The New Word of Mouth Has Gone Digital (and Then Some)

Word of mouth has always been the lifeblood of local business. A glowing recommendation from a trusted mate, a five-star shout-out on a community Facebook group, a handwritten note slipped under the door — these things have moved the needle for small businesses for decades. But the game has shifted again, and this time, it’s not just your neighbours talking about you. It’s artificial intelligence.

Large Language Models — or LLMs, the technology powering tools like ChatGPT, Google’s AI Overviews, and Perplexity — are increasingly becoming the first port of call when people want a recommendation. “Hey, what’s a good accountant in Brisbane?” or “Which café in Fitzroy has the best brunch?” These kinds of queries are no longer just landing on Google’s traditional results page. They’re being answered by AI, and that AI is pulling information from somewhere. Spoiler alert: a huge chunk of it is coming from your customer reviews.

If you’re a small to medium business owner and you haven’t yet thought about how your online reviews connect to what AI says about you, now is absolutely the time. This isn’t some far-off tech trend — it’s happening right now, and it’s quietly shaping whether your business gets recommended or overlooked entirely by the most powerful discovery tools on the planet.

How LLMs Actually Learn About Your Business

To understand why reviews matter so much in this new landscape, you need a rough sense of how LLMs work. These models are trained on enormous datasets pulled from across the internet — websites, forums, news articles, social media, and yes, review platforms. When someone asks an AI a question about local services or businesses, the model draws on all of that training data to construct its answer. It’s not just reading your Google Business Profile in real time (though some tools do have search capabilities bolted on). It’s also working from what it already knows, which includes the patterns, sentiments, and language found in publicly available reviews about your business.

Think of it this way: if hundreds of people have written reviews mentioning that your bakery makes “the best sourdough in Fremantle” or that your plumbing business is “reliable, on time, and reasonably priced,” that language becomes part of the AI’s understanding of who you are and what you do well. Those specific phrases and sentiments get absorbed into the model’s knowledge base. When someone later asks an AI to recommend a reliable plumber or a great sourdough bakery in your area, the AI draws on that accumulated understanding to shape its response.

This is quite different from traditional SEO, where you could optimise a webpage with keywords and meta tags and expect a fairly predictable outcome. With LLMs, you’re not directly writing the content they learn from — your customers are. That shift in control is significant, and it has real implications for how you manage your reputation online.

The Four Things That Make Reviews Count for AI

Not all reviews are created equal when it comes to influencing what LLMs say about your business. There are four key factors that seem to carry the most weight: sentiment, volume, recency, and keyword relevance. Understanding each of these will help you take a more strategic approach to review management — which, frankly, most businesses are still treating as an afterthought.

Sentiment is the emotional tone of a review. AI models are very good at reading sentiment — they can tell the difference between a glowing endorsement and a lukewarm compliment dressed up as a positive review. Consistently positive, enthusiastic reviews paint a clear picture for the AI. Conversely, a pattern of negative sentiment — even if your overall star rating is decent — can influence how the AI frames your business when answering a question. If multiple reviews mention slow service, for example, the AI might note that as a caveat even while recommending you.

Volume matters because it provides the AI with more data points to work from. A business with 400 reviews gives the model far more to analyse than a business with 12. More reviews mean a richer, more nuanced understanding of what your business offers and how customers experience it. If you’ve been neglecting review generation because your star rating is already decent, this is your nudge to keep going — because volume builds credibility not just with humans, but with machines.

Recency is increasingly important as AI tools gain the ability to access more current information. Reviews from three years ago tell a different story than reviews from last month. If your business has gone through changes — new staff, a renovation, a shift in focus — older reviews might not reflect your current reality. Fresh reviews signal to both humans and AI that your business is active, relevant, and still delivering on its promises.

Keyword relevance might be the most interesting factor for marketers to wrap their heads around. When customers use specific, descriptive language in their reviews — mentioning your location, your services, your specialisations — that language becomes part of the signal the AI uses to categorise and recommend your business. A review that says “great café” is less useful than one that says “best flat white in Port Melbourne, with amazing gluten-free options and super fast service.” The specificity gives the AI something to work with.

Why This Changes Your Approach to Reputation Management

For years, the advice around online reviews has been pretty standard: respond to your reviews (especially the negative ones), aim for a four-star-plus average, and try to get a steady stream of new ones coming in. That advice isn’t wrong, but it’s no longer enough. In the age of AI-powered search and recommendations, reputation management needs to be elevated to a genuine marketing strategy — not a box-ticking exercise.

