CloneBrake — when a store gets copied
From a manual-response proposal to a named product with evidence-first review, checkout authorisation and a payment-linked case workflow.
Currently: outcomes unvalidated.
- My focus
- Product definition, branding and workflow design
- 01Find suspected copies
- 02Review and explain the evidence
- 03Owner authorises at checkout
- 04Operator confirms the paid case
- 05File supported complaints and track outcomes
Why this exists
Small retailers who find copies of their stores can struggle to assemble evidence and coordinate an effective response. The September proposal began with a manual incident-response pilot. It evolved into CloneBrake: a detection-first product concept that helps an owner understand what has been copied before deciding what to do.
What I’m testing
Can clear evidence help a retailer decide whether to act, and can a controlled workflow turn that decision into supported complaints? Detection quality, operating effort, willingness to pay and enforcement outcomes still need validation.
Key decisions
- Choose the name through evidence. Brand Launch Agent compared candidates and checked domain, social-name and company/product collisions. CloneBrake was selected and the matching domain purchased; those steps do not imply full legal clearance.
- Lead with the comparison. The revised design shows the copied material before the agent’s recommendation. Customer reports explain the evidence in sentences rather than exposing a confidence score.
- Separate authorisation from confirmation. The newer workflow replaces approval of every complaint with owner authorisation at checkout and an operator check before filing.
- Make the paid-case boundary explicit. The design gives paid cases their own confirmation queue and starts the proposed filing clock at payment, including the team’s check.
- Keep validation staged. Reporting, outbound validation, self-service and monitoring remain separate steps rather than one assumed launch.
- Start with a small, checked batch. The first outreach trial is scoped to ten individually reviewed emails and manual handling of initial paid cases, with readiness gates before outreach and payment.
- Separate outreach from case messages. A dedicated mailbox handles prospecting and replies while receipts and refunds use the service communication channel.
- Check the report before releasing it. Outbound reports should contain only human-confirmed copies with matched evidence. Review warnings about possible stockists or shared supplier images, and preview the same report component the owner will see.
How the design evolved
The original plan kept the owner involved in approving every complaint. The later design moves that authorisation to checkout, while retaining operator confirmation of the evidence. A case that cannot be confirmed should be refunded rather than filed. Routes that require the owner to submit directly remain a separate handoff.
The visual direction evolved too. The October redesign uses annotated comparisons and plain-language explanations to make evidence easier to inspect. Its operator layout places the comparison beside a decision panel, so checking the evidence does not require scrolling past the recommendation.
Evidence and learning
The project history preserves the manual-pilot proposal, detection-first plan, naming decision and subsequent workflow revisions. The naming exercise is also a concrete internal use of Brand Launch Agent: availability observations changed the shortlist before a name was chosen.
Product and design decisions are supported by the working records. Engineering tasks are tracked separately; this case study does not yet establish a complete working service, public launch, achieved filing times or successful enforcement outcomes. Naming checks were observations at the time, not a continuing guarantee of availability.
Project evolution
Recent decisions and releases, newest first.
- Product decision
Tightened the report release checks
Required human-confirmed copies and visible matched evidence before an outbound report can be released, with an explicit review of warnings.
- Product decision
Separated outreach from service communications
Chose a dedicated outbound mailbox, separate from receipts and refund messages, with send-time checks and reply handling.
- Product decision
Scoped a small first outreach trial
Defined a first batch of ten individually checked outreach emails, followed by manual handling of the first paid cases.
What’s next
Validate the evidence review and paid-case flow with real cases, measuring detection quality, operator effort and owner understanding. Test willingness to pay and filing performance before treating the proposed service window or broader rollout as proven.
Related writing
Don’t Bet the Guarantee on the Model explores the separation between model reasoning, reliable evidence and controlled actions.