As tech giants race to undertake AI, Salesforce is preserving tempo. Simply as Microsoft CEO Satya Nadella and Google CEO Sundar Pichai revealed that AI now writes about 30% of their firms’ code, Salesforce experiences an identical development. AI generates roughly a 3rd of the code in the present day, with expectations to hit 50% by year-end 2025. That shift has already translated into an increase in productiveness.
“We’re already seeing 30% productiveness enhancements, nevertheless it’s not nearly code technology,” mentioned Muralidhar Krishnaprasad, president and CTO of Salesforce, in an unique interview with AIM, including that the corporate can be utilizing AI for check case technology. He defined that AI can now generate these check circumstances routinely by understanding the code’s logic and person flows. This protects time and helps builders catch bugs earlier within the growth course of.
Earlier this yr, Salesforce CEO Marc Benioff introduced that the corporate wouldn’t rent any extra software program engineers in 2025. He defined that the corporate is ready to accomplish rather more with fewer engineers because of AI-driven automation and productiveness positive factors.
Krisnaprasad additionally highlighted the affect of AI even in post-deployment levels. Whereas automated deployment is one side, AI also can monitor efficiency, detect points, and take corrective actions, equivalent to producing incident experiences. In accordance with Krishnaprasad, this type of AI help spans your complete growth lifecycle—from writing code and producing check circumstances to deployment and post-deployment monitoring.
He added that lately, he now not has to chase down engineers for solutions. As a substitute, he merely goes to Slack, the place all design paperwork and movies are listed. He asks the agent a query—like “How does this characteristic work?”—and will get an on the spot response. “These are immeasurable, brutal enhancements.”
Salesforce is utilizing AI internally, however apparently, it introduced final yr that it will rent a number of thousand workers to promote its AI merchandise.
Information Cloud is the Coronary heart
Krishnaprasad dismissed the narrative that software program as a service (SaaS) is below risk. “Individuals attempt to sensationalise as a result of they don’t have SaaS platform themselves,” he mentioned, taking a dig at Microsoft.
He additional added that each firm wants a sure set of 5 to 6 capabilities that Salesforce supplies, and they’re going to evolve with the assistance of brokers. He added that Salesforce has remodeled from software program as a service to agent as a service (AaaS).
The corporate launched Agentforce 2 in December final yr, providing out-of-the-box autonomous brokers that combine simply with Salesforce Buyer 360 platform, Information Cloud, Slack, Tableau, and third-party techniques by way of MuleSoft and the AppExchange.
Talking of Information Cloud, Krishnaprasad defined that it lets customers combination structured knowledge and index unstructured knowledge. “We additionally built-in MuleSoft in order that any API you have got within the enterprise will now be out there to brokers,” he mentioned. He added that by combining knowledge and APIs throughout the platform, brokers turn out to be “tremendous good” and context-aware throughout domains like gross sales, service, advertising, well being, and automotive, making them genuinely helpful for finish customers.
Information Cloud reached $900 million in annual recurring income (ARR) for fiscal yr 2025. Since October final yr, the corporate has closed 5,000 Agentforce offers, together with greater than 3,000 paid ones.
Krishnaparasad mentioned Salesforce brokers assist with the whole lot from figuring out segments, creating campaigns, producing leads (by way of SDRs/BDRs), nurturing them, and shutting gross sales.
With a view to enhance accuracy and cut back hallucinations, Salesforce makes use of its Atlas Reasoning Engine and Retrieval Augmented Technology (RAG) to ship personalised responses grounded in each structured and unstructured knowledge.
Information Cloud performs a key function by integrating disparate knowledge sources to create a unified view, breaking down silos with out requiring full knowledge duplication.
Past integration, Information Cloud additionally creates unified buyer profiles, serving to keep away from fragmented experiences, like being focused with adverts for merchandise one has already purchased.
Krishnaprasad famous that the platform serves because the foundational system for logging all interactions, together with agent flows and conversations. “We additionally use Information Cloud as our logging substrate for all agent flows. Each single dialog you’re having is saved in Information Cloud,” he mentioned.
Mannequin Context Protocol and Google A2A
Krishnaprasad additional shared that Salesforce is experimenting with the Mannequin Context Protocol (MCP) and has launched new software program that permits clients to transform any API into an MCP with the clicking of a button.
Furthermore, he added that Heroku can now host and run any MCP server, and that the corporate’s Agentforce platform helps MCP. Heroku is a cloud platform as a service (PaaS) that permits builders to construct, run, and scale purposes rapidly with out managing infrastructure.
Salesforce is working with Google on the Agent2Agent (A2A) protocol to set business requirements. The corporate has additionally partnered with Google to combine Gemini AI into Salesforce’s Agentforce, permitting brokers to course of photos, audio, and video, deal with advanced duties, and supply real-time insights utilizing Google Search and Vertex AI.
Reflecting on these developments, Krishnaprasad in contrast the rise of AI brokers to the early days of cloud computing. He mentioned AI brokers are actually at an identical inflection level because the early days of cloud computing. He believes they might ship even higher affect—as much as 10 instances extra—when it comes to productiveness and job creation.
