HEALTHCARE
Chorus Careflow
Care coordination with clinical-grade oversight. Human judgment stays central where it must; low-risk administrative work can graduate.
TYPICAL: LEVEL 1 — HUMAN-IN-THE-LOOP → LEVEL 2 — HUMAN-ON-THE-LOOP
Overview
Neuro-symbolic reasoning plus a graduated trust model — flexibility and auditability together.
Agentic Workspace
One workspace — files, chat, meetings, calendar, apps, dashboards — where humans and agents collaborate.
Neuro-Symbolic AI Harness
Symbolic guardrails make LLM reasoning deterministic and auditable.
Open Foundation
Spring Boot, Spring AI, Embabel, Pulumi — inspectable, deployable, extensible.
Governance, Compliance & Trust
ISO 9001/27001 discipline plus Open Policy Agent enforcement on every agent action.
Flywheel — Agentic Coding Harness
Spec to delivery to DevSecOps monitoring. One system, not three separate tools.
HEALTHCARE · LEVEL 1–2
Chorus Careflow
Care coordination with clinical-grade oversight — human judgment stays central where it must.
ASSET MANAGEMENT · LEVEL 2–3
CadenceCAM
Asset operations that earn their way to autonomy.
CUSTOMER EXPERIENCE · LEVEL 1–3
CatalystCXP
Customer experience that scales — routine resolved autonomously, novel cases escalate.
Neuro-Symbolic · Graduated Trust
The Agentic Platform for Human-Machine Collaboration
Everyone else is selling empowerment — the promise that AI or low-code removes the expert from the loop. Dynamo doesn't remove the expert. It removes the wait.
Every workflow runs at exactly the level of autonomy it has earned — from expert-supervised execution to policy-governed autonomy — inside a fully open foundation you can inspect, deploy, and extend yourself.
NEURO-SYMBOLIC AI CORE
AGENTIC WORKSPACE
OPEN FOUNDATION
DYNAMO · RUNTIME TOPOLOGY
TRUST LEVEL 2
Human Intent
CAPTURED
Dyna
STRUCTURING
Structured Requirements
ATDD · 14 CRITERIA
REASONING
Neuro-Symbolic Core
Policy Gate
OPA · EVALUATED
Flywheel
SPEC → DELIVERY
Agent
EXECUTING
DevSecOps Gates
BUILD VERIFIED
Production
DEPLOYMENT READY
Observability
RUNTIME OBSERVED · LEARNING LOOP OPEN
{{ ev.label }}
{{ ev.actor }}
THE GRADUATED TRUST MODEL
Autonomy isn't a setting you flip once.
It's a level you assign per workflow — and raise as trust is earned. Select a level to see how the same deployment workflow behaves, and how the expert's role changes with it.
{{ level.status }}
{{ level.definition }}
AGENT
{{ level.agent }}
HUMAN
{{ level.human }}
WORKFLOW · DEPLOY SERVICE UPDATE
{{ level.humanMode }}
{{ st.i }}
{{ st.label }}
{{ st.state }}
EXPERT ROLE
{{ level.role }}
RISK POSTURE
{{ level.risk }}
AUDIT
Every action recorded
The human never disappears. The role changes: approve every step → approve critical decisions → define policy and monitor exceptions.
PLATFORM ARCHITECTURE
Three ideas. One platform.
The trust model governs the workspace. The neuro-symbolic core supplies the reasoning and the constraints it runs inside. Nothing here is an add-on.
01
REASONING
Neuro-Symbolic AI Core
Symbolic guardrails make LLM reasoning deterministic and auditable — the foundation the trust model runs on.
NEURAL
×
SYMBOLIC
DETERMINISTIC OUTCOME
02
GOVERNS EVERYTHING
Graduated Trust Model
Three explicit autonomy levels that every workflow can be assigned to — and can graduate through as trust is earned.
L1 IN-LOOP
L2 ON-LOOP
L3 AUTO
03
WHERE WORK HAPPENS
Agentic Workspace
Files, chat, meetings, calendar, apps and dashboards in one environment — humans and agents in the same place.
SPRINGBOARD · TRUST LEVEL PER ROLE
LIVE EXECUTION TRACE
Every decision leaves a record of who made it.
