INTELLIGENCE
Models
Reason across enterprise context.
The Agents Are Arriving.
Governance Has to Arrive First.
Enterprise AI is crossing a boundary: from answering to doing, copilots to operating agents, recommendations to actions, isolated automation to multi-agent organisations.
SHIFT
The question has changed.
Yesterday: Can AI do this task? Today: Should this agent be allowed to do it? Under whose authority? Within what limits? Who is accountable? What happens when it is wrong? These are organisational questions, not model questions.
STACK
Capability is accelerating. Authority is the unfinished layer.
Models provide intelligence. Agent frameworks provide orchestration. Protocols provide interoperability. ERP provides records. Financial rails provide execution. Enterprises still need durable identity, mandate, delegation, human hierarchy, policy and evidence.
AGENCY
Frameworks
Plan and operate across tools.
RECORD
ERP
Remain systems of record.
EXECUTION
Financial rails
Move authorised value.
GOVERNANCE
★GARD
Identity, mandate and authority.
ORGANISATION
KYD
Humans + agents in one hierarchy.
CONSEQUENCE
Finance makes the problem consequential.
An incorrect generated paragraph can be edited. An unauthorised financial action may already have consequences. Finance requires governance before action, not merely observability after it.
HIERARCHY
From human-in-the-loop to human-in-the-hierarchy.
Putting a person behind every agent decision defeats much of the value of agency. Removing people from authority is unacceptable. Humans retain organisational authority. Agents receive bounded delegation. Exceptions reach named humans. Material judgement remains human.
THE MISSING LAYER
The enterprise stack has intelligence, agency and rails. It still needs governed authority.
GARD's category thesis is that finance needs a distinct layer between agentic intent and consequential enterprise action.
CATEGORY MAP / CONCEPTUALAGENTIC GOVERNANCE
01ModelsIntelligence
02AgentsAgency
03Enterprise softwareSystems of work
04Financial railsExecution
05 / GARDGoverned authorityIdentity · mandate · hierarchy · evidence
GARD isn’t putting AI into Finance. GARD is making Finance agentic.
★GARD / THESIS — HUMAN AUTHORITY · AGENTIC EXECUTION · CONTINUOUS GOVERNANCE
BUILD
We’re building for an enterprise that doesn’t fully exist yet.
Humans and agents will increasingly work inside the same organisations. Someone has to build identity, authority, hierarchy and governance for that world. We’re starting with finance.
Reported context — enterprise agent exposure and financial-crime loss
Enterprises running agents reporting an agent-related incident65%
Cloud Security Alliance
Of those incidents producing direct financial loss35%
Cloud Security Alliance
Asset-misappropriation cases that are billing schemes86%
ACFE Report to the Nations
Year-on-year growth in reported cyber-enabled losses+26%
FBI IC3
0%REPORTED EXTERNAL CONTEXT — NOT GARD MEASUREMENT100%