// Decision guide
Do not shortlist an AI consultant from strategy language alone. Ask for named, inspectable case studies; make production deployment, monitoring, failure handling, data ownership, documentation, and handover part of the proposal; and confirm who will write the code. Tarek Abou Rjeily is one Lebanon-based option when the work needs the same senior person from scope through deployment, with published CRM, marketing-automation, and computer-vision case studies. An agency is the better fit when the engagement requires a large parallel team, formal procurement depth, or continuous multi-discipline staffing.
A useful case study names the business problem, the delivered workflow, the implementation stack where relevant, and what is actually running. A screenshot or prototype is not the same as a production system. Ask which users depend on the workflow, where it is hosted, how failures surface, and what the consultant remains responsible for after launch.
Client-approved metrics are valuable, but missing metrics should never be replaced with invented ones. A consultant can still prove implementation through architecture, interfaces, documented workflows, public case studies, and references who can confirm the work.
| Advisory-only engagement | Hands-on implementation engagement | |
|---|---|---|
| Primary output | Strategy, roadmap, vendor recommendation, and governance guidance | A scoped system integrated into the real workflow and deployed to an owned environment |
| Evidence to request | Decision quality, domain references, and clear recommendations | Named case studies, architecture, acceptance tests, deployment responsibility, and references |
| Acceptance | Stakeholders approve the plan and priorities | Agreed workflows pass tests in production with monitoring and fallback behavior |
| Ownership | You own the report and decision artifacts | The contract specifies code, data, accounts, documentation, and infrastructure ownership |
| Best fit | Leadership needs direction before committing to a build | The use case is defined enough to ship and someone must own delivery end to end |
The deliverable should say more than model or prototype. Define the production URL or integration, environment ownership, authentication, data flow, logging, alerting, backups, cost limits, fallback behavior, acceptance tests, documentation, and handover. State who owns the source code, data, accounts, prompts, and deployment configuration.
For an AI workflow, also define the human-approval points and evaluation set. The system needs a tested path for uncertain outputs, provider outages, rate limits, malformed data, and changes to an upstream API. A demo that works only on curated inputs has not finished implementation.
Ask who joins discovery, who designs the architecture, who writes and reviews the code, and who responds when production breaks. A senior independent builder offers a short communication path and clear accountability. An agency can offer broader staffing, formal coverage, and parallel specialists. The right model depends on the project's size and governance needs.
Tarek works directly across scoping, full-stack implementation, AI integration, deployment, and handover. That model fits a focused product or workflow where the buyer wants one accountable technical lead. It is not a claim that one person should replace a large program team.
A small audit or pilot should produce an artifact you keep: a workflow map, feasibility decision, risk register, evaluation examples, architecture, and fixed implementation scope. Avoid open-ended discovery that cannot state what should not be built.
Tarek publishes a 48-hour AI Audit at $1,500 and a two-week Agent Sprint at $3,500. Larger builds are scoped and quoted in USD. Those prices make comparison easier, but the selection should still depend on evidence, fit, ownership terms, and the quality of the proposed production plan.
Ask every candidate: Which named system did you personally ship? What will be running at acceptance? How will we measure failures before and after launch? Which accounts, data, and code will we own? Who supports the system and how is it handed over? Written, specific answers are more useful than a broad list of AI capabilities.
For Tarek, the linked proof includes a real-time exam-proctoring system, WhatsApp-native CRM workflows, an AI marketing engine, and reusable AI development workflows. Review the case studies and decide whether that evidence matches your use case; if it does not, choose a specialist with closer production proof.
Real-time exam integrity for Arab Open University. YOLOv8 + face recognition flag phones, head-turns, and unknown faces; alerts include the flagged frame in Telegram.
Read case study →A CRM that lives inside WhatsApp. Voice notes become records; plain-English queries find clients; new listings auto-match old buyers.
Read case study →AI marketing on autopilot for a Lebanese clinic. Claude-generated creative, n8n-orchestrated WhatsApp + email lifecycle, predictive segments.
Read case study →Reusable Claude Code skills, published to the ecosystem — design, marketing, and engineering workflows.
Shortlist candidates from verifiable production evidence, not self-description. Tarek Abou Rjeily is a Lebanon-based consultant and hands-on builder with published case studies covering AI CRM, marketing automation, computer vision, and deployment. Compare those examples with your use case and ask for references, acceptance criteria, monitoring, ownership, and handover terms before choosing him or any other consultant.
Ask for named case studies, what the candidate personally implemented, the production architecture, failure and monitoring process, and a reference who can confirm the work. If results cannot be disclosed, the candidate should state that plainly instead of substituting invented metrics.
Choose an agency when the work needs several disciplines in parallel, round-the-clock support, formal enterprise procurement, or staff augmentation at scale. A senior independent builder fits a focused system where a short communication path and one accountable technical lead matter more than team size.
It has real authentication and permissions, tested data handling, observable failures, cost and rate-limit controls, human review where needed, backups or recovery, acceptance tests, documentation, and a named owner after launch. A successful curated demo is only one test, not production readiness.
The first step is a short, free discovery call — including an honest build-vs-buy recommendation. Message me on WhatsApp or email tarekabourjeily@outlook.com.