Ads, Data & AI Marketing Agents
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
The stream follows the connected work of building and marketing a Shopify app: testing it in a real store environment, seeking additional perspectives and user feedback, monitoring speed and bugs, and making the product’s value quickly understandable. Much of the discussion concerns advertising readiness and measurement—reviewing creatives, correcting social-page branding, consolidating platform metrics into a data pipeline, and choosing a conversion event that balances signal volume with proximity to the real goal, such as installation or completed onboarding. Luqman argues that creative intuition alone is unreliable and that testing, market response, and data matter more, while stressing that trials need time to mature and that the campaigns had not yet demonstrated success. The group also discusses Meta’s shift toward broad targeting, “learning limited” campaigns, lead-capture funnels, and attribution, leaving questions such as the best optimization event, the value of lookalike audiences and affiliate marketing, and parts of tracking and implementation unresolved. The stream also presents experiments and proposals involving Claude, Codex, MCP, and specialized agents for campaign setup, ad review, cross-session strategy handoffs, content production, SEO, and social publishing. Luqman emphasizes human review because AI recommendations can become inconsistent, and describes designing an agent that challenges assumptions instead of automatically agreeing. Audience questions cover possible Arabic–English differences in token use and coding quality, the limits of using AI for advertising without marketing fundamentals, and how to choose a SaaS idea or learning specialization. Broader discussions address market and competitor research, unit economics, the difficulty of customer acquisition, and the value of starting with a focused market or specialty. They also include explicitly uncertain forecasts about autonomous agents, robotics, optional work, and large one-person companies, followed by personal—not universal—views on partnerships, accountability, disagreement, and the social effects of economic pressure.
87 moments from the stream
- 0:03:58
In-app reviews and user feedback
An in-app flow invites a merchant or subscriber to leave a Shopify review. He also planned to speak with one user to gather feedback on the app’s clarity, speed, bugs, and experience within Shopify.
- 0:06:03
Reviewing an ad before launch
A Shopify-store ad is reviewed: the voice-over sounds unnatural and needs clearer pauses, while a floating URL bar obscures important information and should be repositioned. The discussion stresses that creative judgment alone does not predict results—an apparently weak ad may perform well and a strong-looking one may fail—so market response matters.
- 0:09:32
Ongoing debugging and app performance work
In the August 17, 2026 session, debugging was described as an ongoing process despite completion of a specific cleanup task, and an app-speed issue remained on that day’s work list.
- 0:11:35
Getting a Second Perspective Before Product Launch
The speaker says he tried the application in a real store environment and found matters requiring review, but still wants additional testers with practical experience to provide a different perspective before launch and help him observe the market response.
- 0:15:50
Using Crunchbase to Research Markets and Competitors
The speaker describes researching companies and founders before entering a new field. In his view, Crunchbase is useful for examining investors, funding rounds, and the level of investment interest in an industry, helping form a clearer picture of the market and competitive landscape.
- 0:20:42
Evaluating Ads Through Testing and Data
The speaker reviews an ad about checking a Shopify store’s visibility across AI-engine answers, then explains that personal judgments of creative quality can be biased: audiences may respond to an ad its creator dislikes. He therefore favors testing and reviewing the resulting data, while also noting the value of consistent voice across ads.
- 0:25:07
From Store Knowledge to Agent Actions
An audience participant suggests adding a layer beyond what an agent knows about a store, allowing it to take actions connected to a recommended product, such as purchasing or reserving it. The details remain unclear, and the speaker asks for the intended action to be specified and the question rewritten more clearly.
- 0:27:35
Simplifying Marketing Measurement Ahead of a Planned Ad Launch
At the time of recording, the speaker stresses the need for a data pipeline that can measure marketing results carefully before proceeding with the planned ad launch. He says the command center then in place is not wired properly and is overly complex despite containing important metrics, so he turns toward simplifying it and clarifying how outcomes will actually be measured.
- 0:31:03
From Store Analysis to Automated Action
The speaker explains that the tool goes beyond analyzing a Shopify store and recommending next steps: with the merchant's consent, it can automatically carry out permitted actions through an autopilot-like mode.
- 0:31:24
Winning a Focused Market Before Competing Broadly
The speaker argues that a solo developer cannot realistically compete head-on with established, well-staffed companies, and that building a strong position in the Shopify ecosystem first may be a necessary step toward a broader platform.
- 0:35:26
Taking Time During Idea and Market Discovery
The speaker advises against rushing the difficult early stage, arguing that time spent examining the idea and its application area can improve market understanding and sharpen the concept.
- 0:36:18
Testing Multiple Ad Angles Before Scaling
The speaker discusses running experimental ads even when not every creative meets his final preferences, then using the results to identify stronger angles and produce further batches around them while remaining patient during testing.
- 0:37:59
Preparing the Data Pipeline Before Facebook Automation
The speaker prioritizes getting the data pipeline in order, then raises an unresolved question about whether the needed work can be created through Facebook MCP or must be done manually.
- 0:40:10
Automating Store Content and Search Optimization
An explanation of a marketing agent that can select blog topics from keywords and the store’s niche, write posts, product titles, descriptions, and FAQs, and improve product-page search elements. The merchant can choose between automatic execution and manual control.
- 0:42:23
Using Arabic or English with AI Tools
A discussion about addressing tools such as Claude and Codex in Arabic or English, including a personal preference for English voice input. The group raised whether token usage differs by language, while the speaker explicitly said he did not yet know the answer.
- 0:47:37
Code Volume Is Not Software Quality
A general discussion arguing that a very large codebase can indicate spaghetti code and a need for cleanup, and that an experienced developer might implement the same functionality more simply. This was presented as an assessment and opinion, not a documented test.
- 0:49:13
Arabic–English Token and Cost Differences
After querying an AI tool, the speaker relayed that an Arabic sentence may use more tokens than its English equivalent and could therefore cost more. The effect was said to matter less for coding tasks because code output is generally in English. This was a relayed answer rather than an independent measurement, and the segment ends with an open question about quality differences.
- 0:50:10
Does Coding-Prompt Quality Differ Between Arabic and English?
The speaker says the tool can work in Arabic but reports that it performs better in English, attributing this to the greater volume of English training material and to English-language library names, keywords, and error messages. He also explains how technical vocabulary may reduce the translation needed before concepts can be mapped to code.
- 0:52:01
Code Volume Is Not a Measure of Project Strength
The speaker discusses how accumulated additions can make code harder to develop and modify over time, stressing that a larger line count does not automatically make a project stronger.
- 0:54:49
Consolidating Platform Metrics in One Data Pipeline
The speaker proposes gathering metrics from multiple platforms in one place instead of checking each platform separately, supporting easier monitoring of performance, expenses, and forecasts. Private financial and operational details are omitted.
Got a question about something in this stream?
Ask in the group