Highlights
- Salesforce was once a small, nimble company that disrupted the legacy software industry with its cloud-based CRM solutions.
- While it has moved from success to success over the years, the tech world is changing faster than ever. The evolution of AI is challenging the cloud giant’s ability to stay ahead.
- Salesforce is no stranger to AI, both in terms of older machine learning tools and more recent generative AI and agentic abilities. Its Einstein offerings have been around for over a decade and have proven that the cloud giant can adapt to new technology demands.
- AIforce is a newer offering from Salesforce, which is going to be showcased in depth at Dreamforce 2026. This shows Salesforce’s latest response to the AI revolution.
When Salesforce was first launched in 1999, the company had one main goal: to introduce the world to cloud-based CRM and business systems, moving away from legacy tools like Oracle that required on-premises installations and infrastructure management. 27 years later, the company has achieved exactly what it set out to do, so where on Earth does it go from here?
The obvious answer is all around you if you’re even slightly familiar with Salesforce: artificial intelligence. But as Salesforce evolves from a CRM offering into an enterprise AI layer for leading AI vendors, the closer it gets to becoming the invisible (arguably) unglamorous backend technology it originally sought to replace. Now, is this a good thing or a bad thing?
Salesforce’s Original SaaS Mission
The Salesforce of today is a far cry from the Salesforce that experienced its humble beginnings in a one-bedroom apartment atop Telegraph Hill, San Francisco. Not only has rent in the area tripled or quadrupled in the area, but Salesforce has also had its own monumental growth, now being worth approximately $200B.
The purpose of creating this new company was simple: to create business software applications in a completely new way, believing that enterprise software “should be as easy to use as buying a book on Amazon or searching on Google”, according to Benioff. This eliminated the need for multimillion-dollar upfront costs, implementations that could take years, and the ongoing complexities of maintenance and infrequent but painful upgrade cycles.
Not only that, but Salesforce also intended to reinvent the UI and UX of business software, changing the way businesses interacted with their day-to-day programs. Software like Oracle and Siebel sat in the background. Salesforce’s signature interface quickly became well-known – a connected application that wasn’t as clunky, outdated, or complex as its predecessors.
Nearly 30 years later, it’s safe to say that Salesforce has achieved its mission. Not only has it been the number one CRM since 2012, but it also led all CRM vendors with a 20.0% market share in 2025. Its latest earnings results were glowing, essentially reviving a beaten stock, and this week, it will announce a whole new set of features and products (or at least branding) at its annual conference, Dreamforce.
Realistically, where does it go from here?
Venturing Into AI
Salesforce has been working on its AI products and strategy since 2014, launching Einstein – the company’s very first AI product – in 2016. A decade later, the AI products you’ll see showcased at Dreamforce, demoed to you by your AEs, and picked apart on LinkedIn and Reddit are very different from the original Einstein release. But there is no denying that Salesforce has had skin in the game for quite some time.
The launch of ChatGPT in 2023 arguably kicked the entire tech sector into motion, spurring the release of several generative AI solutions in order to fit into this new tech world that had seemingly evolved overnight. Eventually, Salesforce’s solution was Agentforce – its proprietary AI agent platform launched in September 2024.
Agentforce, as a product and a big part of Salesforce’s evolving strategy, has been through a lot since its launch. Its pricing has changed multiple times, its adoption numbers have been questioned, its functionality has been scrutinized, and it was regarded as inaccessible for a while.
However, it has persevered, and this year, Salesforce has taken its AI portfolio to the next level with an extended partnership with Anthropic, the release of several Claude tools, and a handful of AI-oriented acquisitions.
Is Salesforce the Siebel of Tomorrow?
When Salesforce released Agentforce, it was clear that the company was trying to become a real player in the enterprise AI space, and it seemed like it would do anything to get there. The AI race is a brutal one, and the pressure to innovate has only become more intense over time.
Now, Salesforce has reached an interesting position in its journey where its modern AI strategy may actually be taking it right back into the past – becoming the very backend technology that it worked so hard to replace.
Enter AIforce
Now, when we write this, we don’t mean that Salesforce is rebranding itself as Oracle or Siebel, waving goodbye to SaaS. Instead, Salesforce appears to be becoming an enterprise AI layer for other software companies – particularly AI vendors. It started with the growing number of Claude products, particularly Claudeforce. It’s now likely going to be taken a step further with AIforce.
What is AIforce? Well, in part it includes what Salesforce is calling its Trusted Enterprise AI Harness, which brings business data and workflows to agents through MCP servers, APIs, and CLI tools. Although more information on what exactly AIforce is and does is expected to be discussed at Dreamforce, SF Ben understands that AIforce is an umbrella term for all things Salesforce AI, which includes Headless 360 and a new Control Plane for the AI Enterprise. It also makes Claudeforce possible.
What’s particularly interesting is that AIforce is a demonstration of Salesforce evolving into this backend software. Salesforce has quietly built and shipped the enterprise version of exactly that: AIforce, developed over the past three years, ties together the data layer, permission layer, and interface layer so that outside AI agents – Claude first, presumably others later – can plug into a company’s Salesforce data without touching Salesforce’s own UI.
The company that killed “install it and manage it yourself” software is now in its own battle to avoid being branded “legacy technology”.
For months, we’ve been talking about the ‘SaaSpocalypse’ and what it means for solutions like Salesforce. Maybe now we should be focusing less on that and more on whether the new version of Salesforce is good or bad for the people who work with it.
Final Thoughts
The world of technology has always been a fast-paced and ever-changing one, and this rings truer today than ever before. While it was in its infancy, Salesforce had the luxury of being more nimble and adaptable than its larger competitors.
Today, Salesforce is a cloud behemoth that is plagued by the same rigidity that its earlier competitors were. It’s also facing an economic issue in which it has to work out how to rebuild its software model whilst the old one, on many avenues, remains successful. The question is whether or not Salesforce has been able to adapt to the AI landscape fast enough, or whether it will become the hunted rather than the hunter.
All signs point to the fact that Salesforce is taking as many steps as it possibly can to ensure that it doesn’t become the next Oracle. While it is a much larger company than it once was, with many moving parts that need to be orchestrated carefully, it’s clear that they do not intend to let that stop them from evolving with the times.
