Thursday, October 09, 2025
Home Innovation Salesforce Salesforce AI Research Unveils...
Salesforce
Business Honor
28 August, 2025
New tools from Salesforce enhance AI performance, testing, and data quality in enterprises.
Salesforce AI Research has released three important breakthroughs aimed at improving the application of artificial intelligence (AI) in business environments. These technologies are geared towards assisting companies to become agentic AI enterprises, whereby AI agents carry out activities independently, streamlining efficiency and decision-making processes across different departments like customer service, sales, and marketing. One of the innovations includes CRMArena-Pro, an advanced version of Salesforce's previous CRMArena tool. This new simulated enterprise setting provides a way for companies to subject AI agents to real-world, multi-turn situations. These situations involve important enterprise activities such as sales forecasting, service case routing, and Configure, Price, Quote (CPQ) processes. Synthetic data is used by CRMArena-Pro to permit risk-free, efficient testing of AI agents without compromising privacy and security, similar to a digital twin of an enterprise. It enables businesses to model difficult scenarios, such as supply chain outages or customer service escalation, prior to deploying AI agents in actual environments.
Salesforce has also launched the Agentic Benchmark for CRM, a metric that is meant to measure the performance of AI agents on some key metrics such as accuracy, cost, speed, trust, and safety. Of interest, sustainability has been included as a new dimension, which tracks the environmental sustainability of AI systems. This addition assists companies in guaranteeing that their AI models are efficiency-optimized, using the least amount of computational resources. The benchmark tests agents in various business applications, including customer support and marketing, to assess their performance prior to deployment. Account Matching is a new product in Salesforce's Data Cloud that seeks to enhance the quality of data upon which AI agents act. It directly identifies and merges duplicate records and resolves inconsistencies in data systems so that AI agents have access to accurate, combined information to make decisions.