From Data Chaos to Revenue Visibility
How AI-Powered HubSpot Operations Gave Gold Safe Exchange the Reporting Foundation Their Revenue Deserved
100+
Hours of manual work replaced by AI
$3,500+
In junior labor costs avoided
18
New possible clients when savings are reinvested in marketing
The Situation
GSE (soon to be Ridgemont Capital) is a growth-stage retirement income firm. They sell precious metals IRAs and provide retirement income planning. They had a sales team actively closing deals, revenue moving through HubSpot, and a CRM that had been in use for years. What they did not have was any ability to see what was driving that revenue.
The HubSpot instance had grown without governance. Each rep had added their own terminology, their own workarounds, their own stage names. The data existed, but it was organized around individual habits rather than shared standards. Reports could not be built because the underlying fields did not mean the same thing from one record to the next.
Deric Ned, GSE's Founder and CEO, commissioned a full operations engagement to fix the data infrastructure, build the reporting dashboards, and put governance in place so the team could finally understand what was driving revenue, and make decisions from that understanding.
The Challenge
What the audit revealed was not a data problem. It was a trust problem. The CRM held years of sales activity, but none of it had been organized around consistent definitions. Lead statuses mixed rep names, process stages, and meeting outcomes into a single field with no shared meaning. The deal pipeline collapsed sales, fulfillment, and liquidation into one undifferentiated sequence. Custom properties that had been documented as live did not exist in the system. Closed Won deals, nearly every one of them, carried no product line, no rep attribution, no context that would allow anyone to understand what had actually closed and why.
The result was a CRM that told the team deals were closing, but could not tell them which products were growing, which reps were performing, or where leads were coming from. Revenue was real. Visibility into that revenue was not. The foundation for decision-making simply was not there.
The Fix
illumination and GSE deployed Claude AI in Cowork mode (in a sandbox environment to protect the data) as the primary HubSpot operations resource for this engagement. Rather than bringing in a junior admin or operations contractor, AI was given direct read and write access to the HubSpot instance and instructed to execute the full scope of work autonomously, with strategic oversight from illumination.
This is the AI Operations model. It is not AI generating content or producing summaries. It is AI doing the hands-on systems work: auditing the CRM, mapping what exists against what was documented, identifying gaps and inconsistencies, building clean properties, remapping contacts and deals, and specifying every report the business needs to see. The work that would typically require weeks of junior ops time, with all the coordination, error correction, and ramp-up that entails, was completed in days.
The engagement was designed around a clear sequence. Audit first. Understand the state of the data before touching it. Remediate from that understanding. Build governance so the fixes hold. Every step was grounded in specific observations about what existed, not assumptions about what should exist.
The Outcome
The immediate savings tell part of the story. More than 100 hours of manual operations work was replaced by AI, and over $3,500 in junior labor costs was avoided. Work that would have taken four weeks with a traditional contractor was completed in days.
But the more important number is what those savings make possible. At a customer acquisition cost (CAC) of $200 per client, reinvesting that $3,500 back into marketing does not just recover the cost of the engagement. It funds 18 new deals. The savings compound. The ops spend becomes pipeline.
The Compounding Effect
CRM drift of this kind, where reps add values, rename stages, and skip fields, tends to recur every three to six months without active governance. The audit report and governance framework that came out of this engagement are designed to prevent that recurrence. The value of the work is not only what was fixed. It is what will not break again.
That is the case for AI Operations as a model. Every dollar not spent on manual data remediation is a dollar that can go toward acquiring the next customer. And unlike traditional junior ops work, the output does not require correction cycles, handoff delays, or ramp-up time. The work was done right the first time, at a fraction of the cost, in a fraction of the time.
Revenue visibility was built where none existed. Twelve reports across two dashboards are now fully specified and ready to activate. For the first time, GSE can see revenue broken down by product line, by sales rep, by lead source, and by month.
The CRM data is clean and reportable. Every Closed Won deal is now tagged with product line and sales rep. Every contact carries a standard lead status that maps to a shared definition. Commission tracking is unblocked. The property infrastructure for commission amount, paid status, and date is live. The operational bottleneck that had been sitting between sales activity and compensation clarity has been removed.
Governance is in place. The engagement produced a written audit report, an internal pipeline reference sheet, and a step-by-step remediation checklist. These are not just records of what was done. They are the tools the team needs to maintain what was built.
What This Enables Next
The infrastructure built in this engagement is the foundation for a much larger revenue intelligence capability. With clean data and reporting infrastructure in place, GSE can now track revenue by product line to understand which products are growing and which are stalling. They can measure sales rep performance objectively and build commission models tied to closed revenue. They can attribute marketing spend to closed deals by tracking first-touch source through to Closed Won. They can monitor lead quality by asset level and age range to validate whether targeting is working. They can run a separate fulfillment and liquidation pipeline without contaminating sales pipeline metrics. And they can use lead status and deal stage data to trigger automated workflows, follow-up sequences, and rep accountability alerts.
None of that was possible before this engagement. All of it is possible now.
“When you understand the work that holds you, you begin to understand how to build work that can hold others.”