Samix TechnologyAgentic AI overhauls

AI agents for businesses

I automate tasks in your business with AI agents, in production, not in a demo

Triaging incoming email, preparing quotes, matching invoices, chasing customers: a well-scoped AI agent takes over these repetitive tasks, on your data and under your control. I design them, put them into production and hand them over to your team.

What an AI agent can do for you

An AI agent is not a chatbot: it is a program that reads, decides and acts in your tools, within rules you set. It helps wherever someone spends their days copying, sorting and chasing. For example:

  • read incoming email, sort it, and prepare a reply or a task for the right person;
  • turn a quote request into a quote ready for approval in your software;
  • extract supplier invoices, match them to orders and flag the gaps;
  • chase customers and unpaid invoices at the right time, in the right tone;
  • keep a tracking sheet or a CRM up to date from what arrives by email.

Three ways to start

AI agents workshop

€800 excl. VAT

1 day

Find the tasks in your business that an AI agent can take over, and where to start.

  • 3 to 5 processes reviewed with you
  • Ranked by gain and by risk
  • A concrete start plan for the first one

Inbound requests agent

€3,000 excl. VAT

fixed, 3 days

An agent that triages and prepares your incoming emails, quote requests or invoices, on your data.

  • An agent in production, not a mock-up
  • Human approval before anything is sent or done
  • Documentation and handover to your team

Hours bank

€100 excl. VAT / h

pack of 20 h (€2,000 excl. VAT)

Help on demand to extend an agent, add another one or unblock your team.

  • Use it at your own pace
  • Transparent time tracking
  • By email, video call or on site

Prices exclude VAT. On-site work available in the south of France.

The proof: Samix HQ, a company run by AI agents

Samix Technology itself runs on a team of AI agents, Samix HQ. Each agent has a role (prospecting, bookkeeping, marketing, development, server operations), its own tools and its own limits. They share a knowledge base, hand work to each other and ask for my approval before anything that commits the company: an email to a client, a purchase, a change in production. Every action is logged.

It is not a demo: it is what runs my business every day. The rules that make it reliable (scoped access, human approval, an audit trail, rollback) are the ones I apply for clients.

The same methods underpin Agentic DBA, my AI agents for SQL Server operations, which monitor, diagnose and prepare fixes for a DBA to approve.

Why trust me

I am Evan Barke, a SQL Server and data consultant for more than 15 years and an author of technical articles on SQLShack. I have spent my career on production databases where a mistake is expensive, and I bring the same rigour to AI agents. An agent that touches your data must be as safe as a careful DBA: minimal access, no sensitive action without approval, and always a way back.

How I work

  1. Scope: the tasks to hand over, the data involved, the approval rules.
  2. Build on your real data, with your team, rather than on a demo example.
  3. Go to production with human approval, then loosen the reins task by task as results come in.
  4. Hand over: documentation, access and handover, so the agent belongs to you.

Let's talk about your tasks

Email me directly at evan@samix-technology.com, book a call, or describe what you need below: I reply within one business day.

We use your details only to answer your request. They are never sold or used for automated prospecting. See our privacy policy.

Samix Technology, Evan Barke (sole trader, EI), SIRET 914 474 358 00019, Saint-Cyprien, France. Legal notice.

Trusted by

  • Synertrade
  • Azura Group
  • MagicOrange
  • Daly Credit Solutions

Frequently asked questions

Do we need a technical team to use an AI agent?

No. I handle the setup and hosting, and your team uses the agent from tools it already knows: its inbox, a shared sheet or a simple approval page.

What about the confidentiality of our data?

The agent only reaches the data its task needs, through a dedicated account. Which AI model is used and where it is hosted is decided during scoping, according to your confidentiality requirements.

Can an agent send an email or pay for something without us?

No, unless you decide so for a specific task. By default the agent prepares and someone on your team approves with one click before anything is sent, paid or changed.

What if the agent gets it wrong?

Sometimes it will, like a person. That is why every action is logged, sensitive actions go through human approval, and the agent flags what it is unsure about instead of guessing.

Where do we start?

With the one-day workshop (€800 excl. VAT): together we find the tasks most worth automating, and you leave with a plan, whether or not you carry it out with me.

A SQL Server problem to solve?

Describe your setup to a SQL Server expert: we reply within one business day with a first read of the situation and a clear proposal.