Agentic AI Engineering
Designs and builds multi-agent systems with real responsibilities: AI roles, reusable Skills, MCP tooling, memory and execution across Mac, Windows, cloud servers and local models.
Explore in depthAgentic AI Engineer · Certified AI Consultant & AI Coach · International Speaker
Builds agentic AI systems that run real businesses — and teaches leaders, teams and builders how to design, use and trust them. Twenty years of hands-on technology, more than a decade of executive search and coaching, and a working human-and-AI operating system behind every recommendation.
David Potgieter's career began with his hands: stripping, upgrading and rebuilding PCs in Pretoria from 2004, then designing CCTV, access-control and network systems for commercial clients. In 2009 he moved to Dubai, where he ran an opening restaurant team, coordinated large-scale facilities at Emirates Group and built a recruitment practice that has conducted more than 100,000 interviews.
Today he is the founder of iMPLEMENTAi.ae, where he architects agentic AI systems — AI roles, reusable Skills, Model Context Protocol tooling, durable organisational memory and local as well as cloud models — and runs his own businesses on them. That hands-on engineering is the foundation for his AI consulting, corporate training, keynotes and one-to-one executive coaching.
The thread through it all is practical: understand the real system, improve it, test it against reality and turn what works into something reusable.
Designs and builds multi-agent systems with real responsibilities: AI roles, reusable Skills, MCP tooling, memory and execution across Mac, Windows, cloud servers and local models.
Explore in depthTurns unclear business needs into scoped AI use cases, roadmaps and working workflows — integrated with the systems a business already runs on.
Explore in depthTeaches AI on real work through workshops, build labs, a public 17-module course, one-to-one executive coaching and keynotes for audiences of up to 10,000+.
Explore in depthBuilt FolderDesk, with 124+ personally authored AI Skills, a 140+ Skill catalogue and 13 AI roles running Talent Bridge and iMPLEMENTAi.
Cut one AI routing workflow from 92 to 18 seconds and from 20 to 5 tool calls.
Rebuilt recruitment into a 21-stage AI-assisted lifecycle, releasing capacity equivalent to two full-time recruiters.
Created a free 17-module course with labs and a capstone, taught by David AI Coach, an AI tutor built on his method.
Coached 54+ executives, including C-level leaders, through a 7-step Executive Career Method over five years.
10,000+ interviews and 1,000+ placements personally; 100,000+ interviews and 10,000+ placements company-wide.
On stage before 10,000+ in the Philippines, 1,000 in the United States, 4,500 and 2,000 in Nigeria, and audiences across the UAE.
Trusted by major Dubai healthcare and education groups for recruitment, training, workforce development, succession planning and AI integration.
David is a low-code-first agentic AI engineer: his strength is system architecture, integration, orchestration and AI-assisted implementation — reading, reviewing and directing code rather than hand-writing every line. Everything below is built and used in production across his own businesses.
Organisational knowledge lives in durable, owned sources, so models and tools stay replaceable.
Every fact and method has one canonical owner; agents read across owners and write to the right one.
Repeatable methods become versioned Skills that any capable agent can load when a task needs them.
Language models handle judgement; scripts and systems handle what must happen identically every time.
A task is finished when its result is checked against the live system, not when an agent reports success.
Agents get real autonomy on routine work; people keep authority wherever the stakes are high.
Consulting starts with the questioning discipline of 10,000+ interviews: extracting what a business actually needs before choosing any tool. Recommendations come from systems David builds and operates himself.
Structured questioning to surface real requirements, constraints and the work worth automating.
Use cases, architecture, success measures and where humans keep authority.
Agents, Skills, automations and integrations with the systems already in use.
Representative tests against real data and live systems before anyone relies on it.
Team training, handover and lessons fed back so the system keeps getting better.
Use-case discovery, prioritised pilots, return-on-effort thinking and governance for leadership teams.
Ongoing Chief-AI-Officer-style direction for SMEs that need senior AI judgement without a full-time hire.
Redesigned processes with AI assistance, deterministic automation and human checkpoints where they matter.
Purpose-built AI Skills and agents for ChatGPT, Claude and Codex, delivered ready to use.
Source-linked knowledge built from documents and archives — such as the AI Estimation Brain proposed for a major Dubai real-estate developer.
AI connected to Microsoft 365, Google Workspace, WhatsApp, CRM, ATS and finance systems through APIs and MCP.
David has trained people since 2007 — hospitality teams, recruiters, executives and now AI builders. Every session follows the same practical loop, on real work the learner brings.
Learners bring a report, a hiring process, a client email or a workflow — and leave with it improved by AI.
Explanations, analogies and exercises adapt to each person's background and demonstrated ability.
Learners practise checking AI output against sources, so confidence comes from evidence, not fluent answers.
Clear AI priorities, realistic pilots, success measures and accountable adoption.
Faster everyday work with ChatGPT, Claude and Copilot-style tools; reusable prompts and workflows.
AI-supported role design, screening, interviewing, assessment and documentation.
AI built into their own practice, content and client delivery.
Working agents, Skills, tool connections, tests and a capstone project.
Keynotes, half- and full-day workshops, multi-week cohorts, hands-on build labs and one-to-one coaching.
ChatGPT and Claude for research, writing, documents and analysis, with output verification.
Use-case discovery, pilot selection, risk, governance and human accountability.
Clear briefs, reusable instructions, knowledge bases and grounded answers.
Job design, sourcing, screening, structured interviews and evaluation.
No-code to low-code agents, Skills, tools, APIs, MCP, memory and testing.
Permissions, source evidence, testing, failure diagnosis and oversight.
A personal AI working system built around one leader's role and decisions.
Practical AI and agentic systems for conferences and large audiences.
A complete curriculum with lessons, labs, workshops, rubrics, templates and a graduation capstone. It is taught by David AI Coach — an AI tutor that runs inside the learner's own ChatGPT, Claude or Codex, follows David's teaching method and resumes from saved progress.
One-to-one coaching over five years for senior professionals and C-level leaders across the GCC.
Deep-dive discovery, executive CV, profile optimisation, job-search strategy and interview coaching — delivered as 4- and 8-week Platinum programmes.
A one-page career dashboard followed by a detailed career journey — now encoded into AI Skills so quality reproduces every time.
From leading a Toastmasters club in Dubai to international stages with audiences of more than 10,000.
Largest audience addressed to date.
Invited twice to deliver on stage.
An audience of 1,000 people.
Corporate, community and Toastmasters audiences; led the Emirates HQ Toastmasters club (member 2014–2019).
Recruitment is intelligence extraction: deep questioning, pattern recognition and matching the right person to the right need. That discipline shapes how David designs AI and how he coaches.
Conducted personally across the recruitment career.
Delivered personally to UAE and regional employers.
Talent Bridge company-wide record, with 10,000+ placements.
AI-assisted recruitment from vacancy intake to post-placement follow-up.
AICoaches.com · 2025
AICoaches.com · 2025
Anthropic · 2026
OpenAI · 2026
Phoenix Educational Institute, Dubai · 2014
South Africa
“To equip 100,000 people with the AI knowledge, skills and tools to work better — and to help many more achieve their goals and freedom.”