The Pull
Trolley was trying to understand what customers were saying across the business and use that signal to make better product and GTM decisions.
Their current setup blocked them because Salesforce had account context, Momentum had calls, Slack had customer-channel feedback, Notion had taxonomy, and product planning had initiatives — but none of it automatically became one customer-signal model.
BuildBetter unblocked them by turning raw customer data into structured, mapped, enriched, source-backed signal.
The Catalyst
“Everyone, as soon as they see it, they’re like, oh, this makes so much sense. But it’s taking that step with how you treat this data and making it first-class and the level of analysis and how you build out the data model for it — it makes a ton of sense. No one’s solved that.”
— Barnett Klane, VP Product, Trolley
Before BuildBetter
Trolley had customer and company signal across:
- Salesforce accounts, contacts, cases, and custom metadata,
- Momentum call recordings,
- Slack customer channels,
- CSV imports and product data,
- Notion product taxonomy,
- product planning workflows,
- PMM, content, competitive, and executive reporting workflows.
Each system held part of the story. None created the analysis layer.
The Blocker
A support case might show a recurring issue. A call might reveal an important blocker. A Slack customer channel might contain real feedback. Salesforce might show ACV, TPV, vertical, or plus-customer status. Notion might define the product taxonomy.
But unless those pieces connect, the team cannot answer:
- which product domains are affected,
- which verticals care,
- which high-value accounts are asking,
- whether an H2 initiative is supported by customer evidence,
- what should route to product, PMM, content, or leadership.
The blocker was not AI. The blocker was the data model that makes AI useful.
“All these AI tools — they kind of are like, ‘oh, this just magically happens when you feed the data.’ But there’s actually a ton of configuration and build needed to do that. None of the AI tools can do that out of box.”
— Barnett Klane, VP Product, Trolley
The BuildBetter Workflow
Momentum calls
+ Salesforce accounts / contacts / cases
+ Slack customer channels
+ product taxonomy
+ product and CSV data
↓
BuildBetter signal extraction
↓
Salesforce business metadata
↓
Product-domain routing
↓
Reports, Slack workflows, PMM research, executive reporting, MCP / automation
What Changed
With BuildBetter, Trolley moved from fragmented customer data toward a structured enterprise signal layer.
The clearest changes:
- calls, Salesforce cases, Slack channels, and taxonomy could connect into one customer-signal workflow,
- product feedback could be organized by product domain and category,
- customer signal could be filtered by account and business metadata,
- PMM and content teams could explore customer intelligence for messaging, personas, competitors, and content,
- executives had a path toward source-backed reporting,
- workflows could route specific signal types into Slack and other systems.
“This is obviously a huge asset for not just the product management team, but the marketing team, some of our other C-suite.”
— Barnett Klane, Trolley
From Anecdotes to Evidence
Before BuildBetter, a dedicated researcher could run two or three research activities a month, each reaching maybe twenty people. Now PMs run multiple research activities a day, drawing on every customer conversation the company already has.
“Before, one researcher could do two or three research activities a month, and could only talk to maybe twenty people per activity. Now PMs can do multiple research activities a day. It’s minutes versus a month turnaround. It’s a complete paradigm shift in terms of user research.”
— Barnett Klane, VP Product, Trolley
The same shift showed up in how decisions get made.
“There was a lot of anecdote-based prioritization by executives. What BuildBetter has really helped is retraining that — let’s actually see what the trend is here, let’s see that account, and we get much richer data. It used to just be a question in a meeting. Now I’m pulling up the agent, getting the report, and we’re having the answer right there.”
— Barnett Klane, VP Product, Trolley
When Trolley’s VC firm surveyed its portfolio on which AI products were actually moving the needle, the answer was easy.
“My VC firm did a survey of the AI products we use, and I put BuildBetter as the biggest impact.”
— Barnett Klane, VP Product, Trolley
“Product-wise, y’all have really high agility compared to larger teams. The speed and focus stand out. Hats off.”
— Barnett Klane, Trolley
