The Content King Is Dead, Long Live Context

I spent three days at UNBOUND 26 about a month ago, and if I had a dollar for every time someone said "context," I could have covered my hotel. (AI still won, but it was close.) The more sessions I sat in on, the more I noticed that almost every good talk came back to the same thing: data. Clean, structured data that people and AI can actually trust.

For years, "content is king" meant publishing more and hitting your keywords. That doesn't work the way it used to. AI search pulls answers from information that's organized well enough for a machine to extract and trust, so structure matters more than volume.

Context has two audiences now

Your readers want answers that solve their problem. AI tools want content they can pull from and cite. Both run into the same issue: data spread across too many places.

Evelyn Welter from HubSpot called this the "context gap," the space between what your CRM knows and all the other tools your team uses every day. Her answer is connectors that let your data flow out to AI assistants and pull in outside data, so the answers you get are based on real records.

Amanda Kopen, also from HubSpot, shared that (based on HubSpot's data) 61% of AI citations come from peer sites and 17% from earned media. So your website is only part of the picture, and the pages you do own need to be easy to pull from: lead with the answer, add FAQs, and use schema and tables.

AI has a different job at every stage

Sangram Vajre of GTM Partners broke down where AI fits based on company size:

• ~$100K: Use AI to build the product. Keep sales founder-led.
• ~$1M: Use AI to build pipeline.
• ~$10M: Use AI to scale your GTM motions.
• $100M+: AI becomes the product.

No matter the stage, AI can only work with the data you give it.

AI makes bad data worse

Darren Smith from Personio walked through four layers of a marketing system, and I wrote down every one:

• Process: "You can't AI a bad process."
• Data: "Data doesn't create clarity. Structured data does."
• Technology: "Tools don't create leverage. Systems do."
• Governance: "The hidden career moat."

His team spent most of a year rebuilding a messy HubSpot portal. At the time, it felt like cleanup. Looking back, it was the best AI prep they could have done.

Sarika Garg from HubSpot sees the same thing on the revenue side, where quotes, contracts, and billing end up in different tools after a deal closes. My favorite quote of the week came from her: "AI on half the data doesn't give half the truth. It confidently gives wrong answers."

AI can find the mess, but people have to fix it

Tyler Samani-Sprunk from Simple Strat has run about 150 HubSpot audits, and every one starts with a person logging into the portal. A dashboard can tell you a field has 40,000 blank values. It can't tell you whether that came from a broken form, an old campaign, or a field nobody needed in the first place. He always starts with lifecycle stages and pipelines, since vague stages like "on hold" are usually where data starts to go sideways.

A few numbers worth sharing

• Mixing email formats (plain text, digest, personal note) lifted opens from 18.7% to 31.2% and clicks from 2.6% to 5.3%.
• Meta Advantage+ campaigns achieved a 22% higher ROAS and a 32% lower CPA than manually targeted campaigns.
• One team cut email volume by 25%. Open and click rates went up, but pipeline went down. Keep an eye on totals, not just rates.

What we're taking away

• Audit before you automate. AI can find the duplicates, but a person still has to decide what they mean.
• Clean up your data first. The same structure that gets you cited by AI is what your own AI tools need, too.
• Write for two readers. Clear answers for people, schema for machines.
• Get mentioned beyond your own site. Peer sites and earned media drive most AI citations.
• Let the data guide your formats, and look at totals, not just rates.

The teams getting this right aren't publishing more. They're taking a hard look at what they already have and building a foundation that both people and AI can trust.

Not sure what's hiding in your HubSpot portal?

That's exactly where we like to start. At GrowthTech, we look at your data, lifecycle stages, and integrations first, so every AI tool you turn on has something solid to work with. Feel free to grab a time that works for you >>