AgentPay / LABS

Intelligence,
out in the
world.

Exploring how agents perceive, converse, act, and learn. Connecting human intent to machine action, with control at every step.

Concept illustration of a silver quadruped robot on a terracotta plateau in a quiet landscape
SOFTWARE AGENTS ↔ PHYSICAL SYSTEMSResearch in progress · London

From conversation to consequence.

Concept artwork of a quadruped robot observing a greenhouse environment

01 — Embodied fieldwork

Quadruped research: observation, navigation and site-specific workflows.

Coral and sage ribbons crossing in a sculptural representation of continuous conversation

02 — Continuous interaction

Listen while speaking. Interrupt, correct and keep the task in context.

One research harness.
Different physical environments.

Research concepts · hardware validation remains a next step.

Notice what changed. Make the next step clear.

A quadruped revisits inspection points, flags a visual change and asks a person to confirm the task. The harness records the observation, decision and outcome.

WHAT TO MEASURE Detection quality · intervention rate · time to resolve

A conversation is
only the beginning.

The middle layer connects camera observations, human intent, model reasoning and fleet execution. Every outcome becomes evidence for the next experiment.

ILLUSTRATIVE ARCHITECTURESELECT A NODE TO EXPLORE ↗
ROBOT FEEDBACK →Versioned rollout ↩

Fast loop: observe → decide → bounded skill. Learning loop: episodes → training → validation.

A missed detection becomes a synchronized episode: video, robot state, action and outcome. Review the cause before changing the policy.

Agency needs
boundaries.

AgentPay gives software agents a payment interface. Card or crypto. One key. You set the limit.

ILLUSTRATIVE POLICY
AgentRequestPolicyReceipt
Purpose
External analysis
Budget
Defined by the operator
Authorization
Required by policy
Settlement
Card or crypto

Notes from the lab.

Build. Observe.
Understand.

AGENTPAY / RESEARCH NOTE