Google’s New AI Call for Pricing Feature: First Impressions

If you’ve searched for a local service lately, like a plumber or appliance repair, you might have noticed a shiny new button in the search results: Have AI get prices.

It sounds like magic. You click a button, and Google’s AI acts as your personal assistant, calling around town to get quotes while you sit back and relax. But as we recently found out during a live test in Homestead, Pennsylvania, the reality is a bit more complicated.

Here is exactly how this feature works, where it breaks down, and why the future of local search isn’t about AI making phone calls, it’s about AI talking to AI.

How It Works

The process is designed to be frictionless for the consumer. Google effectively acts as a middleman, collecting your job details and then unleashing an army of Duplex bots (their voice AI) to dial the phones of local businesses.

Step 1: The Trigger It starts in the Local Pack. When you search for a service (e.g., “appliance repair pittsburg pa”), you’ll see the new call-to-action button.

Step 2: The Interview Google needs to know what to ask. Instead of making you type out a paragraph, it uses a quick, structured wizard to gather the essential data points: Appliance type, Brand, and Issue.

Step 3: The Urgency & Location Next, it establishes the constraints. Do you need someone today, or are you flexible? And crucially, where is the job located?

Step 4: The Handoff Once you hit submit, Google confirms they are on it. You get a confirmation email, and the AI starts dialing in the background.

The Reality Check: Why It Often Fails

In our test, the results were a mixed bag. We received quotes from some smaller shops, but major players, who we know service the area, were marked as could not reach.

Why does this happen? The problem isn’t the AI’s voice; it’s the medium of the phone call itself.

1. The IVR Wall

Google’s AI is built to talk to humans, not computers. If a business uses an automated phone menu (Press 1 for Service, Press 2 for Parts), the AI often gets stuck. It doesn’t know which button to press to reach a human, and after a few seconds of confusion or hold music, it times out and hangs up.

2. The Gatekeeper Problem

When the AI does reach a human, it announces itself:
“This is an automated service from Google calling to check prices…”
Many busy front-desk staff instantly hang up, assuming it’s a spam call or a telemarketer. In our test, this likely caused valid businesses to be listed as unreachable even though we could manually dial and connect ourselves.

3. The Rigid Map Failure

We also saw businesses rejected because they don’t service the area.

A human caller might say, “I know I’m 5 miles outside your radius, but I’ll pay an extra trip charge.” The AI cannot negotiate. If the dispatcher says “No,” the AI accepts it as absolute truth and marks the lead as dead.

4. Lack of Detail

A customer using this service cannot add any additional commentary. With appliance repair jobs, it’s crucial to have a model and serial number before dispatching a technician, in many cases it will adjust the overall quote. The current interface is not asking the right questions to ensure the fastest and most accurate repair process. Especially with their own LLM, like Gemini, Google should have no problem letting customers provide as much detail as possible, but that costs money though… which leads to the next point.

5. Actually Calling or Just Guessing?

A customer has no certainty that a phone call took place. In some of our tests, a service call quote did not match when we manually reached out to confirm info. Google will need to provide call transcripts with the actual voice recording for customers to truly trust these results. Businesses also should be getting these call transcripts, likely be best to show in the Google Business interface, because otherwise how can they improve their process to make this feature work better?

It’s also possible, although tough to prove, that Google is not actually calling each business on the list. This stuff does cost money to do and there appears to be no limit to how often a customer can request quotes. The best way to save money is to just send a bot out to grab pricing and availability from your Google Business listing or website.

The Future: Agent-to-Agent Commerce

The current version of this feature is a bridge technology. It’s trying to force a digital AI to interact with an analog-like phone system. It’s clunky, resource-intensive, and prone to error.

The real revolution will arrive when we move from AI-to-Phone to Agent-to-Agent.

The API Vision

Instead of a bot physically dialing a phone number, Google is moving toward a future where your Personal Agent (on your phone) sends a data packet directly to a business’s Service Agent (in the cloud).

Imagine this workflow:

1. Consumer Agent: Sends a JSON request: { "service": "repair", "zip": "44057", "urgency": "48h" }
2. Business Agent (MCP): Instantly checks the company’s live calendar and service territory database.
3. Response: Returns a confirmed slot and price in milliseconds, without a single phone ringing.

Why We Aren’t There Yet

Google is currently building the infrastructure for this via Vertex AI and Google Cloud Agent protocols, but most local small businesses aren’t tech-ready. They don’t have APIs; they have landlines.

Until local businesses adopt smart CRMs that can talk back to Google, we are stuck with the “Robot calling the Human” method. It’s better than nothing, but as our test showed, it’s far from perfect.

Want to ensure your business stays up-to-date with the newest technology trends? Reach out to Fluid today and join us as we help service companies navigate the future!