5 questions to ask AI vendors before buying a tool

There are many ways to use AI in marketing, and it feels like for every smart plan, 10 AI vendors have come up with a tool to deal with it.

At the beginning of this wave, I took more phone calls and answered more emails than these days. Over time, I realized that I was asking vendors a few of the same questions to check that their equipment was suitable for shipping.

If you’re in the same boat and overwhelmed by approaching salespeople, here are five questions to help you determine if it’s worth your time, along with my reasoning for asking them and what I’m looking to hear — or not hear.

This question should help you understand the purpose of the tool and – in particular – how the value it creates maps to real business results.

If a salesperson can’t clearly articulate the challenges or use cases the tools address, they weren’t purpose-built to solve the real problem your team is facing, whether you’re in-house or working for an agency. Beware of salespeople who try to impress you with heavy language but can’t explain the business benefits those features bring.

If the salesperson identifies at least one team problem that the tool solves and explains how it improves business results, it’s a good idea to keep talking. A good follow-up question is to ask for a case study that shows how the tool was used and the results it delivered for an organization similar in size and vertical to yours.

Look for benefits such as “increase output” or “identify gaps in tracking to speed up troubleshooting.” But don’t rush to invest in tools that promise to “save time” (even if they really do) unless you have a plan for how you’ll use that extra time.

The answer to this question should tell you whether the vendor built the device for advertisers or just of advertisers.

Technical chops are important, but so is understanding how the media consumer spends their day. If the marketer does not have personal experience in media buying, they must explain how they researched the media buying market and included that information in the tool.

If they have a shallow understanding or no expertise, that’s a red flag. It’s okay if a lawyer doesn’t have that expertise directly, but someone on their team does, and you should be able to reach that person as quickly as possible if you intend to continue speaking.

When a salesperson has a story or background that led them to identify a problem that you can’t understand firsthand and decide to create a solution, it’s compelling. A creative intent that addresses your existing challenges is a solid foundation for a tool that can make a difference in your team’s performance.

3. What lessons, real use cases, and results can you share?

I touched on the lessons of several sections above, and they are essential in a new, rapidly developing industry. I will be looking to understand if the seller has an attractive track record with customers like me or if we can be early adopters.

If you fall into the latter camp, there are pros and cons, as with any advertising beta you can get to it before your competitors. You might get a leg up by finding an important growth accelerator before your competitors do, you might spin your wheels trying to exploit bugs, or you might find that a tool doesn’t live up to its promises.

If you can’t trust the tool or there’s a risk you’ll need to provide detailed feedback to make it work, it might not be the best way to spend your time and money unless you believe it could be a game changer.

If you’re obviously going to be an early adoptive parent and the seller isn’t willing to flex on contract terms that help reduce risk, that’s a non-starter. Most established instruments will likely deliver more value consistently, although they will have less wiggle room as a result. New tools that take a hard line on pricing and contract terms, however, are unlikely to be good long-term partners.

For established retailers, you need to see specific, comparable examples with real numbers from marketers in the same space, the same size, or the same use case.

If they’re startups, the best answer is to be honest: “You can be one of our first customers in this vertical. Here’s what we’ve seen elsewhere, here’s what that relationship will look like.” That is obviously a green flag.

4. Who owns my data, and how is it used to train models?

It’s interesting how easily people are sharing data with AI and AI tools in the race to gain a competitive edge. This is something I would strongly caution potential buyers to consider before signing anything.

View any feedback that suggests your data is being used to train shared or third-party models without your express consent. Another red flag is vague or misleading answers or terms of service that contradict or misrepresent what the seller is verbally telling you.

You own your data, full stop.

The vendor must be able to clearly explain its data management practices, including where your data is stored, how long it is stored, or used for model training (and, if so, to improve your situation only), and what happens to your data when you stop using the tool. This needs to be in the contract, not just a verbal guarantee. If it is not there, insist that it is installed before you sign.

5. What does implementation really look like, and what does success require from our team?

Before you make money, you need to understand the true cost of adopting this tool. Those costs include more than just the price. Time, internal lift (including integration, training, and QA), and any potential disruption to your existing martech stack.

If this requires resources that your team doesn’t have, or if you honestly can’t devote the time to getting the tool right, it’s not worth the investment yet. Many marketers can avoid spending money on martech if they ask this question and think carefully about the answer.

No tool should be one-size-fits-all, but ease of use and intuitive design are essential for your team to adopt and stick with it.

Don’t let the AI ​​hype rush your decision

I know firsthand that many of these tools sound too good to be true – and they often are. You need to balance the desire to grow and the curiosity with a little caution.

Remember that we are still in the early stages of AI adoption. If the tool seems too expensive or difficult to get into, or the contract is tighter than it should be given the work record, a better solution will probably emerge in the next few months.

When in doubt, request a free trial. Assuming it won’t cause too much work for your team to integrate the tool, that could be the right next step in determining whether you’ve found the next competitive advantage.