INTENT
“Prepare the production release for the claims-processing service.”
{{ traceState }}
{{ t.i }}
{{ t.label }}
{{ t.actor }}
INSPECTOR
{{ traceLabel }}
{{ traceDetail }}
RESPONSIBLE
{{ traceActor }}
CONFIDENCE
{{ traceConf }}
RISK
{{ traceRisk }}
TRUST LEVEL
2 · ON-LOOP
POLICY
OPA v42
EVIDENCE
ATTACHED
Automation moves through everything that doesn't require judgment. Steps 07 and 11 are held for expert review, because that is where judgment lives.
NEURO-SYMBOLIC AI HARNESS
Reason probabilistically.
Operate deterministically.
NEURAL
Understanding
Reasoning
Context
Planning
SYMBOLIC
Rules
Policies
Constraints
Permissions
Auditability
Dynamo Neuro-Symbolic Harness
Neural interpretation proposes. Symbolic constraints decide what is permitted. The trust model decides who confirms it.
Pure-LLM agent frameworks are flexible and hard to audit. Pure rules and BPM engines are auditable and rigid. Dynamo runs both layers as one system, which is what makes graduated autonomy safe to offer at all.
REASONING PATH
01Natural-language intent
02Neural interpretation
03Structured representation
04Symbolic constraints
05Policy evaluation
ALLOWED
ESCALATE
BLOCK
06Agent execution
07Audit trail
THE DIFFERENCE
Everyone else is promising to remove the expert. We're removing the wait.
System development is expert work. What should change isn't who does it — it's how fast the loop runs between what you need and what gets built.
THE EMPOWERMENT LOOP
Business request
Interpretation by a tool, not a person
AI-generated output, at volume
Review burden moves downstream
Uncertain production readiness
Independent 2026 engineering-analytics research from Waydev and LinearB converges on the same finding: AI increases the volume and variability of generated code, and when coding accelerates the bottleneck doesn't disappear — it moves to review.
THE DYNAMO LOOP
Customer intent, in plain language
Dyna — Level 1, Human-in-the-Loop
Structured ATDD requirement & acceptance criteria
Flywheel — automation where judgment isn't needed
Breezeware expert engineering
DevSecOps gates → trusted delivery
The feedback loop that gets faster is the human-to-human one: customer intent to expert delivery, accelerated by automation everywhere automation is safe.
Faster feedback loops
NOT “GENERATE MORE CODE”
THE PRODUCT FAMILY
Four names, one loop.
Dynamo is the platform. Dyna is how customers tell us what they need. Flywheel is how our engineers build it. CARE is where customers experience the whole loop.
CARE PORTAL — CUSTOMER VISIBILITY SURFACE
01
Customer
Describes what they need
02 · LEVEL 1
Dyna
Plain language becomes stories and acceptance criteria
03
Dynamo trust layer
Assigns the autonomy level, enforces the policy
04
Flywheel
Spec to delivery to DevSecOps monitoring
05
Expert engineering
Breezeware engineers own the judgment calls
06
Production
Verified against the customer's own criteria
We're not a tool you license and hope your team can make work. We're the team, and the tool, together.
A prospect evaluating a licensed platform has to ask “can my team build this well with this tool?” A prospect evaluating Dynamo asks “can Breezeware deliver what I need?” — a structurally different vendor relationship.
TECHNOLOGY LICENSING
TECHNOLOGY-ENABLED SERVICE
Access to a platform
A delivered outcome, built by experts
Customer or SI partner builds
Breezeware engineers build, via Flywheel
Risk sits with the customer at signature
Risk sits with Breezeware, through delivery
Build quality invisible to the vendor
Visible in CARE, end to end
AGENTIC WORKSPACE FOR HUMANS
Humans and agents occupy the same operating environment.
Files, chat, meetings, calendar, Springboard, dashboards — one workspace, where the trust level is visible and adjustable at every surface.
DYNAMO WORKSPACE
TRUST LEVEL 2 · HUMAN-ON-THE-LOOP
WORKSPACE
Chat
Files
Meetings
Calendar & Events
Springboard
Applications
Dashboards
OPERATIONS
Agents
Tasks
Approvals
Policies
Observability
ACTIVITY — RENEWALS AUTOMATION
PM
“Start the Q3 renewals workflow for the enterprise segment.”
