A few years ago, I moved to San Francisco for an internship. Originally from South Asia, I had never been to the West Coast before, and in fact had only come to the U.S. for the first time a few months earlier to attend college. On a rare break from work, I spotted a distinctive tall building downtown while out with a friend. “What’s that building?” I asked. “That’s Salesforce Tower,” my friend replied. “It’s a software company that everyone hates.”
I came to learn that Salesforce is both a symbol of tech success and user pain, and I’m still struck by the irony that one of the most disliked products in tech also has the most dominant presence in the San Francisco skyline.
Salesforce was once a scrappy startup taking on a huge incumbent called Siebel Systems. The company became known for its aggressive business practices and even PR stunts, like crashing the Siebel Systems user conference. Yet over the years, Salesforce earned the top CRM spot – its product was simply better. Companies wanted to move from clunky on-prem software to SaaS solutions. But the advent of AI has made Salesforce susceptible to emerging startups. Its technology has fallen behind.
I’ve spent the past two years talking with Salesforce users about what their ideal AI-powered CRM would look like. Here are their top complaints:
Salesforce was designed for managers to track sales activity, not for reps to sell more. Managers love Salesforce because it is designed primarily for them. It allowed managers to track sales activity and success. It’s not designed to help reps sell, and they know it.
Reps spend most of their time updating Salesforce instead of selling. Salesforce’s own data shows that sales reps spend less than 30% of their time selling, and the bulk of the balance on administrative tasks… like updating Salesforce. Reps want a CRM that leverages AI to automatically update fields, follow up and real-time deal updates. Salesforce’s long-awaited AI features were a disappointment. Despite the hype, Agentforce agents average only about a 58% success rate on simple tasks. Salesforce’s own data shows real world accuracy isn’t keeping pace with the marketing gloss.
Salesforce’s AI doesn’t go far enough and it’s often inaccurate. AI can make sales reps’ lives infinitely easier, but Salesforce has not delivered on its AI promise. AI should be able to automate the top of the funnel completely and tell salespeople what to do next in real time so they can spend less time on admin tasks and more time closing deals. It should be able to understand email conversations and create accurate proposals. Instead, many sales reps find themselves having to double-check everything produced, slowing them down instead of helping.
What do reps want? A system of actions instead of a system of records. AI that can be trusted to drive itself, moving prospects along through the funnel automatically so that reps can come in and close when prospects are ready. Much of the early work in the sales process can and should be automated, freeing reps from clunky workflows and tedious manual forms and endless clicking through fields.
While many enterprises are struggling to achieve a return on their AI investments, Sales has been a bright spot: companies adopting agentic AI report an average revenue increase of 6% to 10%, showcasing AI’s tangible impact on sales performance. As a result, 35% of chief revenue officers are expected to establish centralized “GenAI Operations” teams by 2025. But as is common, most innovation is taking place at the margins, driven by startups who can act quickly and are motivated to bring change to a stagnant arena. Sales leaders should be looking beyond Salesforce for the next wave of innovation in CRM.
