ChatGPT Just Changed Marketing Forever. Here’s What to Build Now.
ChatGPT Just Changed Marketing Forever. Here’s What to Build Now.
Google CEO Sundar Pichai has called 2027 the year of the AI agent. A few months ago, I predicted that your website won’t matter in three years. Last week, ChatGPT took a massive step in exactly that direction.
After eight years of marketing for home service businesses, I believe this is the biggest first-mover opportunity since Google Business Profiles launched. In this post, I’ll cover what ChatGPT launched, why fewer people will visit your website in 2027, and how to get found, trusted, and booked by AI agents.
The three eras of getting found online
We’re moving through three distinct phases right now, and each one changes how customers find a home service business.
Phase 1: The website era
For the last decade, the path was simple. Someone searched online, landed on a website, and turned into a lead. That model is slowly dying.
Phase 2: The AI search era
About six months ago, we entered phase two. People started using ChatGPT, Claude, and Gemini to find answers. The tipping point came when Google made the biggest change to its search homepage in 25 years, replacing the familiar search button with AI Mode.
That matters because roughly 87% of people use Google to find local businesses. Google has also shared how behavior is shifting inside AI Mode:
- About 63% of users now talk to it by voice rather than typing.
- Around one in five upload photos or files alongside their questions.
- The average search is now about three times longer than before.
This is the shift from keywords to conversations. So the question becomes: is getting found still about ranking for keywords, or is it about being the recommendation? I expect phase two to stay the norm for another 6 to 12 months.
Phase 3: The agent-to-agent era
Phase three is where agents talk to agents, a setup Google calls agent-to-agent (A2A) protocol. When I first described this, people told me I was crazy. Today it’s running.
For it to work, you need both sides of the equation: a personal agent acting for the customer and a business agent acting for you.
The personal side is filling up fast. Beyond Gemini Spark, there’s Muse (all over Meta’s platforms and TV commercials), Grok’s bot, Dots, Duet, Instinct, and others.
You might say nobody uses personal agents. I won’t fight you on that today. But consider this: Google is funneling its huge local-search audience into AI Mode, and Gemini already learns from people’s calendars, Gmail, and search history. When Spark arrives as a proactive agent working on their behalf, those habits carry straight over. Meta, X, OpenAI, and Google will all push their users into these tools whether or not those users go looking for them.
What ChatGPT just launched
The missing piece was the business agent, and ChatGPT just filled it. On November 10, OpenAI introduced businesses in ChatGPT. Think of it as a Google Business Profile, but inside ChatGPT.
Here’s what it includes:
- A claimable business profile. You claim your business and keep its information current. Verification is real: when I claimed ours, I went through a text message check, DNS records, and a government ID upload.
- An insights dashboard. You can see mentions, profile views, and the discovery topics that bring people to your business.
- A business agent builder. After claiming, eligible businesses can build, test, and launch their own agent inside ChatGPT. When someone is chatting, your profile can pop up with a Start chat option so they can talk directly to your business.
The agent builder is rolling out gradually by business type, so you may not have access yet. We don’t on our account. But the capability is live, and it changes how people find and interact with businesses.
The goal is no longer the click
Look at what changed in roughly four months:
- Search shifted from keywords to conversations.
- Personal agents started acting for customers.
- Google rolled out A2A protocol so agents can talk to each other.
- Businesses got their own agents inside ChatGPT.
- Next comes booking, where agents schedule the job directly.
Put those together and the website era is effectively over. Why would anyone visit a website when an agent can hold the conversation and book the appointment?
The goal is no longer the click. The goal is to be the recommendation and own the conversation.
Here’s why this is a first-mover moment: right now, you don’t have a ChatGPT agent. Neither does the company with 10,000 Google reviews, or the one that’s been around for a hundred years. Everyone starts on a level playing field.
To take advantage of it, ask yourself (or your marketing partner) three questions:
- How do I get AI to recommend my business?
- Does AI actually recommend my business, and how do I know?
- How do I train an agent to interact on my behalf?
Question 1: How do I get AI to recommend my business?
If AI never recommends you, nothing else here matters. So does SEO still matter, and does it help you get recommended by AI?
Google answered this directly in its help doc on optimizing for generative AI search. The short answer is yes, SEO is still relevant. But one line in that guidance matters more than the rest: creating unique, compelling, and useful content will likely influence your presence in AI search more than any other suggestion in the guide.
That’s a big deal in the age of AI slop. Anyone can ask an AI tool to write the “best SEO blog post” with perfect title tags, meta descriptions, and keywords. If everyone can do that, it’s no longer a moat.
Create non-commodity content
Google draws a clear line between two kinds of content:
- Commodity content is based on common knowledge, could come from anyone, and adds little unique insight. Google’s example is “seven tips for first-time home buyers.” Most AI-generated content lands here.
