Managed AI virtual assistants
Customer-specific digital workers that operate approved SaaS, CRM, browser, desktop, phone, messaging and legacy systems.
- Controlled workspace
- Communications identity
- Permissions and approvals
- Managed operation
AI FORGE / AGENTS X
Don’t add AI to every system. Give your business one assistant that can operate them all.
Agents X creates persistent digital workers with controlled access to browser, desktop, phone, messaging, APIs, SaaS, CRM and legacy tools, while your team keeps control of consequential decisions.
✓ Controlled private alpha✓ Customer-specific identity✓ Cross-system operation✓ Human approval built in
BROWSER✦DESKTOP✦PHONE✦WHATSAPP✦CRM✦APIS✦LEGACY SYSTEMS✦
THE FRAGMENTATION PROBLEM
Your software does not work as one business. Your people make it work as one.
Most AI is being added separately inside each application. That creates more disconnected assistants, more subscriptions and the same broken hand-offs between systems.
CRMits own AI and partial contextISOLATED
EMAILanother AI and another workflowISOLATED
PROJECTStasks still moved by peopleISOLATED
LEGACYno useful AI interface at allMANUAL
AGENTS Xone persistent operating context across the stackCONNECTED
A DIFFERENT MODEL FOR BUSINESS AI
One intelligent operating layer across the stack.
Your assistant carries the working context between approved systems. It can use an API when that is the safest route, operate a browser or desktop when no suitable API exists, communicate through approved channels, and report the outcome back to the right person.
- PersistentRetains role, context, tasks and follow-through.
- Cross-systemWorks between software instead of inside one app.
- PermissionedReceives only the access required for its role.
- ManagedMonitored, supported and improved by AI Forge.
AI FORGE / TWO CONNECTED CAPABILITIES
Operate what the business has. Build what it needs next.
AI FORGE UK is the systems architect and managed-service operator. Agents X is its flagship operating system. When the existing stack is inadequate, AI Forge can build a new agent-friendly platform as a separately scoped project.
New systems designed for humans and agents
Custom SaaS and business platforms built when current tools cannot support the required workflow. The same assistant operates it alongside the old stack.
- Agent-friendly architecture
- Human control surfaces
- Clear event and audit records
- Separately scoped builds
A CONTROLLED WORKING ENVIRONMENT
More than a chat window.
A real place to work.
Each capability is provisioned for the role, not switched on by default.
Computer and browser
Operate approved web and desktop systems through normal interfaces when APIs are unavailable.
Phone and messaging
Use approved telephone, voice, WhatsApp, Telegram and messaging channels for the role.
APIs and webhooks
Choose reliable machine interfaces where they offer the safest and clearest execution path.
Knowledge and context
Work from customer-specific documents, processes, permissions and operating instructions.
Tasks and follow-through
Track work, prepare next actions and move relevant information between systems.
Human approval
Pause sensitive communications, changes, commitments or payments for the right person.
OLD SYSTEMS / NEW SYSTEMS
Keep what works. Replace only what holds the operation back.
One assistant can become the operating bridge between the software a business already has and an AI-native platform built for what it needs next.
See how AI-native systems fit ↗EXISTING STACKLegacy browser appCRM and SaaSDesktop software
AX
AI-NATIVE STACKAgent-friendly workflowsHuman control surfaceEvents and audit trail
FROM REQUEST TO CONTROLLED EXECUTION
Capable for daily work. Controlled around risk.
The assistant does not use autonomy for its own sake. It chooses the safest reliable operating route, records what happened and keeps the right human responsible for consequential outcomes.
Follow the operating flow ↗01Request or business eventCONTEXT RECEIVED
02Assistant understands role and permissionsPOLICY CHECK
API / WEBHOOKBROWSER / DESKTOPPHONE / MESSAGEOPERATOR ESCALATION
03Human approval when requiredSAFE PAUSE
04Action, record, status and follow-throughVISIBLE OUTCOME
A BROAD OPERATING SYSTEM / A FOCUSED START
Start with one valuable role.
Expand after it proves useful.
The assistant receives only the systems, information and permissions required for the first job.
Enquiry & follow-upOperations & administrationRecruitment coordinationFounder operationsProperty operationsResearch & reportingKnowledge & documentsSupport preparation
BUILT BY OPERATING THE MODEL
An internal reference implementation.
Not a concept slide.
AI Forge operates the underlying paradigm in daily work: a dedicated computer environment, real communications infrastructure, telephone capability, WhatsApp, browser and desktop operation, messaging, tool use and agent-directed execution.
Agents X is in controlled private-alpha development. Customer pilots begin with a tightly scoped role, explicit permissions and founder-led supervision. Capabilities are provisioned individually according to security, data and approval requirements.
Persistent task, run, event and approval recordsFOUNDATION LIVE
Customer and operator control surfacesPRIVATE ALPHA
Tenant-shaped worker and workspace architectureIN TESTING
Automated platform verification41 TESTS PASSING
Carefully scoped founder-led pilotsREADINESS GATED
STRAIGHT ANSWERS
Before an assistant
gets to work.
Broad capability should be introduced through explicit scope, not vague promises of autonomy.
Try the live Discovery Assistant ↗Is Agents X a chatbot?+
No. Chat or messaging can be one interface, but the product is a managed operating layer that can work across approved software, communications channels and computing environments.
Do we need to replace our current systems?+
Usually not. Agents X can operate useful existing tools. If a system cannot support the required workflow, AI Forge can separately scope a new AI-native platform.
Can it act without approval?+
Only within the boundaries designed for the role. Sensitive outbound actions, infrastructure changes, financial or legal commitments and other consequential decisions require explicit controls or escalation.
Is this ready for every business?+
No. Agents X is in controlled private-alpha development. We consider carefully scoped founder-led pilots after the relevant security, workflow and readiness gates are satisfied.
ONE ROLE / ONE VALUABLE WORKFLOW
Design your first
AI virtual assistant.
Show us where work crosses systems, channels and people. We’ll help define a sensible first role.