HexaBox Talk to us
HexaBox symbol

HEXABOX

Unleashing the value of Artificial intelligence

Capability is exploding. The value stays locked until AI can act under control
whichever model wins. HexaBox is that control.

Keep reading ↓ The system →
Controlling the action

We control the action.
Not the intelligence.

Enabling the enterprise layer to harvest the efficiency, productivity and profitability wins of autonomous AI. Governed fleets of agents, working 24/7 at machine speed.

Only 8% of enterprises report established return on their AI investments (KPMG Global AI Pulse, 2026). The gap is not capability. It is governed execution.

Create governed agents with HexaBox.The Agent Factory builds them from a job description, with the declared rules inside. The next one inherits the connectors and rules already in place — a fleet grows one description at a time, and an agent type built for one customer travels to the next.
Connect existing agents to HexaBox's action controls, human approvals and verifiable record.Agents built by anyone else are connected during implementation — and then pass the same gate, land on the same record and appear in the same seat.
On top of what you already run.HexaBox does not replace your AI, your platforms or your business systems — any model, any vendor, your cloud or your customers' own servers. The intelligence is a replaceable part: the newest model for the hard cases, a lean one for the routine, swapped as they improve. The rules, the record and what the fleet has learned stay.
Your brand in frontThirty invention families · patent‑pendingMade in Sweden
The system

Three parts. Create, control, run.

Three parts, in plain words. The full mechanism is on The system page.

Create · the Agent FactoryDescribe the job in your own words — any specialist, any sector. The factory builds the agent's structure from proven parts, with the declared rules and approval requirements inside. The factory is itself a HexaBox agent — the first one built, and the one that built Eihwaz, our AI-driven cyber-security fleet.
Control · the Control PlaneEvery action an agent proposes carries an authority level. Before a consequential action can execute, the gate checks identity, mandate, level, evidence and approvals. The outcome — execute, hold or deny — is recorded.
Run · the Driver SeatYour operators run the fleet from one screen: what ran within mandate, what waits for authority, one button that suspends further agent actions across the fleet. The same seat, limited to one customer, is that customer's own view.
The intelligence evolves · the rules stay in place

AI is getting better every month. Your system gets better with it.

Keep pace with AI. Keep control of the actions. HexaBox separates the choice of AI model from the rules that govern its actions — Anthropic, OpenAI, a fine-tuned small model, your own. New models are introduced as they become useful, while the declared mandates, approval requirements and record stay in place. What you build today keeps its value as the models improve. How that works →

What sets it apart

The moat is the system, not the component.

Build→Govern→Authorize→Approve when required→Prove

The same declared rules govern how agents are built, which actions they may take and when human approval is required. Operators see those rules in the Driver Seat, and their application can be verified in the record. HexaBox connects construction, action, authority and record through the same governing rules — there is no second place where a rule can be written differently.

It is also what makes change manageable. Change a rule in one place. For Factory-built agents, rule changes trigger regeneration and re-signing; for connected agents, the execution controls and the record follow the updated declaration. What was re-signed is visible on the record.

One declaration keeps the rules consistentThe governance requirements your customer declares — their sector's rulebook turned into policy by your people — are built into the agent, checked at the action and shown in the seat. One source, every part.
Enforced at the actionNot a document someone has to remember: the control plane checks each consequential action against the declared rules before it executes.
The record makes it verifiableWho proposed, who approved, what ran — one tamper-evident chain per customer. How the rules were applied can be verified by their auditor without trusting you or us. Evidence is what we ship; certification is what an auditor grants.
One action, step by step

The agent proposes an action. HexaBox checks whether it is authorized to execute.

Agents act autonomously within their mandate. Actions requiring additional authority wait for the required approval. Three illustrated scenarios show how Eihwaz handles a routine lookup (pre-authorized — executes and is recorded), a laptop isolation requiring approval (a named person approves) and an unauthorized data transfer (outside the mandate — never runs).

