It takes a lot of work to manage your Google Business Profile. Between creating and publishing posts, responding to reviews, and updating new business information, you’re bound to run into a problem now and again that requires assistance. Currently, you have to create a support ticket. However, I’ve noticed Google testing an AI chat support option to help users find resolutions to their GBP-related problems.
This is great in theory, but it could also be a sign that human-related support (which has always been notoriously difficult to access in the first place), could be even harder to access, especially if Google finds the AI chat button to be successful.
A New Take on Google’s Long-Standing Idea
The desire to implement an AI Chatbot to assist with support is not a new idea. Years ago, when I was leading the Google Business Profile Support team, I tried (and failed) to implement an automated chat widget for the GBP support website. I’ll spare you the details, but the TL;DR was that it was not good enough to launch.
This latest iteration, which appears to be Gemini-backed, is really just the latest evolution of a system Google’s been trying to implement for a while. The difference between now and then? The rapid development of artificial intelligence and the overall quality of these large language models, or LLMs.
Though Google’s use of AI can be tracked back to the late 1990s, developments in the technology have caused the industry to boom. Just take a look at the newly introduced AI Overview, which rapidly checks thousands of websites to deliver answers in an easy-to-skim (and understand) format, eliminating the need to surf the web. Ah, remember when we all used to say that?
Unsurprisingly, Google (just like other companies around the world) wants to take advantage of AI to improve their efficiency. In fact, a study conducted by Exploding Topics found that 85% of companies have made AI a key component of their business plans. When it comes to GBP management, Google believes that AI support will reduce the time it takes to resolve the most common questions and problems its users have.
From my perspective, this isn’t an isolated experiment. This is Google creating a blueprint. They are likely using GBP support as the pilot program for an AI-driven support channel that can be scaled across all their products. The end goal is to make AI the first, and perhaps only, line of defense for users, drastically cutting the company's investment in costly human support.
Reward Does Not Come Without Risk
It’s clear that Google is only in the testing phase of this new AI chat support feature. The fact that the AI chat button shows up on my profile intermittently is proof of that. To avoid hallucinations (AI-generated information that is false or contains misleading information), Google is taking the time to train the data both internally and externally. In layman’s terms, they're making sure the data is clean, or free from errors. Plus, if Wolf River Electric’s lawsuit against Google for publishing incorrect and misleading information that damaged their reputation tells us anything, they’re also probably taking the time to invest in extra training to avoid liability. Read more about that case here.
The problem with LLMs has always been trying to train the data to be accurate. These systems still require hands-on human development, so if they are fed incomplete, biased, or false information, then they’re going to dole out results containing just that. Furthermore, if it isn’t trained with the right data, the AI system will try to fill in the gaps with what it believes to be relevant. When people complain about AI, it’s likely because they encountered these hallucinations, which presents incorrect information as factual. Whether leading to reputational harm or financial risk (like with the Wolf River Electric case), hallucinations create real headaches for users.
And the same could be the case for GBPs.
Hallucinations could make it difficult for the AI support system to offer the correct solution, especially if it cannot correctly identify the problem. Furthermore, if your GBP contains false information as a result of a hallucination, the AI support system may deny your support request altogether, leading to an even bigger headache. In cases like this, the support needed to resolve issues directly with the AI system could mean more complex cases requiring manual review; aka the very thing Google aims to disband with the rollout of this new feature.
Final Thoughts
With so much at risk, Google is taking the time to thoroughly train the LLM to avoid hallucinations because it has the potential to lead to more issues. However, once they get the kinks worked out and see a return of investment on it, expect to see even more challenges in getting the human-related support you need for GBP management.
Bottom line, if they are already testing this new feature out, it means they’ve already started to scale back the human support they offer. If you’ve dealt with them before, you already know that getting to access actual people to help you on Google support is pretty tricky. You have to know how to work the system to be able to get a person (this blog offers my personal tips to do just that). If implemented, the AI chat support button will be a game changer in more ways than one.
Time will tell if and when this new feature gets the official green light. Either way, it shows the direction Google continues to move and how GBP managers will, yet again, have to adapt.