HUMAN · 14:02
AG
Agent started. Referenced Renewals-Playbook-v7.md and 212 open contracts.
AGENT ACTIVE · CONTEXT SYNCHRONIZED
PO
Policy gate evaluated the pricing-change request against the discount ceiling.
POLICY EVALUATED · CONSTRAINT SATISFIED
EX
Two accounts exceed the ceiling. An expert joined the 14:30 review and approved one, escalated one.
ESCALATING TO EXPERT · EVIDENCE ATTACHED
FW
Flywheel task created for the contract-terms change. Dashboard updated.
EXECUTION AUTHORIZED · AUDIT RECORDED
SPRINGBOARD — CONTROL PLANE
Production Deployment
AGENT: RELEASE-AGENT · ENV: PRODUCTION
LEVEL 2 — HUMAN-ON-THE-LOOP
ALLOWED
Run tests
Build artifacts
Scan dependencies
Prepare deployment
REQUIRES HUMAN APPROVAL
Production deployment
BLOCKED
Modify security policy
Springboard is where trust levels are assigned per role and per workflow — a no-code task for the customer.
OPEN FOUNDATION
Inspect it.
Deploy it.
Extend it.
Certifications are attested. An open platform is provable: your security team can read the code that enforces access control and data handling, and run it on infrastructure you control.
EXPERIENCE
Agentic Workspace
AGENT SYSTEM
Dynamo
REASONING
Neuro-Symbolic Harness
POLICY
Open Policy Agent
APPLICATION
Spring Boot · Spring AI · Embabel
INFRASTRUCTURE
Pulumi
CLOUD
Customer-controlled environment
GOVERNANCE, COMPLIANCE & TRUST
Policy is code, evaluated on every agent action.
Here is an actual evaluation. The agent requests a production database write at Level 2 — Human-on-the-Loop.
REQUEST
WRITE production database
environmentproduction
agent_rolerelease-agent
permissiondatabase-write
trust_level2 — human-on-the-loop
approval_present{{ gov.approvalVal }}
OPEN POLICY AGENT — RESULT
{{ gov.result }}
{{ gov.reason }}
Decision, inputs and evidence written to the audit trail
Same policy bundle enforced in every environment
Operated under ISO 9001 and ISO 27001 discipline
FLYWHEEL — SPEC TO DELIVERY
Two lanes. Automation moves; judgment decides.
The same spec-to-delivery and DevSecOps harness Breezeware runs its own company on.
Intent
Dyna
Structured stories
Acceptance criteria
Architecture
Implementation
Tests
Security
Policy
Deployment
Monitoring
Feedback
SOLUTIONS
Same model, three industries.
Each solution sits where its risk profile earns it — and each can graduate.
ASSET MANAGEMENT
CadenceCAM
Asset operations that earn their way to autonomy. Proven high-volume flows run at Level 3; exceptions fall back to a checkpoint.
TYPICAL: LEVEL 2 — HUMAN-ON-THE-LOOP → LEVEL 3 — AUTONOMOUS OPERATIONS
CUSTOMER EXPERIENCE
CatalystCXP
Customer experience that scales. Routine interactions resolve autonomously; novel cases escalate to a person with the context attached.
PER WORKFLOW: LEVEL 3 FOR ROUTINE, LEVEL 1–2 FOR ESCALATIONS
BUILT IN THE OPEN
Nothing to take on faith.
RELEASE
v2.14.0
Latest tagged platform release
BUILD STATUS
passing
main · DevSecOps gates green
POLICY TESTS
312 / 312
OPA bundle regression suite
MODULES
open source
Workspace, harness, IaC, policy bundles
ENVIRONMENTS
customer-run
Kubernetes-native, cloud-agnostic via Pulumi
Figures shown are placeholders wired for live repository data.
Start with the level of autonomy you trust.
Dynamo lets your workflows earn their way from expert-supervised execution to policy-governed autonomy. We don't skip levels either.
LEVEL 1
Human-in-the-Loop
Approve every step
LEVEL 2
Human-on-the-Loop
Approve critical decisions
LEVEL 3
Autonomous Operations
Define policy, monitor exceptions