We are constantly bombarded with advertisements and articles touting the global use of artificial intelligence at work. And it’s true — AI offers huge advantages in marketing, recruiting, workforce management and other emerging areas. But while there are critical uses for AI-driven software at larger enterprise companies, it isn’t a one-size-fits-all solution. AI is no small investment, and the decision to put funding toward it shouldn’t be made lightly.

But how do you know if — or when — AI is right for your startup? “AI-based software can do a surprising number of things for the business, from building products to finding customers to everything else,” says Vin Vashista, chief data scientist at Pocket Recruiter. “But it’s only a good fit if you have the right use case for it.”

Here are some things you should consider before investing in AI-driven software at your startup.

Determine if AI is a Good Fit for You

Before investing in expensive AI software, it’s important to know if it will actually be useful for your startup’s needs. For most local businesses — like restaurants or yoga studios — AI may not be a practical solution, Vashista says. “AI is marketed as a panacea,” he says. “But when you’re looking at everyday small business, AI-based products can be cost-prohibitive.” For companies like these, Vashista suggests using Facebook, Google or other sites that have analytic software embedded.

The best candidates for AI are tech-oriented startups that anticipate rapid growth, can afford the cost on the front end and are familiar with ways to use and manage AI-based software. “It has to fit within the business model,” Vashista says.

Select the Right Vendor

Once you have determined that AI software is a good fit for your startup, the thought of selecting a product may be overwhelming. But there are some clear signs of a good product and vendor. For example, a good software will have a proven track record on things like bias and data sourcing. “The vendor that’s selling you the software isn’t liable for those things — you are,” Vashista says. “Your business is liable for any outcomes tainted by bias or unethically gathered data.” It is critical that any software you choose has been vetted and found to be equitable and ethical.

Another track record to look for is return on investment. Vendor representatives should be able to point you to specific use cases for their software. Good vendors will work with you to find the best solution and will be honest if they don’t have the right fit for you. “Watch out for vendors who talk about how the solution helps everybody,” Vashista says. “When they start getting vague about ROI and specific use cases, that vagueness is really a tip-off that something’s not right.” Good vendors are honest about what they can do for you.

Use AI to Support Your Growth

AI-based software’s primary use is in marketing and growth. It can help you determine which markets are easiest or cost the least to acquire, Vashista says. While not everyone should use AI from the start, it might be time to invest once you hit the right growth threshold to move into new markets. “AI can provide insights that analytics can’t,” Vashista continues. “AI can provide a more detailed path to generating additional revenue, increasing average occurring revenue and looking at the right customer metrics.”

As your startup grows and expands into new markets, the ROI on AI-based software for your specific use case is likely to increase. Once you’ve outgrown the embedded analytics offered by Google or other companies, Vashista says, it is usually the best time to consider investing in more robust and detailed AI-driven software.