Gyorgy Bolyki. AI that fits how your business runs.
I run the AI adoption side of Magrathean alongside the IT consultancy. My work focuses on agentic AI in real codebases — Codex, Claude Code, MCP, GitHub Copilot — with the discipline of someone who has also had to keep a regulated Microsoft 365 estate audit-ready.
Practice areas
Tooling, workflow, identity and review stay in one conversation.
What I help with
- Agentic coding workflows in Codex and Claude Code
- GitHub Copilot rollouts on regulated estates
- Custom Model Context Protocol (MCP) servers
- Eval suites for codebase-specific agent use
- CLI checks and review discipline
- M365 + identity guardrails around AI tooling
Tools I set up
Codex agent
Codex is a repo-aware coding agent. Used well it ships real diffs; used poorly it creates churn.
Claude Code
Claude Code is a terminal coding agent from Anthropic. It works best when the team designs the workflow around it.
Copilot rollout
Copilot is easy to introduce and easy to introduce badly. The rollout is half technical, half policy.
MCP design
MCP lets agents call your tools. Good MCP design is API design for a model caller.
Results from in-house IT
The same controls work that AI rollouts depend on, measured before and after.
- Before72%
- After96%
- Before25–30
- After12
Why the work is needed now
UK business use of AI has nearly tripled since late 2023, and the controls around it have to keep pace.
- Sep 2023: 11.9%
- Dec 2023: 11.8%
- Mar 2024: 13.8%
- Jun 2024: 14.8%
- Sep 2024: 17.8%
- Dec 2024: 18.2%
- Mar 2025: 20.6%
- Jun 2025: 25.1%
- Sep 2025: 27.2%
- Dec 2025: 28.7%
- Mar 2026: 32.1%
- Jun 2026: 34.9%
Start with one workflow
The AI readiness score helps separate a useful pilot from a tool-first project.
Why one operator helps
- 01
Hands-on
Hands-on across the full agent stack, not a vendor reseller.
- 02
Context-aware
Carries the M365 and identity context, so the rollout is not split across two consultants.
- 03
UK-based
Watford-based, mostly remote across UK SMEs and engineering teams.
Questions teams ask first
Do you only work with engineering teams?
No. The AI practice starts with a business workflow and can cover sales, operations, finance, HR, support and leadership alongside engineering use cases.
Why include Microsoft 365 security in an AI rollout?
Identity, device trust, sharing and data boundaries determine what an AI workflow can safely reach. Those controls are part of the rollout, not a separate afterthought.
Keep exploring
