Salesforce is letting customers drive its AI roadmap — and that’s actually smart

Salesforce has been around long enough to know that the best product ideas don’t always come from the C-suite or a strategy room. Sometimes they come from a frustrated admin in Omaha who just wants the damn thing to work.

That’s the thinking behind Salesforce’s latest approach to building its AI features. The company is essentially crowdsourcing its product roadmap — letting enterprise customers submit feature requests, vote on them, and even influence what gets built next.

It’s not a new concept. Salesforce has long had an IdeaExchange where customers can suggest and upvote features. But what’s changed is the scope and speed. The company is now explicitly tying that feedback loop to its AI development, which has historically been more top-down and driven by what the market thinks customers want, rather than what they actually need.

And honestly? I think this is a smarter move than most AI companies are making right now.

Too many vendors are chasing the same shiny objects — agents, copilots, RAG pipelines — without asking whether their customers are even ready for them. Salesforce is effectively saying, “Tell us what’s broken, and we’ll build the AI to fix it.”

That’s refreshing. But it also puts a lot of pressure on the customer base to know what they want. Enterprise users aren’t always great at articulating their needs, especially when it comes to AI. They might ask for a chatbot when what they really need is better data reconciliation. The risk is that Salesforce ends up building a bunch of narrow, one-off features that don’t compose into a coherent platform.

Still, I’d rather see a company take this approach than the alternative: building a bunch of AI features nobody asked for and then trying to sell them as a “platform shift.” We’ve seen that movie before, and it usually ends with expensive shelfware.

There’s also the question of how much influence the loudest customers get. If you’re a Fortune 500 with a dedicated account team, your vote probably counts more than a mid-size nonprofit’s. Salesforce says it’s weighting feedback by relevance and impact, not just company size, but I’ll believe that when I see it play out.

What I do like is the transparency. Salesforce is publishing which features are in development, which are being considered, and which got rejected. That’s more than most AI companies do. It also gives customers a reason to stay engaged — they can see their input actually shaping the product.

This isn’t a silver bullet. Crowdsourcing a roadmap works best when the customer base is large and diverse enough to surface real patterns. Salesforce has that. But it also means the company has to filter out noise, manage expectations, and avoid building a Frankenstein product that tries to be everything to everyone.

Still, if I were a Salesforce customer, I’d rather have a say in the direction of AI features than be handed a black box and told to trust the roadmap. This approach at least gives me a seat at the table.

Whether Salesforce can execute on what customers actually ask for is another question. But at least they’re asking.