LEAN TOOLS

HubSpot success kit

Maximize your HubSpot ROI with this all-in-one resource kit:

• HubSpot Savings Guide: See how to save over $10K on HubSpot in your first year!
  
• HubSpot Onboarding Walkthrough: Crush it with HubSpot from day one with our Onboarding
  Guide.
• Five-Star HubSpot Guide: Identify key opportunities and see the steps needed to get your perfect setup.

HubSpot's AEO setup article lists 25 daily prompts for Marketing Hub Professional and for a HubSpot AEO subscription, and 50 for Marketing Hub Enterprise. HubSpot's articles on prompts and recommendations add a requirement: a Super Admin has to turn on several AI settings first, so find that person in week one.

If Marketing Hub sits beside a site hosted in another CMS, the edits happen there. HubSpot's SEO scan still covers externally hosted pages, and its traffic reports cover them once the HubSpot tracking code is installed.

If your account has none of the AEO tools, the loop still runs. You track fewer questions by hand and keep the results in a spreadsheet.

Start with the questions sales already hears

With 25 prompt slots on Professional or 50 on Enterprise, you can't track everything a buyer could ask, so the prompt list is a priority decision. HubSpot generates suggested prompts from your brand details, competitors, products, and ideal customer profiles. Treat those as a first draft. The questions your sales and support teams hear on calls are strong candidates because a real buyer asked them while deciding.

Calls aren't the only place to look. When we run AEO in HubSpot, we build the question set from sales calls, CRM insights, Search Console demand, and competitor prompts. Use whichever of those your team can reach this month, and add the rest as access opens up.

Write each prompt the way the buyer phrased the question. HubSpot's prompt documentation adds a locale point: use the language your buyers write in where they are, so Spanish for buyers in Spain. HubSpot also lets you put prompts in groups. We'd group them by the page that should answer them. Then a change to one page maps to a known set of prompts when you review results.

Suppose a B2B software team on Marketing Hub Professional, with its site on HubSpot's CMS, keeps hearing one question from finance buyers: does the product connect to NetSuite, and what does setup cost? The site has an integrations page that names NetSuite and says nothing about the setup fee or how long setup takes. That question goes into a tracked prompt, and the integrations page becomes its assigned page.

Give each question a page and an owner

Look for an existing page that should answer the question before you plan a new one. Editing that page keeps its URL and the traffic history HubSpot holds for it. Write a new page only when no current page has the job.

Five roles cover the work: a strategist, an editor, a subject expert, a HubSpot or web owner, and an analyst. On a small team, one person can hold several of them. The table shows one month for the NetSuite example.

The Owner column is what keeps the month moving. Each decision belongs to one person, so the NetSuite edit doesn't stall between the marketer who wants it and the engineer who knows the fee.

The subject expert's approval is in the schedule because an answer engine that cites your page repeats what the page says. If the setup fee on the page is wrong, the wrong fee now reaches buyers who never visit your site.

HubSpot can hold the task list. Each AEO recommendation has an assignee and a status you can set to in progress, done, or dismissed, and completed recommendations stay tied to the prompts they affect. For technical work, the SEO recommendations tool has a role filter that separates items a marketer can handle from items that need a developer. Its scan reads page HTML only, so content loaded through scripts, such as CTAs, isn't part of what it checks.

Clear access problems before you edit copy

A page an engine can't reach gains nothing from better copy, so the web owner checks access in week one, before the editor starts. Google and OpenAI each document their own control.

For Google, a page has to be indexed and eligible to show with a snippet, and the site has to be included in Search generative AI features. Search Console has a Search generative AI setting under Settings, with Include as the default. A property with a parent follows the parent's choice by default, and its owner can change that, so confirm the setting on the property that covers your domain.

For ChatGPT search, OpenAI's crawler documentation says a site that opts out of OAI-SearchBot won't be shown in ChatGPT search answers. That crawler is separate from GPTBot, which collects training data, and robots.txt can allow one and block the other. OpenAI says a robots.txt change can take about 24 hours to register for search.