AI AGENT PROPOSES SECURITY AGENT · 03:12 Isolate the laptop ransomware · high confidence HEXABOX identitysigned · valid mandatewithin scope levelT3 · named person evidencesigned assessment approvalwaiting FIVE CHECKS · EVERY ACTION · EVERY PATH EXECUTEwithin mandate · on the record HOLDa person decides DENYoutside mandate · never runs ✓ a named person approvessigned · single use · recorded HEXARECORD who proposed · who approved what ran · one chain per customer checkable without HexaBox
The gate · three illustrated scenarios from Eihwaz, our cyber-security fleet1 of 3
Four levels of authority
T1Pre-authorized — executes and is recorded.
T2One person confirms.
T3A named, logged-in person approves.
T4Two independent people, two keys — never automatic.

The action decides its level — never the one who asks. An approval is bound to the exact action and approver, valid once, signed with a key the agent never holds.

The gate, in full →
Not a black box

See every agent at work. The Wind Shield.

The view out through the windscreen of the Driver Seat: sources in, the specialists at work, the gate, the actions out — every event travelling the chain. Here it draws a simulated fleet so you can touch it: drag to orbit, tap a container to open its graph. Behind the login it draws your fleet, live, and shows problems as they are.

  1. A case comes inFollow an event from a source on the left, through Fleet Command and Dispatch, to the specialist that reads it.
  2. An action needs a yesAfter Door 2 and the SOC pipeline the proposal reaches the gate — the ring marked 'a person decides'. Pre-authorized steps pass; the rest wait.
  3. The decision is on the recordWhat leaves towards 'Response actions' has been decided, and the Ledger at the bottom holds the receipt. The autonomy gauge counts what ran on its own.
The Wind Shield: a fleet drawn as containers — sources on the left, command above, the specialists, the gate where a person decides, the actions on the right

A simulated fleet with simulated events — on this public site by design. Zoom out to the fleet of fleets, in to one agent's graph, down to one part and its contract.

Open full screen ↗The four zoom levels →
Evidence

See how agents are built. See Eihwaz in operation.

The factory · construction

Watch the Agent Factory build an agent's structure.

Describe a job on this site and the real factory builds the agent's structure in about a minute: its parts, its declared rules, its approval requirements. What you see is how an agent is constructed — connecting it to a customer's systems and confirming their requirements is the implementation that follows.

See the factory at work →
Eihwaz · operation

See the architecture applied operationally.

Eihwaz is our AI-driven cyber-security fleet: a security operations centre (SOC) built on HexaBox, in production through a managed security provider. Every alert read and investigated, the response proposed, every consequential action gated, analysts in the Driver Seat.

Eihwaz, in production →

It is testable. Four tests — change a rule, revoke a mandate, alter a record, try to get an action past the gate — are run in the partner evaluation. They are described on The system.

Partnership

You bring the customers and the sector knowledge. We bring the system.

HexaBox is licensed to the companies that already deliver to every sector — IT-service companies, consultancies, cyber-security companies. You bring the customers, the implementation and the sector knowledge that turns a rulebook into a policy. HexaBox brings the control layer, the Agent Factory and the evidence. Your brand in front, your fleet, your customers kept separate.

01
DemonstrationYou see the generated agent structure, its declared rules and its approval requirements — on this site, or with us on your screen.
02
Customer implementationThe agent is connected to the customer's systems and the relevant business requirements are confirmed — your sector knowledge, our connectors.
03
DeploymentThe solution is validated and taken into operation within the agreed scope: your operators in the Driver Seat, the customer in their own seat.
The partner program →Eihwaz, ready to implement →
Contact

Let's talk about growing your business.

Tell us which sectors you deliver to and what your customers are asking for, and we show you what a governed-agent offering could add to your business — under your brand, with the controls defined. Want Eihwaz, our AI-driven cyber-security fleet, in your offering? It is ready to implement.

Explore a partnership Eihwaz, ready to implement

or write to contact@hexabox.one

We control the action. Not the intelligence.