TECHNOLOGY / WORLD RUNTIME
Build around the world your business actually operates in.
A supplier delay, a city incident, a team of AI agents: each can set off decisions beyond one team or system. With i5 World-Runtime, your developers can model the situation, replay how it changed, and compare responses before committing to one.
WORLD RUNTIME / IN PRACTICE
Explore a decision.
Port expansion / THE DECISION
More capacity. Which tradeoff?
Compare terminal expansion plans against cost, environmental constraints, and the authority to commit.
- Horizon
- 5 years
- Alternatives
- 3
- Decision owner
- Infrastructure team
Capacity gains only make sense alongside the conditions needed to deliver them.
- Demand forecast updated
A higher cargo forecast triggers a capacity review.
- Community feedback added
Local employment and construction impacts join the decision context.
- Permitting risk flagged
The alternate site conflicts with the example environmental constraint.
- Mitigation plan added
The current-site plan includes an emissions reduction package.
- Alternatives compared
Three fixed projections are ready for human review.
THE NEXT DECISION
Recommend current-site expansion
Take the capacity case and mitigation plan to the infrastructure team. A recommendation does not release funding or grant a permit.
Evidence & assumptions
Illustrative data adapted from supplied workbench designs. Values, event sequences and policy outcomes are fixed examples, not a recorded runtime run or measured results. The linked runtime source explains the underlying pattern; it does not verify these figures.
- Capacity and emissions changes are relative to the current terminal baseline.
- Costs are illustrative capital estimates; mitigation is assumed to work as modeled.
Supply network / THE DECISION
A delayed vessel. Three ways forward.
Port congestion puts electronics deliveries at risk. Compare service, time, and the added cost of a reroute.
- Disruption
- 48-hour delay
- Production
- 2 lines at risk
- Commitments
- 5 priority orders
The same shipment connects port capacity, production plans, and customer commitments.
- Order confirmed
Priority deliveries enter the shipment plan.
- Shipment departed
Components are in transit to the receiving hub.
- Weather advisory issued
Planners review exposure along the route.
- Congestion delays vessel
The team compares reroutes before changing delivery commitments.
THE NEXT DECISION
Recommend the Jakarta reroute
Protect service with the shortest projected delay, then obtain approval for the extra transport cost.
Evidence & assumptions
Illustrative data adapted from supplied workbench designs. Values, event sequences and policy outcomes are fixed examples, not a recorded runtime run or measured results. The linked runtime source explains the underlying pattern; it does not verify these figures.
- Delay is measured against the original delivery plan.
- Service level means the projected share of commitments met. Alternate port capacity is assumed available.
City response / THE DECISION
One incident. Connected services.
An industrial fire affects roads, transit, and hospital access. Compare coordinated response plans.
- Incident
- Riverside fire
- Transit
- 3 lines affected
- Residents
- 120k in area
Changing a route can protect hospital access while moving disruption elsewhere.
- Fire reported
The incident establishes a common operating context.
- Road closures initiated
The affected corridor is closed to through traffic.
- Air-quality advisory issued
Responders consider wider resident exposure.
- Transit service reduced
Three lines need alternative routing.
- Hospital access impact detected
Access becomes a constraint on response plans.
- Response plans compared
Alternatives balance access, disruption, and cost.
THE NEXT DECISION
Recommend coordinated rerouting
Bring transit and emergency services together around the access plan before routes change.
Evidence & assumptions
Illustrative data adapted from supplied workbench designs. Values, event sequences and policy outcomes are fixed examples, not a recorded runtime run or measured results. The linked runtime source explains the underlying pattern; it does not verify these figures.
- Hospital access is an illustrative share of trips maintaining access, not a clinical measure.
- Example thresholds: access at least 70%; impacts above 50,000 residents require review. These are not operational safety standards.
Agent coordination / THE DECISION
A campaign goal. Explicit authority.
Agents propose campaign changes. Compare potential reach with budget and publishing permissions.
- Goal
- Product launch
- Proposals
- 3
- Conflict
- CMS write access
A promising proposal still has to fit the permissions granted to the agents.
- Research updated
Market evidence joins the campaign context.
- Launch proposal drafted
Content and channel choices are proposed within the current scope.
- Targeting alternative added
An agent proposes a smaller adjustment to the existing campaign.
- Policy conflict flagged
A proposed expansion would need additional CMS permissions.
- Write-access request held
The review lead can inspect alternatives; no publishing occurs.
THE NEXT DECISION
Review the launch proposal
Keep the reach projection, budget, and permission checks together for the human reviewer. Hold the expanded-access proposal.
Evidence & assumptions
Illustrative data adapted from supplied workbench designs. Values, event sequences and policy outcomes are fixed examples, not a recorded runtime run or measured results. The linked runtime source explains the underlying pattern; it does not verify these figures.
- Reach changes are illustrative projections against the existing campaign.
- The agents have read access to customer data. New write permissions and publication require human authorization.
01 / Supply networks
Compare reroutes before changing the plan.
A delayed shipment affects inventory, production, and delivery promises. Your team could build a planning service that connects those dependencies and compares reroutes against the same disruption.
The public supply-network scenario rebuilds state from events and compares reroute alternatives, keeping cost and loss tradeoffs visible alongside a simulation-backed recommendation.
Your team would connect its own ERP, transport, and partner data, then validate the model before using it for operational decisions.
Inspect the supply-network example02 / City operations
Coordinate around one incident.
A major incident can affect transit, utilities, and hospital access at once. Your team could build a shared incident model that lets participating agencies compare response plans and see which decisions need approval.
In the public city-operations scenario, policy blocks plans that lose hospital access, warns about long transit detours, and requires approval for large resident impacts.
This uses scenario data. Agency connections, data-sharing agreements, and public-safety controls would need to be developed and validated with the participating organizations.
Inspect the city-operations example03 / Multi-agent AI
Keep delegated work within agreed bounds.
Several agents can pursue different objectives while acting on the same work. Your team could build a review service that records their proposals, compares alternative plans, and checks delegated authority before actions proceed.
The public multi-agent scenario blocks unreviewed expansion of write access, flags conflicting objectives, and sends larger delegations for approval.
Your team chooses and connects the models, agent frameworks, and tools. The example demonstrates coordination and review; it does not establish that an agent’s answer is correct.
Inspect the multi-agent example04 / Start with one decision
These are implemented examples in the repo. Start with the example closest to your problem, model your own domain, and test its decisions against familiar situations.
Build through documented interfaces such as the Public API, Python SDK, and extension contracts.
Existing systems retain their authority over business records. Your team defines the connections, policies, and people responsible for action.
Plan your first implementationLET’S TALK
What could work better?
A stubborn problem. A promising idea. Start wherever you are.