Services

Your workflow.
A system built around it.

We scope and integrate agents around your documents, business software, websites or technical equipment. The audit determines what can run locally and where cloud capability would help. You own the hardware; we install the system and maintain it through an agreed care plan.

Private workflows

ExecAI Private

Office, finance operations, websites and customer care

  • Knowledge assistants, document preparation and reports
  • Invoice extraction and bookkeeping preparation for review
  • Customer support, website and inventory workflows
  • Integration with your tools; local or hybrid after audit
  • Named users, per-role access, audit

Typical setting: back-office, professional services, online business or technical team.

Technical operations

ExecAI Infrastructure

Energy, machines and technical installations

  • Solar, inverter, batteries, transfer switch, meters
  • Per-cell diagnostics and drift tracking
  • Predictive thresholds: “critical at 16:00”
  • Daily and monthly reports
  • Questions answered with the figures of the base

Typical site: off-grid home, farm, telecom site, small plant.

Physical security

ExecAI Guardian

Physical security and site supervision

  • IP cameras integrated, on-site detection
  • Zones, trajectories, confidence, passage files
  • Natural-language search over weeks of history
  • Alerts and a dated journal
  • Human verdict on every contested call

Typical site: remote property, yard, warehouse, worksite.

Engagement

Four phases. Each one ends with something you can check.

01

Audit — one or two days

On site, or by video walkthrough for a first filter. We map the workflow, systems, privacy needs and network constraints, then assess whether a local model meets the task’s requirements. Output: a written scope with a fixed price for the pilot.

02

Pilot — four to six weeks

One site, one use case, one measurement agreed in advance (time saved, incidents caught, unplanned outages avoided). The system runs for real during the pilot; it is not a demo licence.

03

Rollout — quoted per site

Hardware sized and installed, integration with existing equipment, users created, documentation written, your team trained on the console. Priced after the audit, never guessed before.

04

Care — monthly or yearly

Monitoring of the system itself, updates, model and threshold tuning, journal review, and an on-call path. Supervision that nobody maintains quietly stops being supervision.

The commercial model

You own the system.
We keep it alive.

The hardware is bought by you, in your name. The installation is a fixed-scope project, quoted after the audit. Care is a subscription — month or year, cancellable at term. No per-seat licence on your own infrastructure, no fee proportional to the number of cameras you already paid for.

This is deliberate. A system that only works while a supplier is paid is a hostage, not a tool. We would rather be kept for the quality of the care than for the impossibility of leaving.

What is not included

  • Hosting of your data. Nothing runs in our premises; the base stays on your site.
  • The hardware itself. Cameras, batteries, meters, computers — bought by you, at cost, with our specification.
  • Your network and your electricity. We specify what the system needs; your installer provides it.
  • Promises we cannot measure. If a claim cannot be verified during the pilot, it does not go in the proposal.
Pilot first, always. A pilot costs less than a mistake and settles the question with evidence — including the possibility that the answer is “AI does not help here”. We will say so.
Questions

The ones that come up first.

Do we need an internet connection?

A fully local workflow can operate without Internet when its models, tools and data are all on site. For example, energy supervision and local alerts can continue at a remote installation. Cloud inference, connected websites, remote phone notifications and remote support need a working connection.

Can we keep our cameras and equipment?

Usually yes. Standard IP cameras, MQTT-capable meters, serial or network inverters: we integrate what you already own before proposing anything new.

Where does the data actually go?

In a fully local workflow, prompts, documents and operational data are processed on your infrastructure without being sent to a model provider. For a hybrid workflow, the selected context is sent to the configured cloud provider. The audit and proposal specify those data flows before deployment.

Does local AI need an API subscription?

Local inference requires no model-provider API subscription and has no per-token or per-request provider charges. You still budget for hardware, electricity and the agreed maintenance plan. Cloud-model usage is charged separately when included in the project.

How do you choose local or cloud?

We test the model against your actual task, then weigh quality, privacy, connectivity, latency and operating cost. The result can be fully local or hybrid, with local agents handling routine work and cloud models reserved for selected tasks.

What hardware do you need?

For most sites, a small industrial computer and a disk. For video, the same machine is usually enough for a handful of cameras; we size it during the audit and give you the specification.

Is it secure?

Named users through single sign-on, per-role access, hardened services with restricted write access, and no inbound connection required from the internet. The details are part of the audit document.

What happens if we stop working with you?

The system keeps running; only the care subscription ends. Your data is exportable in open formats. That is written into the contract, not promised in conversation.

Tell us the site. We will tell you the pilot.

Two sentences are enough to start: what the site is, and what you cannot see today. We answer with a scope, a price and a date — or with a reason to stop.