A worked campaign example
Worked campaign angle for the advisory engagement: “The judgment layer.”
Engagement shape: a 90-day advisory retainer with three workstreams and one content component.
Workstream 1 — AI integration review. Aziz audits the company's AI feature plans against what actually runs in production, separating features worth shipping from demos. Their stated position gives the buyer the lens: AI coding tools sped up implementation on their own builds, but review, taste, and responsibility stayed with the engineer; platform defaults can commoditize anything built on access to AI alone.
Workstream 2 — build versus buy and cost model. Aziz reviews the internal tools the company is about to license and prices building them instead, drawing on their own record: a hiring dashboard built in about an hour on Cloudflare Workers and D1 at zero infrastructure cost, and an outreach CRM that grew from one HTML file into a multi-user system. They apply the same cost-per-output logic they use publicly against an EUR 87,000-a-month European team of eight.
Workstream 3 — one working prototype. Pick the spreadsheet the team dreads updating, describe the task as you would to a contractor, and let Aziz build the narrow version instead of scoping a platform.
Content component — a three-piece series for the company's technical buyers, drawing on Aziz's policy and AI analysis: 1. A decision memo on AI-generated code: what a senior engineer checks before it ships, based on their HalalCodeCheck lesson that plausible code still has to meet production reality. 2. A cost breakdown of one business process: what it costs to run the old way, what a custom internal tool costs to build and maintain, and where the savings land. Aziz's product-grade pipeline example supplies the method — architecture choices tied to a measured outcome like latency falling from 24 hours to under 20 minutes. 3. A platform-risk piece: what happens to a software business when a large vendor ships its core feature as a default, modeled on their Claude for Small Business analysis, and what judgment-based defensibility looks like instead.
Scale discipline: Aziz argues a solo lead plus a vetted network beats a standing team on focus. Scope the prototype and review work accordingly — one accountable contact, milestones agreed up front, no headcount.
Success check at 90 days: at least one AI feature killed or reshaped before build, one build-versus-buy decision made with a written cost comparison, one working internal tool in use, and the three published pieces read by named buyer contacts rather than measured on impressions.
Before signing, settle in writing: rate and minimum engagement (neither is published), how Aziz's network engineers are contracted, who owns the code and data, who accepts releases, and whether co-published content is in scope. If the company wants the content to function as advertising with reach guarantees, that belongs in the poor-fit case.