Inside HubSpot, the site audit in AEO recommendations runs two scans. The technical audit checks structured data markup, canonical tags, page speed signals, and crawlability. The content audit scans your pages against your tracked prompts. Run both on your domain in week one, so a blocking issue gets fixed ahead of the new copy.

Keep your own record of answers and changes

HubSpot's AEO dashboard reports brand visibility, share of voice, and the domains cited most often. Brand visibility is a percentage score for how often your brand appears in answers across your tracked prompts. An answer that names you and quotes the wrong setup fee counts toward that percentage the same as a correct one.

So for each question you worked on that month, the analyst runs the prompt and saves what came back: the prompt wording, the engine, the date, the answer text, and every cited URL. Keep a second log for changes, with the page, what changed, who approved it, and the publish date. The two logs together give you a sequence: the fee went live on the 14th, and the answer saved on the 28th quoted it. That sequence doesn't prove the edit caused the citation. It does give you dated evidence to reason from, which a score alone can't.

HubSpot can add to that record for a page you've edited. In a site audit recommendation, the Impact tab lets you track a URL from a start date you set and see how often answer engines reference that page.

Hold the prompt set steady while you compare. HubSpot measures brand visibility across the prompts you track, so adding or deleting prompts can move the number with no change to any page. If you change the competitors HubSpot tracks, only future runs reflect the change, so start the comparison after that date. HubSpot also advises reviewing several days or weeks of results before reading a trend, because engines return different answers to the same prompt.

FREE GROWTH TOOL: AI GROWTH GRADER

Identify your growth roadblocks

Get 3 FREE AI-Powered reports that reveal what’s costing you leads, conversions, and revenue. Then, get expert insights for how to fix it.

• Marketing maturity score:  Find out where your strategy is weak (and how to fix it).
• Website copy critique: Pinpoint the weak points in your messaging (and what to say instead).
• Landing page audit: See what's blocking conversions on your landing pages (and how to solve it).
  

Get your growth score now – FREE!

Review visibility and pipeline as two separate questions

HubSpot reports answer visibility and CRM outcomes in different places, and one can't stand in for the other. For visibility, use the AEO dashboard and your saved answers. For CRM outcomes, open web traffic analysis and find the AI Referrals row. HubSpot's traffic source definitions count visits from ChatGPT, Claude, Perplexity, Microsoft Copilot, Gemini, and several other assistants in that source, and the report shows sessions, new contacts, and customers beside it. Clicking the session count shows which platform sent the visits.

We measure the AEO programs we run on the same layers: AEO visibility, then AI referral traffic, then the contacts, deals, and pipeline HubSpot records. Reading them in that order shows where the chain stops, which tells you what to work on next.

HubSpot keeps regular Google and Bing search in organic search and counts only AI-specific domains such as gemini.google.com as AI Referrals. An AI Overview sits on a regular Google results page, so under that definition, a click from one is counted as organic search. For Google's AI features, use the generative AI performance report in Search Console, which Google's guide describes as the view of how people discover your content through those features.

A buyer who reads an answer and never clicks creates no session at all, so the AI Referrals row can't show that buyer. HubSpot also reprocesses analytics, which can change past figures. Export the report each month if you want a fixed baseline.

High brand visibility with no AI Referrals contacts is a finding to investigate. Look at which pages the engines cite and whether those pages give a reader a next step. If a few referral contacts arrive through a page tied to a question you never tracked, that question belongs on next month's list.

Decide where the first month starts

Every month after the first opens with last month's results. The first has none, so its questions and pages come from the sources you already have: the questions sales and support hear, plus the pages and access findings from week one.

We can help you make that first choice. We run AEO directly through your HubSpot content, website, analytics, and CRM, so the question list and page assignments you've drafted are the right material to bring. Book an AEO strategy meeting with us, and we'll go through where you stand in AI search right now, which AI models mention your competitors, and the quick wins to start getting cited. With those answers, you can choose month one's questions from evidence, and you'll know whether AEO is the right fit for your business.