How Air achieved a 900% increase in knowledge base updates with no additional headcount

Lauren has personally maintained Air's help center for nearly five years.

Use case: Help center maintenance
Industry: Creative operations

Customer Profile: Lauren Ford, Director of Customer Support, Air
Pageloop features used: Update detection, Find outdated screenshots, Chrome extension, Intercom integration

Air is a creative operations platform used by 2,500 brands to organize, approve, and scale visual content across every channel and team. The company was founded in Brooklyn in 2017 by Shane Hegde and Tyler Strand, and is known in B2B SaaS for its brand work, including a full-page New York Times manifesto by the CEO and a Webby-honored video series.

Lauren has been at Air for nearly five years. As she puts it: "I was their first customer-facing hire. So I founded customer experience." She was recently promoted to Director of Customer Support and leads the support function. What’s impressive about this Support function is that it also contributes a big portion of inbound revenue for Air. Today, Air relies entirely on Pageloop to keep its help center in sync with their product.

"I couldn't stop talking about y'all. This is the dream. Somebody really came into my brain and said, "this is what we need to build."
– Lauren, Director of Customer Support, Air

The challenge

Air is a company that pushes product updates “multiple times a day, all throughout the day”. In the last year, the company has also rolled out a complete UI rebuild called UI 2.0, a full brand refresh, and a company-wide terminology change from "members and guests" to "users." While a seemingly simple change, a single word change can affect dozens of articles, especially in a help center that now spans more than four hundred articles. And, according to Lauren, it can be really difficult to just use keyword search on Intercom to make sure that you find an exhaustive list of articles that you need to review and edit for changes like this.

The help center also has a second audience that Lauren didn't choose: Fin, Air's Intercom AI agent. Air launched Fin in May 2025. In its first two months, Fin was resolving fewer than half of the conversations it handled. When an article described a feature that had already changed, Fin repeated the outdated information to customers. The help center wasn't just falling behind for human readers. It was feeding wrong answers to the chatbot.

"There hasn't ever been, like, a kind of constant help center that matches what our actual product is doing today."

- Lauren

Before Pageloop, Lauren's process for keeping up was to scan the release notes, Slack channels, and search through Intercom by keyword search to find affected articles. With a company shipping rapidly and a growing help center, that approach hit a limit. The discovery work alone was costing her four to five hours a week before any rewriting began.

The solution

Air started using Pageloop in July 2025. Pageloop watches Air's release notes Slack channel and Intercom for changes that affect documentation. When the product team ships something that breaks an existing article, Pageloop scans the entire help center to find any article affected and sends it to Lauren with the suggested action - create a new article or update a set of existing articles. Once Lauren provides any additional context that’s necessary, Pageloop provides a suggested rewrite ready for her review. She approves, edits, or ignores the suggestions. Approved changes are published back into Intercom in the right collection, all from within Pageloop. Of the suggestions Pageloop has sent Lauren, over 95% were ready for review without her needing to attach additional context.

For new articles, Lauren uses Pageloop's Chrome extension. She clicks through a feature in Air's app while narrating what she's doing by keeping the microphone on, and Pageloop generates a draft with screenshots, hyperlinks, and the team's style applied.

The results

Today, more than half of Air's active help center is maintained through Pageloop. Pageloop handles the discovery work, proactively surfaces changes from her connected signals, and passes them on to Lauren for review.

The help center isn’t just a vital part of Air’s customer support functions. It’s also a working tool for other teams at Air. Engineering teams that handle tech support, integrations, and product configuration rely on the help center that Lauren maintains for their own work.

What it means for Air's Fin AI agent

Fin, Air's Intercom AI agent, reads from the help center that Lauren maintains with Pageloop. When the article describing a feature matches the product, Fin's answer matches too. When the article is out of date, the answer is wrong.

When initially launched, the help center at the time was still catching up with UI 2.0, the terminology changes, and months of rapid shipping. Fin was live, but the articles it was reading from hadn't kept pace.

In just a couple of months after Pageloop had been running, the help center stabilized, and Fin's resolution rate climbed to nearly 70%, even as Fin now handles nearly four times the conversation volume it did at launch.

Air went from roughly 1,600 support conversations a month in September 2025 to about 600 by April 2026. Most of the reduction comes from Fin resolving conversations on its own, because the articles it pulls from are accurate.

"The main benefit for Air and for me has been the ability to keep the help center up to date and accurate, which is what informs our chatbot, which is performing beyond our expectations when we launched it back in May."

– Lauren

With Pageloop, Lauren is now able to focus her time where it actually matters: delivering stellar customer support, and not hunting for outdated articles.