Start of Main Content

Most marketing teams can now see whether their brand shows up when someone asks Claude, ChatGPT, Gemini, or most any LLM about their category. Fewer can say which of those answers matter to their business. That gap limits what content optimization can do. A team that sees a visibility gap but can't see the margin, pipeline, or customer value behind it is only running half of a learning loop: it can change content, but it can't tell which changes were worth making.

Velir x Brooklyn Data approaches this from two sides: data and digital capabilities. Both practices drive towards the same conclusion: AI visibility pays off when it is connected to the customer, product, and commercial data a company already owns, and building that connection is a smaller project than most leaders expect.

AI assistants are now a research channel large enough for a CEO to care about, and the visitors they send convert better than average.

Generative AI platforms averaged 9.5 billion web visits a month from June 2025 to May 2026, up 70% year over year (Similarweb). Counting mobile apps as well as the web, Graphite puts AI sessions at 56% the size of search worldwide and 34% in the US, with US AI sessions up more than 300% year over year (Graphite). Graphite also finds that combined search and AI usage has grown since 2023, so AI is adding a channel alongside search.

The traffic AI sends has changed character. AI-referred visits to US retail sites converted 60% better than other traffic in July 2026 (Adobe, via Digital Commerce 360). In March 2025, the same traffic converted 38% worse (Adobe). Growth is slowing as the base gets bigger, from 393% year over year in Q1 2026 to 62% in July, and AI is still a small share of most sites' traffic. What changed is the value of each visit.

For a CEO, that makes AI discovery a competitive question. In a Similarweb study, people whose ChatGPT answer named a brand were 2.5 times more likely to visit that brand's site within seven days than a competitor's (Search Engine Land). B2B buyers start there too: Forrester's 2026 buying study finds that genAI search is where B2B buying begins, in decisions that typically involve 13 internal stakeholders and nine outside influencers (Forrester). The question for leadership moves from "are we visible?" to "which of these conversations are worth winning?"

An AI visibility platform tells you where your brand appears in AI answers. Without the business data behind each answer, content optimization runs on an incomplete learning loop.

Scrunch, for example, reports on each prompt a team tracks: whether the brand appears, its position and sentiment, which competitors appear instead, which URLs the answer cites, and which AI agents crawl the site. The cited pages are where action starts, whether through traditional content optimization or Scrunch's Agent Experience Platform (AXP), which helps make content more accessible and consumable by AI agents.

What the visibility data can't tell you is what each answer is worth. It doesn't know your margin, your customer value, your pipeline, whether the personas behind a prompt match the customers you want, or whether a change you made worked. A raw share-of-voice score counts a prompt about a discontinued product the same as one about your best-selling line. A team working from that score can close a gap and never learn whether closing it paid.

Two examples show how business context changes the priorities:

  • Intent. On transactional queries with an AI Overview, cited brands roughly doubled their organic click-through rate over 2025, from 0.7% to 1.7% (Seer Interactive). Your own conversion and pipeline data tell you which prompts sit close to a purchase.
  • Product accuracy. Retail product pages averaged 66% machine-readable in Adobe's 2026 benchmark, and 66% of US consumers believe AI tools give accurate results (Adobe). Wrong answers about your products tend to be believed. Catalog and margin data tell the product team which errors cost the most, so high-margin items get fixed first.

None of this makes visibility tools less useful. They provide the signal, and the business data tells you which readings matter.

When executives hear "connect your data" or "build a data platform," many picture a multi-year transformation program. For AI visibility, the starting point is data most companies already have.

Five source systems contain most of what's needed: the CMS page list, web analytics, CRM or order data, the product catalog with margin (either in the ERP or an extract), and the AI visibility feed. The joins happen on pages, topics, and products, so nothing requires identifying individual people, which is often a major culprit for data project slowdowns. By connecting these disparate sources, every visibility gap carries a revenue and margin figure, and the content team gets a ranked list where each item comes with its business case.

How long this really takes depends on the maturity level of your data estate. A team whose data is already well modeled can see its first ranked list in about two weeks. A team with no centralized, connected data should plan on one to two quarters. Knowing which of those states best describes your organization is the first step. If you aren't sure, a 2-4 week data maturity assessment can answer that.

The harder part is organizational. Marketing, product, and analytics have to agree on who drives the initiative forward, and that decision matters more than any tool choice. Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data (Gartner). In its 2026 CMO Spend Survey, CMOs put 15.3% of their budgets into AI, but only 30% said they were ready to scale it, and Gartner warned that CMOs "invest in AI tools faster than they build the data foundations" (Marketing Dive).

The investment also pays off outside AI search. The same connected customer, product, and commercial data supports pricing, merchandising, and campaign decisions across an organization.

