AI Not Increasing Sales? Why Only 31% of Small Firms See It

AI not increasing sales at your business yet? If you've got AI tools running and revenue hasn't moved, you're in the majority. A New York Fed analysis published October 8 found that only 31 percent of small business owners in the survey reported increased sales from their firm's use of AI.

That's not a reason to rip it all out. It's a sign that most owners pointed AI at the wrong job. This post walks through what the data says, why the gap exists, and what to change if you want AI to show up in revenue and not just in saved minutes.

What the New York Fed found

The post comes from economists Will Aarons and Asani Sarkar, using the Federal Reserve's 2025 Small Business Credit Survey. A few numbers worth knowing:

• 46 percent of firms with at least one employee said they or their employees were using AI tools. Another 15 percent planned to start within a year.
• Among users, 63 percent said AI was somewhat or very important to how they produce their work.
• 77 percent reported no change in labor costs from using AI.
• Only 31 percent reported increased sales from AI.

Here's the twist. AI users were far more optimistic about the next year. Among employer firms using AI, the share expecting revenue to rise beat the share expecting a drop by 48 percentage points. For non-users it was 21 points. Even after the researchers controlled for firm size, owner traits and past revenue, AI users still had a 14-point edge in expected revenue growth.

So owners believe AI will help sales. Most just haven't seen it yet.

Why AI not increasing sales is the normal first result

When I look at founder-led businesses using AI, the pattern is usually the same.

AI got pointed at internal work first. Writing emails faster. Summarizing meetings. Drafting social posts. Cleaning up a spreadsheet. All useful. None of it touches the moment a customer decides to buy.

Nobody defined the number. "Use AI more" isn't a goal. "Answer every web lead within five minutes" is. If you didn't pick a sales number before you started, there's nothing for the tool to move.

The tool was bought, then left alone. Someone signed up, played with it for a week, and moved on. AI that nobody owns doesn't do much.

Speed without follow-through. You can write ten proposals in the time it used to take to write one. If nobody follows up on them, you've just made more unread proposals.

The Fed post has one more interesting finding. Among AI users, the ones dealing with technology challenges (websites, social media, ecommerce) had a 16-point higher net expectation of revenue growth than users without those problems. The researchers suggest AI may help firms get past tech roadblocks that were holding back sales. In plain terms: the owners who expect the biggest lift are the ones using AI to fix the stuff customers actually touch.

Where AI does move sales

If you want AI to show up in your revenue, aim it at the steps between "someone is interested" and "someone paid." Here are the spots where I see it work.

1. Speed to lead

A lead that waits a day goes cold. AI can reply in minutes, answer basic questions, and book a call. It doesn't replace your sales conversation. It makes sure the conversation happens. We wrote a full guide on automating lead follow-up if you want the setup.

2. Follow-up that actually happens

Most deals die from silence, not from a "no." AI can draft the follow-up email right after a call, remind you on day three, and write a different angle on day seven. You still hit send. You just stop forgetting.

3. Better ads and landing pages

Paid social is where AI saves real money for small teams. You can test more headlines, more images and more offers, then spend on what's working. The test still needs a human eye, but the volume goes up a lot.

4. The website basics

The Fed finding about tech challenges lines up with what I see. A slow site, a confusing checkout or a contact form nobody monitors will cap your sales no matter how good your ads are. AI can help you rewrite pages, find broken flows, and draft FAQ answers.

5. Pricing and offer tests

AI is good at helping you think through offers: a bundle, a guarantee, a payment plan. It won't tell you what will work, but it will help you write three versions to test.

How to tell if AI is increasing sales or just keeping you busy

Pick one number before you start. Then track it every week.

Good numbers to watch:

• Time from lead to first reply
• Percent of leads that book a call
• Percent of calls that turn into a sale
• Revenue per ad dollar on paid social
• Number of follow-ups sent per open deal

If you can't name the number, you can't tell if AI helped. Our guide on how to track marketing ROI walks through a simple setup that doesn't need fancy software.

A test I like: run a change for 30 days on half your leads and leave the other half alone. If the AI-assisted half books more calls or closes more deals, keep it. If not, try a different step.

A simple 30-day plan

Week 1: pick the leak. Look at your last 20 leads. Where did the ones that didn't buy fall off? No reply? No follow-up? Price shock?

Week 2: point one tool at that leak. Just one. If it's slow replies, set up an instant reply and booking link. If it's follow-up, use AI to draft the next two emails right after every call.

Week 3: measure. Compare this week's number to your baseline.

Week 4: keep, change or kill. Then pick the next leak.

That's boring. It works.

When to get help

Some owners can do this alone. Plenty don't have the time, and that's fine. What matters is that whoever helps you owns a sales number, not a list of tasks.

That's how we work at AI Answered. We're a growth partner for founder-led businesses: paid social, sales systems and AI systems on one team. We agree on the target and how we'll track it before we build anything. We don't do retainers with fuzzy value, so what you pay ties to results you can see in your own numbers. If you're comparing options, our post on growth partner vs marketing agency explains the difference.

Common questions

Why is AI not increasing sales at my business? Usually because it's doing internal work, like writing and summarizing, and not the steps where customers decide to buy. Point it at lead response, follow-up and ads.

Is the 31 percent number bad news? Not really. It means most firms are early. The same survey shows AI users expect stronger revenue growth than non-users.

What's the fastest place to see a sales lift? For most small businesses it's lead response time and follow-up. Those are cheap to fix and easy to measure.

Do I need expensive software? No. Most of the wins come from tools you probably already pay for, set up with a clear goal.

The bottom line

AI saving time is nice. AI growing sales is the point. If you pick one number, aim one tool at it, and track it every week, you'll know within a month whether it's working. If you'd like a team that builds this and gets paid on what it earns, see how our growth partnerships work.