How To Figure Out What Customers Care About

Featuring my favorite two-step AI workflow

Here are two ways to figure out what your customers actually care about.

One is to do what everyone in B2B sales and marketing has done forever, which is to schedule a bunch of “user research calls” where you ask people leading questions, nod thoughtfully while they tell you what they think you want to hear, then go back to your team and argue for the rest of the quarter about what the customer really meant when they said your product “needs better analytics.”

The other way is to just... gather up what people are telling your sales team right now, feed it into an AI, and ask the AI to tell you what those people actually care about.

This second approach has the advantage of being (a) much faster and (b) based on things customers said when they didn’t know they were being “researched,” which turns out to matter a lot because people are much more honest when they’re complaining to a salesperson about their actual job than when they’re being asked by a CMO or product manager to cogently describe their ideal workflow, which, by the way, is something they are almost definitely thinking about for the very first time as they are trying to sound smart in front of someone they just met.

Everyone in B2B sales, marketing, and product knows, intellectually, that you’re supposed to “talk about customer pain points” and “lead with outcomes not features.” There are approximately one million blog posts out there about this. I even have a couple. Everyone reads this stuff, and everyone nods along. And then they all go back to their desks and write landing pages with stuff like “AI-powered analytics platform with real-time dashboards” on it because talking about features is so much easier than figuring out what actual human problems you can solve.

Why is it easier? Because to talk about problems convincingly, you need to actually understand what the problems feel like to people, not just what the problems are. You need to know that when your customer says “we need better reporting,” what they actually mean is “my VP asks me for numbers and I have to manually pull data from four different tools and when he does that I get acid indigestion and I’m worried I’m going to get fired if I send him the wrong number again.” That’s a different thing! And understanding that difference (and then linking what your product can do to those visceral “you’re reading my mind” challenges and pain-points) is literally how you win.

Which brings us to the AI topic. And the exact workflow I’m using with pretty much all my portfolio companies right now to stop talking about features and start talking about the problems we can solve for people.

My Favorite AI Workflow

Step 1: Boot up your AI. Load in transcripts of 10 customer calls with sales and your best product marketing stuff (pitch deck, talking points, messaging guidelines, etc.) Then ask AI to build you a table.

Column 1 = Problem/challenge (as articulated by customer)

Column 2 = Negative impact on them personally

Column 3 = Negative impact on their business/team

Column 4 = How we can help (instead of X we can help you Y)

Finish the table, then ask AI to prioritize the rows based on how often and how intensely the pain/challenge is mentioned in the call transcripts.

Now you’ve got something. Now you’ve got a stack-ranked list of the biggest, most frequently-mentioned problems all your target accounts probably have. Basically, you have everything you need to be a customer mind-reader.

Step 2: Then ask the AI to write whatever sales/marketing/customer-getting stuff you want, using what’s in the table. I’m talking stuff like:

• Linkedin posts for your CEO.

• Pages for your pitch deck.

• Prospecting sequences.

• Discovery questions.

• Landing page copy.

• Webinar outlines.

• Sales talk-tracks.

Doesn’t sound like you? Don’t like the tone? Train it on your preferred writing style and update your settings. Then try again. Refine and finish.

Congrats. You just did a month’s worth of customer-getting work in less than a day. And now you immediately stand out to your prospects because you can guess correctly re: what they’re going through (and can position what your product does without feature-dumping like your competitors do).

Final Thoughts

The really amazing part about this is not that AI can write your sales or marketing copy. Everyone knows that’s a thing now. The really amazing part is that it forces you to actually organize your customer intelligence in a way that most companies never do. Most B2B companies now have tons of recorded sales calls stashed away, and maybe only a few people have listened to some of them, and there’s definitely a Slack channel where sales occasionally shares “interesting feedback,” but none of that stuff is structured or prioritized or accessible to the marketing team.

So what happens? Marketing writes copy based on what they imagine customers care about. Sales has conversations based on what customers actually care about. These two things slowly diverge until your website is talking about “enterprise-grade security” while your prospects are asking your AEs “can you just help me not have to rebuild this report every Sunday afternoon please.”

Anyway. The AI doesn’t solve this problem by being smart. It solves it by being patient and methodical enough to do what your short attention span would never allow you to do: read all your sales call transcripts. It also solves it by being a little dumber than you. Because it’s less familiar with your work and your product than you are, it forces you to be concrete and literal. It won’t let you get away with vague handwaving about “customer needs.” You give it the transcripts, you force yourself to structure the output, and suddenly you have something that looks suspiciously like actual, useable customer research.

So, try it. Go ask your sales team for 5-10 good call transcripts. Run the workflow above. Build some customer-facing stuff with it. Notice how much more human and clear and relevant you sound. Think about how much better people are going to respond to that stuff when you put it in front of them.

Now go be nice to your implementation team.

They’re about to get a lot busier.

---

Still figuring out how your team can put AI to work? There’s lots more of my favorite AI workflows (and tips for how to build an AI-enabled GTM team) at my AI 101 page here.