A Salesforce cleanup that crawled for months finished 300 records in three days
A custom GPT for brand ownership and trademark research helped clear about 300 Salesforce records in three days, after months of slow progress.
This student was deep in a messy Salesforce cleanup after a database migration, checking brand ownership and trademarks one record at a time. They built a custom GPT that turned scattered research into one structured, sourced answer, and kept a human firmly in charge of every decision. Their story shows that using AI well is about judgment as much as speed.
Brand Owner Trademark Research Assistant
Before
Their team had migrated from an outdated relational database into a highly customized Salesforce system, and the data needed cleaning. Each record meant researching who owned a brand, its trademarks and the right contacts across many sources. The work had progressed slowly over several months.
“work that had previously progressed slowly over several months.”
In their own words
The turning point
In the program they set out to design their own custom GPT for this research. They built it for partial and imperfect inputs, the way real data actually arrives. It followed a tiered research model that put primary legal sources first, separated verified facts from inferences and unknowns, and demanded a source for every conclusion.
“Importantly, the system supported—but did not automate—decision-making.”
In their own words
After
With the assistant, they processed 50 to 100 records in about three hours and finished roughly 300 records in three days. The tool pulled multi-source research into a single, structured response, cutting manual lookups and making decisions more consistent. Every output was still reviewed and validated by a person, in line with their organisation's AI governance policies.
What it lets people do
The assistant lets a researcher check brand ownership, trademarks and contacts against public trademark databases, government business registries and official company websites in one pass. It shows clearly what is proven, what is likely and what is unknown, so a human can decide with confidence.
What they learned
They came away convinced that AI mastery is about judgment, not speed alone. By auditing sources and respecting uncertainty, they could trust the system without giving up responsibility, using AI as a research governor rather than a decision-maker.
“By auditing sources, validating evidence, and respecting uncertainty, I was able to trust the system without surrendering responsibility.”
In their own words
Brand Owner Trademark Research Assistant was built during AI Mastery.
Read the full submission
Written 9 February 2026 Submitted by the storyteller
Every story is written by a Mindvalley student or by someone our teachers have worked with. How we collect and check stories
More from AI Mastery
