#82: Google Maps Gets an AI Upgrade

Google Maps is central to Google’s offline data collection strategy

Author’s Note: Today’s article is a follow-up to November’s piece, Evaluating AI’s Impact on the Customer Service Industry. Check it out!

Last week, Google announced a major update to Google Maps where AI is being introduced into the core experience. By leveraging Google Maps data, which includes hundreds of millions of places and user reviews, you can now receive hyperlocal recommendations for your next trip or night out directly within Google Maps.

I believe this paragraph from the announcement sums up the new value proposition quite well:

“Your results are personalized based on things like places you’ve searched for or saved in Maps to help you get the most relevant recommendations. So when you ask, “My friends are coming from Midtown East to meet me after work. Any spots with a cozy aesthetic and a table for 4 at 7 tonight?” Ask Maps already knows you like vegan restaurants and finds convenient midway spots with vegan options.”

Personalized, niche, and informed. The makings of an ideal recommendation engine.

Before we dive into what this update means for consumers and businesses, let’s backtrack to discuss the problem that I believe Google is attempting to solve.

AI works best for deterministic problems, or in other words, outcomes that can be reasonably predicted. Sure, there is a long tail of possible questions a customer could ask an AI chatbot (like this), but most questions have something to do with a company’s assortment, return policy, or shipping delays, based on the data I see in e-commerce customer service reports. AI trained on a company’s brand voice and tone, as well as historical customer service conversations and order history, can act as an adequate facsimile for a human customer service representative.

That’s why billions of dollars have flowed into companies developing AI customer service agents. You know, startups like Fin AI, Decagon, and EliseAI. They provide enterprise firms with an option to rely on AI for their customer touchpoints across email, voice, and chat instead of relying on people.

But what about B2C, or consumer AI customer service? Now, you’re looking to solve a probabilistic scenario, where there could be many different answers based on slight variations of the question. Questions about any business or place have a much longer tail than questions that could be realistically asked of a singular business.

Last year, Google quietly shared a feature that lets Google call businesses on your behalf. Maybe you want to ensure the local shore store has your size before running over. Delegate that task to Google (assuming the store’s phone number is online).

This is inherently an example of consumer-facing AI customer service (and an example of agentic voice AI). I believe that Google is looking for niche, offline data to leverage to improve their foundational models and I suspect they launched this product to acquire phone transcript data.

Google, OpenAI, and Anthropic will continually need more data to enhance their models. And eventually, they may see diminishing returns from internet data: they’ll need to find a way to harness real-world interactions and knowledge that only exists in the physical world.

Google’s quest to acquire offline data will lead them to back into creating consumer-facing AI customer service: a highly personalized (and action-oriented) recommendation engine that allows people to enhance their offline experiences with their online presence. Google Maps is an excellent venue for this goal given the sheer amount of user contributions (over 500 million contributors, according to Google), and omnipresence of the product.

Today, Google Maps is a transportation app with crowd-sourced reviews of businesses and places. There’s a massive, untapped opportunity to turn people’s shared maps into a social feed. However, I think Google is more focused on acquiring as much data as they can, particularly data that their rivals wouldn’t be able to capture. In doing so, they are creating AI customer service for consumers that can act on behalf of people (i.e., phone calls) and provide hyperlocal, personal recommendations regardless of their location in the world.

Google states quite plainly, “combining our Gemini models with our deep understanding of the world unlocks entirely new possibilities”, which I believe nods to Google’s interest in further enhancing its offline data collection.

Think about this scenario: you’re on a road trip and you need to make a pit stop. To fill up your gas tank and use the restroom.

“Hey Gemini, find me a gas station within 15 miles that has a restroom.”

Gemini will quickly learn people’s common driving routes and the reasons why they make the stops they do. They’ll combine user reviews with prompt intent to better understand why people choose to go to one restaurant over another.

So, Google Maps gets AI, and it’s a massive data play to learn more about how and why people travel to certain locations across the globe. What’s the business implication?

Generative Engine Optimization (GEO) matters for all businesses. Regardless of size.

Apparently, Google Maps has over two billion monthly users, which makes it hard to comprehend how much influence Google has with this product . At those numbers, AI isn’t just for people who sit behind a desk all day. It’s for everyone.

If you’re a local pizzeria that relies on foot traffic in a busy part of town, you’ll probably still be fine ignoring GEO. But if you rely on customers looking up non-branded keywords in Google Maps (“pizza near me”) to find your spot, you’re going to want to make sure you show up in AI search results because Google Maps now provides AI-generated output. This leads into my core argument: since Google Maps will leverage Gemini, optimizing your presence in Gemini will translate into more frequent and favorable recommendations within Google Maps.

Optimizing the pizzeria’s web presence for AI search results (particularly for Gemini results) will be critical for small businesses thanks to Google adding AI into Google Maps. And this goes for more than just pizzerias. Car dealerships, mechanics, and gas stations will all need to consider how people find out about their business in the first place. Adding content like a frequently asked question (FAQ) guide, directions to the dealership on major highways, or that the gas station has a public restroom, are examples of content likely to perform well on an AI prompt within Google Maps.

If you’re building in the GEO space, please reach out. I’d like to chat more about the implications this Google Maps update has for small businesses.

Either way, consumers are the real winners of AI making its way to Google Maps. As hyperscalers compete to win the AI race, we benefit from their innovation. I look forward to using the new features on my next road trip.