One step shouldn't wait for any of this: start archiving your AI visibility history now. These tools keep limited history, and trend analysis needs months of it to be most effective.

Once visibility and business data are joined, the same foundation can answer different questions for a number of leadership personas. The scale of the payoff comes from the reusability of the insights generated.

  • CEO: AI share of voice weighted by revenue and reviewed quarterly, so the board sees which competitor is winning the conversations that carry money.
  • CMO: Content budget moves to topics where business value and visibility gap overlap, and every reallocation arrives with its business case.
  • Product: The team sees where AI poorly describes products or leaves out key information, ranked by margin, so that the costliest errors are fixed first.
  • Sales: AI answers to buying-stage prompts are matched to CRM win and loss reasons, so reps are briefed on the claims and objections buyers heard before the first meeting.
  • Content and digital: The team leverages AI-powered workflows to identify content gaps, optimize existing pages, and improve how AI agents access and interpret content, while aligning human experiences with the expectations established during AI discovery.
  • Communications: Earned media made up 84% of AI citations in Muck Rack's May 2026 study (Muck Rack), so many of the sources behind an answer sit outside your site. PR can target the outlets AI reads for priority topics.

The same set of insights-ready data serves all six teams, and each team's results feed the next round of priorities. The outcomes land in the systems those teams already watch: leads and pipeline in the CRM, engagement and conversion in analytics.

 

Ready to Build Your Content Learning Machine?

Velir x Brooklyn Data works with teams on both sides of the loop: where your brand stands in AI answers, and how ready your data is to support decisions like these. Let’s start the conversation.

Together, three pieces form a closed learning loop: Scrunch identifies visibility gaps and improves how AI agents consume content, connected business data helps prioritize opportunities and measure impact, and SitecoreAI enables teams to activate those insights through AI-assisted content creation, optimization, and publishing workflows.

Scrunch Loop

The loop has five steps: observe, prioritize, act, measure, and learn. Scores re-run monthly because AI platforms change without notice. One ChatGPT update in May 2026 moved the share of its referrals landing on homepages from about 25% to nearly 60% in three weeks (Similarweb, summarized by Salience), which reshuffles which pages matter.

The loop works only if it runs as an operating workflow with owners and a schedule. In our work, that means joint project management between the data and digital teams, clear objectives agreed at the start, and several stakeholder touchpoints each week. That rhythm keeps both teams aligned and raises the odds that the work they ship has an impact.

AI doesn't send clean referral traffic, so measurement has to capture influence that arrives through other channels without claiming more precision than the data allows.

The gap shows up in the reports most teams use. GA4 added an AI Assistant channel on May 13, 2026, but it counts Google's AI Overviews and AI Mode as Organic Search and has no history before that date (Magnet; Something Inc.). One buyer who found you through an AI assistant can show up as AI Assistant, Referral, Organic Search, or Direct. In Similarweb's study, 55.9% of the visits that followed a ChatGPT recommendation arrived through branded search, where they look like ordinary search traffic (Search Engine Land), and only 6.8% of ChatGPT answers include citations at all (Salience).

Start by defining success in three measures: time saved, the measure teams have relied on for years; revenue or MQLs generated; and how far the result scales across topics and teams. Agree on the threshold with the CFO before the first change ships.

Then measure in tiers:

  • Weekly. Presence, citations, and sentiment for the topics you changed.
  • Monthly. AI Assistant sessions, branded search impressions, and self-reported attribution, with "AI assistant" added as an option in how-did-you-hear fields.
  • Quarterly. Pipeline or revenue for the topics you changed, compared with similar topics you left alone.

In SitecoreAI, every change ships with a test ID, and the control pages and readout date are set prior to go-live. Report results as ranges with a confidence level. Similarweb describes its own 2.5x finding as a correlation, and you should treat your internal numbers with the same caution. If changed topics don't beat their controls by the readout date, the scores change and so does the plan.

Three moves get your company started in the next 30 days:

  1. Name an owner for AI discovery outcomes, pair them with an analytics lead, and have them start storing AI visibility data in your warehouse now.

  2. Pick one high-value topic cluster for a 90-day pilot.

  3. Agree with the CFO in advance on what result would justify scaling it with deeper investment.

The pilot usually belongs to whoever owns its budget, most often the CMO, with the CFO signing off on the success criteria before work starts. The first quarter tells you where to start. Every quarter after that tells you more precisely where the next dollar should go.

If you want a clear read on where your brand stands in AI answers, or on how ready your data is to support decisions like these, Velir x Brooklyn Data works with teams on both, and we'd be more than happy to start the conversation.

This post expands on our Sitecore Symposium 2026 session, "Data + Digital: Building a Content Learning Machine with Scrunch."

Published:

Latest Ideas

Take advantage of our expertise with your next project.