# HG Fabian — AI & Automation Consultancy

> Everyone's talking about AI. We help you use it.

We turn AI from a buzzword into something that actually runs across your infrastructure. Built by engineers, for engineers — with no product to push.

For ISPs · operators · enterprise infrastructure.

- **AI-first** — genuinely ahead on applied AI, not a vendor's slide deck.
- **Hands-on** — real production infrastructure: networks, servers and the systems around them.
- **Independent** — nothing to resell, just the right outcome for you.

## Why this is hard

**AI is everywhere. A plan for your infrastructure isn't.** You're promised the world. You're rarely told where to start, what's realistic for the infrastructure you actually run, or whose advice you can trust.

- **Everyone has an AI story.** Vendors, headlines, your board. None of them start with your systems, and none of them tell you what's actually worth doing first.
- **Honest advice is hard to buy.** Most "experts" are selling a platform. The recommendation always turns out to be the thing they happen to be paid to sell.
- **Your infrastructure isn't a clean slate.** It's mixed, it's been added to for years, and some of it is past its prime. Off-the-shelf AI assumes a greenfield you don't have.
- **Pilots that never ship.** The demo dazzles, then dies before production — because no one wired it into your real systems, your data, or your people.

## What we do

From "is AI even worth it?" to "it's running." Three stages, with AI woven through — on top of the infrastructure and network engineering we'd do anyway. Start with one, or take all three.

1. **Assess — Find where AI actually pays off.** We look at your infrastructure, your workflows and your goals, then tell you straight: where AI and automation are worth it, where they aren't, and what to do first. You leave with a prioritised plan you can act on — not a hype deck.
2. **Build — Build the AI and automation in.** AI assistants that answer real questions about your infrastructure in plain English; AI-assisted changes that catch mistakes before they ship; automation that turns alerts into action, wired into the monitoring, ticketing and tools you already run; documentation and config that keep themselves current. Built to fit your setup, and able to run on infrastructure you control when the data is sensitive.
3. **Enable — Hand it over, or hand it off.** We train your team to run and extend it, with docs and runbooks they'll actually use. Or, if you'd rather not, we run it for you. Your call — and reversible either way.

## How we work

**A low-risk way in.** You don't have to bet the whole estate to find out if this works. We earn the next step, one piece at a time.

1. **We talk.** A short call about your infrastructure and what's pushing you toward AI. No pitch — if we don't think it's worth it yet, we'll say so.
2. **We start small.** One focused piece that proves its value fast: no big upfront commitment, and nothing you can't reverse.
3. **We scale what works.** What earns its place gets extended. What doesn't gets dropped. You see results before you ever bet big.

**The ground rules:** honest (we'll tell you when AI isn't the answer); incremental (small, reversible steps over big-bang rewrites); no lock-in (you keep the knowledge, the configs and the keys); on your terms (your infrastructure, on-prem or cloud).

## Proof — what we've actually done

No logos, no name-drops — just the work. A sample of what we've built and delivered across real infrastructure.

- **AI put to work in production** — AI assistants and automation that put infrastructure knowledge a question away, and act on it safely.
- **AI servers, built and deployed** — designed and stood up self-hosted AI; capable models running on hardware the owner controls.
- **Infrastructure migration** — moved live production infrastructure onto new platforms with the service staying up.
- **Repeatable deployments** — replaced hand-built, one-off setups with automated, repeatable deployments.
- **Codification** — captured infrastructure and configuration as code: reviewable, repeatable, recoverable.
- **Full-stack monitoring** — end-to-end visibility across infrastructure; blind spots closed before they bite.
- **Redundancy & functionality testing** — proved failover and function actually work, before an incident does the testing.
- **Security analysis** — reviewed infrastructure and configuration for weaknesses, and closed the gaps.
- **Documentation that stays true** — documentation that reflects reality and keeps itself current.

## Why HG Fabian

**Ahead on AI. Grounded in real infrastructure.** Plenty of people can talk about AI. Fewer have spent their career inside the infrastructure it's meant to run on. You get both — and one person accountable for the result.

- **Direct & accountable** — you deal directly with the person responsible for your project; no account managers, no being handed to a junior team. Where extra specialists are the right call, they're brought in deliberately, and accountability stays in one place.
- **Real, not theatre** — we build things that reach production and survive contact with a live environment. When something's hype, you'll hear it plainly.
- **Built to outlast us** — documented and transparent by default, so your team can own and extend it long after the engagement ends.

Plugs into your stack: your monitoring & observability, your ticketing & ITSM, your CI/CD & pipelines, your config & version control, your existing automation, your home-grown tools.

## Contact

**Not sure where AI even fits? That's the conversation.** Book a short call and tell us about your infrastructure. You'll get an honest read on what's realistic, where we'd start — and whether we're the right fit. No pitch, no obligation.

- Email: <info@hgfabian.com>
- For ISPs, operators & enterprise infrastructure teams · based in Sofia, Bulgaria, working across the EU.
- HG Fabian · UIC 203910385