What does that look like in practice? It starts with being intentional about the kinds of reviews you try to generate. When you ask a happy customer to leave a review, give them a little guidance. Not scripted, fake-sounding prompts — but a gentle nudge toward being specific. “We’d love it if you mentioned what service you used and what stood out for you” is the kind of prompt that leads to keyword-rich, genuine reviews. You’re not asking customers to lie or exaggerate. You’re asking them to be specific, which is what great reviews are made of anyway.

It also means thinking carefully about how you respond to reviews — both positive and negative. Your responses are public, they’re indexed, and they add to the overall body of content associated with your business. When you respond to a positive review and naturally weave in a service name or your location, you’re adding useful, keyword-relevant content. When you respond to a negative review thoughtfully and professionally, you’re demonstrating the kind of customer-first culture that AI models can pick up on as a positive signal about your business character.

The Risk of Ignoring This — and It’s a Real One

Here’s where we get a little frank (pun absolutely intended). Businesses that don’t engage with this shift are going to find themselves increasingly invisible in the spaces where their potential customers are spending time. As AI search tools become more mainstream — and they are becoming mainstream, fast — the businesses that show up in AI-generated recommendations will have a meaningful advantage over those that don’t. And the businesses that show up will largely be the ones with the richest, most positive, most specific review profiles.

There’s also a risk on the other end of the spectrum: businesses with consistently negative or thin review profiles may find themselves actively excluded from AI recommendations, or worse, mentioned in a negative context. If an AI model has learned from dozens of reviews that a particular restaurant has a problem with wait times or a particular tradie has a pattern of not following through on quotes, that understanding gets baked in — and it can surface in AI responses even when no one specifically asked about negatives.

The good news is that this is entirely within your sphere of influence. Unlike some aspects of AI search — like what websites get crawled or how algorithms are weighted — the reviews left about your business are something you can actively shape through your customer experience, your engagement, and your review generation efforts. You don’t need to game the system. You need to be genuinely great and make it easy for your happy customers to say so, specifically and publicly.

Practical Steps You Can Start Taking This Week

If you’re feeling a bit overwhelmed by all of this, let’s bring it back to earth. You don’t need a massive budget or a team of AI specialists to start making progress here. You need a plan and some consistency — which, coincidentally, is exactly what we help businesses with at Frankly Organised.

Start by auditing your current review profile. How many reviews do you have across your key platforms — Google, Facebook, industry-specific sites? What’s the general sentiment? Are there recurring themes, both positive and negative? What language are customers using to describe your business, and does it reflect the services and values you most want to be associated with? This audit gives you a baseline and helps you identify the gaps.

Next, build a simple review generation process. This doesn’t have to be complicated. A follow-up email after a completed job, a text message with a direct link to your Google review page, a prompt at the end of a positive customer interaction — these small touchpoints can dramatically increase your review volume over time. The key is consistency. One big push followed by months of silence won’t cut it. You need a steady, ongoing stream of fresh reviews to signal to both humans and AI that your business is alive, active, and continuously delivering.

Also, take a look at your existing negative reviews with fresh eyes. Are there issues being raised repeatedly that you could actually fix? Sometimes the most powerful reputation management strategy isn’t generating more positive reviews — it’s genuinely improving the thing that people keep complaining about. Fix the problem, and the better reviews will follow naturally. That’s not just good for AI — it’s good for business, full stop.

AI Is Listening — Make Sure It Hears the Right Things

We’re at a genuinely fascinating and important moment in the evolution of how businesses get discovered. The tools people use to find local businesses, make decisions, and get recommendations are changing rapidly, and the businesses that thrive in this new environment will be the ones that adapt. Reviews have always mattered. Now they matter in ways that are more complex, more far-reaching, and more consequential than ever before.

The core message here is actually quite simple: take your reviews seriously. Not as a vanity metric or a compliance exercise, but as a genuine reflection of your business and a powerful input into the AI-driven systems that are increasingly shaping your visibility and reputation. Encourage your happy customers to leave specific, descriptive reviews. Respond to every review with thought and care. Fix the things that keep coming up in the negatives. And do all of it consistently, not just in a one-off burst of enthusiasm.

AI is listening to what your customers say about you. The question is: are you making it easy for them to say the right things?

Let’s Get Sorted

If you’re ready to stop guessing and start growing, we’d love to help. Head over to Frankly Organised Contact and let’s get your marketing seriously sorted.