Background
Camille had been a product manager before going solo. She knew how to build software and she knew how investors thought. What she'd underestimated was how much investors would probe operational maturity for a solo-led company. They weren't asking about the product — they'd seen the product and liked it. They were asking whether she could operate the thing at scale, handle customer issues without a support team, monitor AI costs without a data analyst. For four months she'd been managing operations with a combination of Stripe's dashboard, direct database queries, and a spreadsheet she updated manually on Fridays. The answer she was giving in meetings was true but unconvincing: 'I have a system for that.' When investors asked how she managed user accounts, how she debugged AI issues, how she would handle a customer reporting unexpected charges — she was describing manual processes. That gap was visible in the room. What she needed was to stop describing the system and start showing it.
The challenge
Camille needed to be able to show investors a real operational picture, not describe one. User management with subscription visibility, AI cost tracking by user and in aggregate, a way to investigate reported issues without parsing raw logs, and enough system health visibility to answer 'how do you know when something breaks' with something other than 'a customer tells me.' She needed all of this quickly — she had investor meetings scheduled.
How they built it
From spreadsheet to admin panel in a day
Camille had been avoiding building admin tooling because it wasn't the product — every hour on ops infrastructure was an hour not on the thing customers paid for. What she hadn't accounted for was how much time manual operations were actually costing her. When she adopted ShipAI and got the admin panel for free, she stopped managing user accounts through Stripe's interface, stopped running database queries to check usage, and stopped updating her Friday spreadsheet. The time recovered went back into the product.
The investor meeting that changed
Two weeks after adopting ShipAI, Camille had an investor meeting where she was asked the familiar questions. This time, she screen-shared the admin panel. User list with subscription status, usage analytics by account, AI cost attribution, the trace viewer. The investor's follow-up questions shifted from 'how would you handle X' to 'how are you thinking about X at scale.' That's a fundamentally different conversation — one that assumes operational competence rather than probing for it.
Customer support without a support team
Three weeks after launch, a paying customer reported that their usage numbers looked wrong. Previously, this would have meant a database query, a Stripe dashboard check, and probably twenty minutes of investigation. Camille pulled up the user in the admin panel, saw their usage history and subscription status, opened the AI trace viewer and found a duplicate request that had been counted twice. Total time: about four minutes. She replied to the customer with the resolution while they were still reading her acknowledgment.
The round closed
MetricLoop closed its pre-seed six weeks after Camille started the process. Two investors cited the operational tooling specifically in their investment rationale — one said 'you're running this like a team of five, not a team of one.' That was the point. The product hadn't changed. What changed was what investors could see.
Outcomes
Pre-seed round closed
MetricLoop closed a pre-seed in six weeks. Two investors cited operational maturity — specifically the admin tooling — in their investment rationale.
Customer support time reduced from 20 minutes to under 5
Investigating and resolving user-reported issues dropped from 15–20 minutes of manual investigation to under 5 minutes using the admin panel and trace viewer.
Friday ops spreadsheet eliminated
The manual weekly operations review Camille had been doing since launch was replaced by a real-time admin dashboard. She estimates she's recovered two to three hours per week.
AI cost questions answered with data, not estimates
Investor and customer questions about AI cost scaling went from estimates to screenshots. Camille can show per-user token consumption and cost attribution in real time.
In their own words
I was doing everything manually and it was working, in the sense that nothing was on fire. But 'nothing is on fire' is not the same as 'I'm running this well.' When I had real tooling — actual user management, actual cost visibility, actual debugging capability — I stopped describing my operations and started showing them. That changed how investors saw the company.
“I was running MetricLoop out of a spreadsheet and Stripe's dashboard for the first four months. When I started talking to investors, I realized that story didn't hold. Getting actual operational tooling in front of them changed the dynamic immediately.”
— Camille Dupont
Frequently asked questions
What does the admin panel include?
User management with subscription status and history, usage analytics by user and in aggregate, AI trace viewer for inspecting individual requests, application logs, payment history, and system health. All connected to live data — no mock numbers.
How does Camille use the trace viewer for customer support?
When a user reports an issue with an AI response, Camille searches for their recent requests in the trace viewer. Each trace shows the model used, tokens consumed, tools called, latency at each step, and the full input/output. Most support issues are diagnosable in under five minutes.
Can a solo founder realistically scale customer support without a team?
For the pre-seed stage, yes. Camille handles all support herself using the admin panel tools. Her current SLA is a four-hour response and a same-day resolution for technical issues. She expects to need support help when she's past 200 active users — but that's a problem for after the round.