- Non-commodity content is helpful, reliable, and people-first. It’s built on stories and real expertise, the things that actually happen inside your business.
Add high-quality images and video
Google also notes that AI search features pull in relevant images and videos. That means more chances to show up beyond a plain webpage link.
This is why I think of SEO now as search everywhere optimization. You earn AI’s trust by becoming the authority, through the content you create across every format.
Question 2: Does AI actually recommend my business?
This is a separate question from the first one, and it’s the one most businesses skip.
Before Phlash, I worked at a home service business. SEO felt like black magic: black hat, white hat, smoke and mirrors. So I understand the skepticism. You shouldn’t take anyone’s word that the work is getting you recommended. You should see it.
We use an AI visibility tracking tool for this. (We have no affiliation with it; it’s just useful.) We build a set of prompts and topics for each client, then measure how often ChatGPT mentions their brand or cites their content.
For one roofing client, the brand showed up in about 78% of tracked queries. A few examples:
- “How much should I expect to pay to fix a roof leak in [city]?” brand mentioned about 75% of the time.
- “Roof repair in [city]” brand mentioned every time.
- Their content, such as a blog post or social video, was cited roughly 25% of the time.
That’s the difference between hoping AI recommends you and knowing it does.
Question 3: How do I train an agent to act on my behalf?
Agents make recommendations based on what they find online: your website, photos, videos, short-form content, and reviews. Once your own business agent exists, it needs to know your business deeply and uniquely.
The obvious answer is “my website will train it.” Here’s my pushback. Most websites were engineered to rank, with keywords jammed into title tags, meta descriptions, and H1s. They weren’t built to show real authority. A site like that isn’t the best way to teach an agent who you are.
Build a knowledge catalog
What you need instead is a knowledge catalog: a complete, unique record of your business that you can hand to any agent. Some call it a knowledge brain or a digital twin. I believe it will be your biggest business asset going forward.
At Phlash, our agent Turner builds knowledge catalogs for clients from six sources:
- Call transcripts. Real customer questions and the way your team actually talks to customers.
- Sales conversations. Recorded sales calls and discovery conversations.
- SOPs and brand docs. Who you are, what you do, and how you do it.
- Your website. It’s one piece, but probably the least realistic, because it was engineered to rank.
- Reviews. What customers actually say about working with you.
- Social content. Authority-building videos and posts, the kind of content Google says AI search rewards.
We use the catalog two ways. First, it trains the client’s business agent. Second, it fuels non-commodity content: we pull real customer questions and real case studies straight from it. That’s how our clients show up so often in AI answers.
Why this is a first-mover advantage
Remember the Google Business Profile gold rush 10 to 15 years ago? It became a race to collect the most reviews. Today there’s a huge moat between companies with thousands of reviews and newcomers with ten. Winning the map pack or Local Services Ads from behind is very hard.
ChatGPT business profiles and agents launched on November 10. Nobody has a head start. The businesses that become the trusted, AI-recommended choice now, and build their agents early, will hold that position as these platforms grow.
And they are big platforms. ChatGPT and Gemini each report more than a billion users.
The old internet vs. the new internet
| Old internet | New internet | |
|---|---|---|
| What drives visibility | SEO, keywords, backlinks | Trusted content and data AI can rely on |
| How customers find you | Search, visit a website, become a lead | AI answers, and agents act on the customer’s behalf |
| What content is for | Ranking, likes, and shares | Training AI to trust you and recommend you |
| Your core digital asset | Your website, the hub everything points to | Your knowledge catalog, deployed inside agents |
| How transactions happen | The customer fills out a form | Agents talk to agents and book the job |
| Competitive advantage | The best-looking website | The most complete, trustworthy data |
Your AI agent readiness checklist
- Claim your ChatGPT business profile. Verify your business and make sure every detail is accurate. (Claim your profile here)
- Prepare your business agent. Even a Google Drive folder or NotebookLM notebook works as a start. Gemini, Claude, and others will likely let you build agents too, and you’ll need the information ready.
- Start building your knowledge base. Gather call recordings, customer questions, FAQs, and SOPs in one place.
- Publish pricing content. Pricing is one of the top questions people ask AI. You don’t need to post your price book. Be the authority on what makes a price go up or down and what factors drive cost.
- Turn real customer questions into content. Pull the top questions from your calls and answer them more thoroughly than anyone else.
- Set up online booking. Agents will increasingly book directly, and you want to be bookable.
- Track your AI visibility. Don’t take anyone’s word for it, including your marketing company’s. Good data leads to good decisions.
Ready for 2027?
The playbooks are changing. If you own a home service business and want a forward-thinking marketing partner that uses AI to help you grow, let’s talk. At the very least, we’ll point you in the right direction.
