resolution of customer service requests. The Einstein Trust Data Layer, positioned between the user prompt and the LLM who will provide an answer to the request, is a great move to reduce risks in feeding these models with sensitive data which should not make their way to the model.

But allow me to “ground” the excitement which can easily be seen from these demos. Demos are great visions of the future and will perform well in narrow situations today. Even when they end up being decently performant, many questions remain: trust obviously, as clearly mentioned by Salesforce, but also the cost. Cost conversations seem to remain taboo in many places today.

I also wish I saw more clarity around the maturity of each capability presented, all appearing as ready for prime time. And clearly, looking at the lines forming early for each AI session, people were ready to listen in!

If you are interested in our perspective and research, we recently published a series highlighting the opportunities and challenges, as well as the application of Generative AI for the Audience Segmentation use case.

No Space For Acquisition Marketing at Dreamforce 2023

Did you notice? Salesforce DMP (formerly Krux) is no more. Data Cloud conversations have all been about the importance of centralizing your customer data at one place (in Salesforce Data Cloud), but I didn’t hear stories around advertising use cases and customer acquisition needs. Surely, Martin Kihn brought it up in his session — a former Krux executive cannot forget its roots — but outside of that hint, there was silence radio around how Salesforce could support the early interactions between a brand and its prospects/customers.

Don’t Live in a “Forward Looking” Dream, Live in Reality

Salesforce did it again. A great show, on-stage and outside the stage at Dreamforce 2023. But respectfully, the few logos shared during the strategy and roadmap session are here to remind us that, in the past few years, Salesforce CDP/Data Cloud did everything but gain trust from the market and from organizations. And trust, as Salesforce said many times throughout the week, is a critical and invaluable asset.

It’s undeniable that Salesforce built a CRM with limited competition and acquired great tools such as Salesforce Marketing Cloud (ExactTarget), Tableau and Slack. When it comes to the customer data foundation for CX, you will have to remain patient if you decide to go with Salesforce. A sweet deal to sign might quickly give you a sour taste.

ActionIQ’s composable CDP has been powering IT, data and marketing teams, in complement of existing Salesforce investments. Check out our guide — Increase the Value of Your Salesforce Stack With ActionIQ.

If you’re looking to get started with a CDP that can scale a comprehensive, enterprise CX strategy with your business, we invite you to reach out to our experts to talk stack.