Google AI Max and Meta Advantage+ Leads are not magic campaign types. They are collections of automation that let the platforms make more decisions about who sees an ad, where it appears, what it says, and how the budget moves.

That can be useful for a home-service business. It can also make a weak setup harder to diagnose. The right question is not “Should we use AI?” The useful question is “Which decisions are we handing over, and what evidence will tell us if that worked?”

What Google AI Max changes

AI Max for Search campaigns adds an automation layer to an existing Search campaign. Its main parts include:

  • Search term matching. Broad-match and keywordless technology can reach queries beyond the keywords selected by hand.
  • Text customization. Google can create or adjust ad text using existing ads and landing-page content.
  • Final URL expansion. Google can send a searcher to a different page on the site when it predicts that page is more relevant.
  • Controls and reporting. Brand and geographic controls remain available, with reporting that includes matched searches, headlines, and URLs.

The expansion is the point. It is also the risk. A roofing company with clean service pages and clear location coverage gives the system better material than a company with one vague page covering every service in three states.

Final URL expansion deserves a deliberate review. Google notes that when it is enabled, pinned responsive-search-ad assets may not be respected. If a specific disclaimer, offer, or message must stay attached to a specific page, check the setting rather than assuming the old behavior remains.

What Meta Advantage+ Leads changes

Meta Advantage+ leads campaigns use AI across three main levers:

  • Audience. Meta can expand beyond the starting audience to find people it predicts are more likely to take the chosen action.
  • Placements. Delivery can move across available Facebook and Instagram placements.
  • Budget. Spend can move across ad sets based on predicted opportunity.

Meta shows an “Advantage+ on” label when the recommended automation is active. Narrowing audience, placement, or budget controls can turn that designation off. That does not mean narrow is always wrong. It means the campaign is no longer giving Meta the same freedom.

For a local operator, the non-negotiable boundaries still matter: actual service area, services the team can fulfill, schedule capacity, honest offers, and a form or conversation that collects enough information to route the inquiry correctly.

Both systems need better outcome data

Google and Meta compete for ad dollars, but their guidance points in the same direction: send downstream outcomes back to the campaign.

Google’s current lead-generation guidance recommends optimizing toward milestones such as a qualified lead, converted lead, booked appointment, or request for quote. In May 2026, Google also announced journey-aware bidding in beta. When a business tracks the lead-to-sale path, the system can learn from more than one stage of that journey.

Meta recommends connecting CRM outcomes through its Conversions API so lead campaigns can optimize toward quality rather than only volume. The setup can send stages such as qualified, converted, lost, or not qualified back to Meta.

The platforms publish performance claims about these features. Treat those as vendor evidence, not a forecast for your account. Your own qualified rate, booked rate, revenue, and capacity are the numbers that decide whether the automation deserves more budget.

A practical control plan

Before turning on more automation, capture the current operating view:

  • campaign and service area;
  • calls and forms received;
  • qualified and unqualified reasons;
  • answer speed and contact rate;
  • appointments or estimates set;
  • booked work when it is available;
  • the pages, messages, and offers currently in market.

Then change one meaningful layer at a time. Give the platform enough room and time to learn, but put a review date on the test. Search terms, landing pages, call recordings, lead dispositions, and booked outcomes should all be in the review.

What to automate and what to keep visible

Let the platform process more auctions than a person ever could. Let it adapt bids, assemble eligible creative, and explore where the evidence supports it.

Keep the business definition of quality visible. Keep service-area and brand constraints explicit. Keep call and form evidence available. Keep a person responsible for deciding whether cheaper leads were actually better leads.

That is not resistance to AI. It is how AI becomes part of an accountable lead system instead of another report no one can explain.