How Aeqium uses Grain to automate customer feedback with Claude

After implementing Grain and Claude Cowork, Aeqium's customer success team eliminated manual product-request logging, started closing the feedback loop on every shipped feature, and built the foundation for a self-updating customer profile system — all without adding headcount or enterprise CS tooling.

Headquarters: San Francisco, CA
Founded: 2020
Employees: ~20
Industry: HR & Finance Software (AI-powered)
Champion: Will Ruff, Head of Customer Success
Stack: Grain, Claude Cowork, Zapier, Notion

Aeqium brings AI-native automation to customer success with Grain

Most customer success teams sit on a goldmine and never mine it. Every call, every Zoom recording, every off-handed "it'd be cool if it could do X" gets transcribed somewhere, filed away, and forgotten. The pain points are right there in the transcripts — and at most companies, the path from a customer saying something to a product manager acting on it runs through a person manually typing it into a spreadsheet at the end of the week.

Will Ruff, Head of Customer Success at Aeqium, an AI-powered software platform for HR and finance teams, decided to skip that step entirely.

"Almost all of the information that matters about the work I have done is contained in customer communications. It's just a question of how to consistently organize, extract, and update that in a nicely formatted way."

Will runs customer success at a roughly 20-person company. He doesn't have the headcount to staff a manual feedback program, and he didn't want to wait until Aeqium grew into something that needed enterprise CS tooling. So he built his own customer intelligence layer — using Grain transcripts as the substrate and Claude Cowork as the brain.

The starting point was actually a switch. Aeqium had been on Fireflies.ai. Fireflies handled the basics fine. The problem was downstream.

"The reason I actually signed up with Grain was because I was interested in automating the actions that come out of a customer call. Fireflies served the transcribing and note-taking needs, but they didn't integrate with Zapier or Make to pull in the actual transcript content."

With Grain's Zapier integration and Claude Cowork connector, every transcript became a programmable input and the customer success workflows Aeqium had been waiting to build started shipping.

A self-serve product-feedback engine, built by one person

Aeqium runs a biweekly product review with their CEO, Head of Engineering, and Head of Design. Before the automation, Will spent the days leading up to that meeting manually logging every customer request he'd heard, scoping the lift, and flagging quick hits.

Now a Claude skill does the heavy lifting. After his calls, Will runs the skill and it reads new Grain transcripts since the last run, identifies feature requests (filtering out bug reports), compares each one against the existing request database in Notion, and either appends the customer to an existing entry or creates a fresh entry with the customer name, the date, the pain point, the requested behavior, and an AI-generated summary.

"It posts them in with a lot more context than I used to write on these requests, because it's generating an AI summary of the pain point and all that sort of stuff."

The biweekly product meeting still happens. The pre-meeting scramble doesn't.

Closing the loop on every shipped feature

The second workflow is what happens after a feature ships.

When something Aeqium built solves a pain point a specific customer raised, Will runs another Claude flow that pulls the change description, cross-references the original request and the Grain transcript where the customer raised it, and drafts a personalized update email — "Hey, we shipped X. This should help with the workflow you mentioned on [date]."

Will reviews and sends.

"Customers respond pretty favorably when it's like, 'Oh, I asked for this and they delivered it,' and it appears that wasn't independent of me asking. It makes you feel valued."

The change-detection step is still being wired up. Once the changelog feeds into Claude reliably, the rest of the loop closes automatically.

Sharing customer insights across Aeqium with Clips, Playlists, and Trackers

Beyond the automated workflows, two Grain features have become core to how Will gets the voice of the customer in front of the rest of Aeqium: Clips and Playlists for narrative-building, and Trackers for cross-customer pattern detection.

"Clips and Playlists make it incredibly easy to compile and share key insights and narratives across Aeqium."

When Will hears something in a customer call that the product, engineering, or leadership team needs to see, he doesn't summarize it secondhand — he clips the moment, drops it into a playlist, and shares it. The customer's actual voice does the persuading.

Trackers handle the other direction: reviewing topics across many customers at once, instead of moment by moment.

"Trackers enable us to seamlessly review key topics in a comprehensive fashion, across all of our customers and prospects, giving us new insights into the areas we really care about."

Together they turn Aeqium's call library from an archive into an active research surface — one that surfaces patterns Will couldn't have caught from any single call.