Scan Funnel Implementation
Want to back the stream? The «داعم» membership on YouTube is EGP 60 a month: badges, emoji, and I answer your questions first on stream. Become a supporter
The gist
This portion centers on developing Burnish and an evolving proposed acquisition funnel for Shopify merchants. Some variants begin with only a store URL, while others also request a business email, then present a scan or audit report that may assess visibility in AI answers and identify product-data gaps before inviting a later action. The report’s scope, free depth and pricing, scan-engine integration, accuracy and cost, the mechanism for conversion into an install or subscription, and abuse controls all remained in development; the records do not establish completion. Practical work covered waiting and progress states, report depth, calls to action, mobile presentation, simplifying scan setup, and automating product and competitor selection, alongside investigation of UI, engine, worker, and deployment failures without a documented final resolution. Luqman argued that the product should deliver observable value, turn visibility measurement into useful recommendations, educate merchants, and avoid guaranteeing SEO or GEO outcomes, while organizing the roadmap, reducing context switching, and using parallel work where tasks are not tightly dependent. This overview prioritizes the central Burnish and funnel work, but the indexed pages also contain separate discussions of brand and domain naming, the unresolved availability of Stripe in Egypt, Perplexity MCP setup and Hermes evaluation, AI-assistant behavior, a second-hand and unverified claim about Shopify pre-submission checks, projects slowing near completion, and audience questions about starting within Shopify, patience with organic visibility, and web-design learning resources.
81 moments from the stream
- 2:52:45
Making a Coding Assistant More Proactive
A discussion contrasts an assistant that ends with a question, direction, or next step with one that states the problem and stops. The observation becomes a practical instruction for the assistant to keep moving the work forward and suggest solutions or next steps after a scan fails again.
- 3:03:46
Why Projects Slow Down Near the Finish Line
A practical observation that tasks may move quickly when a project starts from scratch, then slow as it approaches completion, based on the speaker’s experience with prior projects not named here.
- 3:08:54
From AI Visibility Measurement to a Deep Diagnostic Report
The discussion distinguishes measuring a merchant’s visibility across AI platforms, whose method seems clearer, from improving that visibility, which remains more ambiguous and needs research. It proposes a report assessing catalog elements such as product descriptions, alt text, images, and tags, alongside component breakdowns, prompts, and competitor comparison, while some implementation details remain under development.
- 3:14:03
Reviewing UI and Scan-Engine Issues
The review found the mobile presentation broken and questioned why some engines were missing despite available APIs. Troubleshooting included restarting the worker, refreshing, retrying the scan, and observing a worker-related error; the captions do not establish that the issue was resolved.
- 3:17:35
Evaluating Scan Economics Under Conversion Uncertainty
The discussion weighs report-generation cost against conversion potential and the longer-term value of leads. The speaker explicitly treats the conversion examples as hypothetical, declines to assert unsupported rates, and considers selling the report itself as an option requiring further thought.
- 3:21:15
Simplifying Perplexity MCP Setup
The speaker notes that Perplexity MCP is now available and contrasts the easier current setup with an older approach requiring manual configuration in editor tools. The captioned setup command is unclear, so no exact steps are inferred.
- 3:23:41
Scan-Level Cost Accounting and AI Engine Integration
A discussion about adding per-scan cost accounting, clarifying whether a theoretical report incurs real usage, and showing an engine-by-engine breakdown. It also identifies the need to finish wiring Grok and DeepSeek before running real reports through them.
- 3:24:58
Assistant Behavior and Dependent Session Workflows
An in-progress opinion that Claude behaves more proactively while GPT feels more reactive to prompts. The discussion also notes that more parallel sessions can usually speed up work, but offer less benefit when the steps are tightly dependent on one another.
- 3:28:27
Researching Proven GEO/AEO Activities Versus SEO
A brief for deep research through Perplexity from a new Codex session, aimed at identifying and ranking activities shown to improve visibility in AI-generated answers. It asks researchers to separate tactics shared with SEO—such as backlinks, blogging, and strong product descriptions—from GEO-specific optimization that does not overlap with SEO.
- 3:33:50
Building a Conversion Funnel Around a Valuable Store-Scan Report
The funnel is reconsidered from a generic “scan my store” call to action toward a clearer explanation of the deep report merchants receive. The discussion covers collecting a Shopify store URL and business email to start generating the report, then balancing enough free value with a reason to subscribe. It also frames scanning as only part of the product: the broader proposed value is helping merchants address visibility problems revealed by the report. How much of the report to reveal and the exact conversion mechanism remain open questions.
- 3:39:18
Reviewing the Scan Report’s Responsive Layout
During the report experience review, the lower section is found to be slightly cut off on desktop, possibly because a call-to-action element is obscuring it. The mobile version is described as completely broken.
- 3:44:31
From Visibility Measurement to Useful Merchant Actions
The speaker argues that measuring a store’s visibility on AI platforms is not sufficient by itself; the real value lies in identifying actions that can genuinely improve that visibility, supported by a report comparing multiple factors.
- 3:47:30
Reframing the Free-Scan Call to Action
The scan is positioned as an in-depth, professional report that delivers real value rather than as a generic “free scan.” The proposed work includes inventorying calls to action across the site, developing alternatives, and clarifying the scan’s store scope.
- 3:51:00
API and OpenRouter Limits for Visibility Scanning
Responding to a suggestion to use an account subscription instead of APIs, the speaker says OpenRouter may assist with some scanning tasks but, in his experience, does not provide the same accuracy when checking prompts inside specific platforms such as GPT. He remains open to the alternative while emphasizing the accuracy tradeoff.
- 3:53:02
Scan Wait and Result Presentation
The speaker notes that the scan may take around two minutes before showing the result, then begins describing the related call to action, but the explanation cuts off before completion.
- 3:53:27
Reviewing and Shortening the Mobile Report
The mobile report review found the content still far too long. Although the report should demonstrate a deep reading of the merchant’s store, its mobile presentation needed shortening and refinement.
- 3:54:34
Evaluating AI Agents and MCP Connections
An exploratory discussion considered the Hermes agent and a possible MCP tool, noting the learning curve and need for deeper review. It raised unresolved questions about model-usage pricing, who bears the cost, and the scope of available access; these pages do not provide verified answers.
- 4:00:58
Diagnosing a 500 Error and Missing Worker Logs
A retry produced a 500 error while the worker showed no logs explaining the failure. The debugging exchange also noted that the assistant was asking for commands to be run manually instead of using its own terminal, slowing the workflow.
- 4:01:56
Aligning the Report Funnel and Calls to Action
The report was clarified as explaining a prospect’s current AI visibility and why Burnish may need to help, rather than deciding whether one specific tier is a fit. The planned funnel work attached plan intent to navigation, revised CTA language, priority, and placement across mobile and desktop, added a report CTA to pricing cards, deferred four destination pages, and then moved toward PDF report design.
- 4:13:59
Report Optimization and the Missing Scan Button
After optimizing the report, an issue appeared inside the Shopify app: after trying to launch the app again, a scan-related button was not visible and the interface appeared blank. Luqman also reviewed the call-to-action copy and considered “Generate my report” acceptable before making further changes.
Got a question about something in this stream?
Ask in the group