A startup that gets brands recommended by AI chatbots just hit a $1.8 billion valuation. Here's why that matters before you submit an 'or equal'

Profound, a startup that helps brands get recommended inside AI chat answers instead of ranked search results, raised a $180 million Series D on September 15 at a $1.8 billion valuation — up from the $96 million round it closed just seven months earlier. That's real money confirming that what an AI chatbot says about a product is no longer a neutral byproduct of its training data. It's a surface companies now pay to influence, the same way they've paid for search rankings for two decades. For anyone who prices a job by researching products, that changes how much weight an AI's product suggestion should carry.

What does Profound actually do?

Profound tracks how often a brand shows up in AI-generated answers, what sentiment those answers carry, and how the brand is positioned against competitors — then helps the brand's marketing team change that. Sequoia and Kleiner Perkins led the new round; existing backers Lightspeed, Khosla Ventures, and South Park Commons also put in money. The company says it now serves more than 700 enterprise customers, including roughly 10% of the Fortune 500. None of that is construction-specific. Profound sells to brands broadly. But the category it's in — generative engine optimization, aimed at getting a product mentioned when someone asks an AI chat tool a question instead of typing a Google search — is exactly the layer building-product manufacturers now have to compete in too.

Why does that matter to someone estimating a job, not marketing one?

Because the research habits of the people upstream of a bid have already started shifting. A 2026 survey of roughly 800 architects and designers, run by rendering company Chaos and architecture platform Architizer, found 64% had already experimented with AI tools in their day-to-day practice, and 86% of those users reported it measurably saved them time. That's not a survey about product specification specifically, but it confirms the underlying shift: AI tools are now a normal part of how design and precon teams work, not a novelty. As that habit extends into product research and "or equal" substitution requests — asking ChatGPT to compare two manufacturers' fire-rated door assemblies, or suggest an equivalent to a specified valve — the answer that comes back is increasingly shaped by exactly the kind of optimization Profound just got funded to sell.

Should you trust the AI's product suggestion in a submittal?

Not on its own. Here's the check before an AI-suggested product reaches paper:

None of this is new diligence — it's the same verification a good estimator already does before accepting a sales rep's claim. The difference is that the claim is now coming from a chat window that feels neutral and isn't.

The takeaway

Use AI to speed up the first pass of product research — that part works and isn't going away. But the Profound round is confirmation, in dollars, that the answer an AI gives about a product is a competed, monetized result, not a fact. Before an AI-suggested substitute goes into a bid or a submittal package, verify it against the manufacturer's own current documentation the same way you would a rep's pitch. The habit costs a few extra minutes per substitution; skipping it costs a rejected submittal or a compliance callback later.

DeepSeek's new model reads a full spec book in one pass for pennies — a reminder that AI is genuinely useful for compressing spec review, right up until the point where its output needs the same sign-off any other secondhand source would get.

Forward this to whoever on your team is fielding "or equal" requests this week.

Friday one chart. Every week, one piece of data that should change a decision on your project. Subscribe at constructionaibrief.com.