Organizational Intelligence Assessment

How ready is your organization for AI agents?

Explore Governance, Planning, Memory, Routing, Execution, Observation, Learning, and Simulation. Identify strengths and next steps through a directional, self-reported profile—not a validated measurement of your entire organization.

Answer 26 questions across eight capabilities in about five to seven minutes. No email is needed to view or download your results. Anonymous responses are submitted to help improve the assessment.

Progress cannot be retained on refresh in this browser. You can still complete the assessment and download your results.

Your progress0 of 26 answered

Capability 1 of 8

0 / 3 questions

Governance

Authority, policies, human oversight, and accountability for AI-assisted work.

Our organization has a documented AI policy with human-review checkpoints.

No policy or guidelines existInitialInformal guidelines, applied inconsistentlyDevelopingDocumented policy exists but is rarely enforcedDefinedPolicy enforced with human-review checkpointsManagedPolicy enforced, audited and continuously improvedOptimized

We can monitor, pause, or halt automated work before it ships.

No visibility into automated actionsInitialWe can review some outputs after the factDevelopingManual checks on critical actions onlyDefinedReal-time monitoring with the ability to pause workManagedAutomated guardrails halt risky work and alert ownersOptimized

Roles and accountability for AI decisions are clearly defined.

No one owns AI outcomesInitialOwnership is ad hocDevelopingSome teams have defined responsibilitiesDefinedClear owners and escalation paths existManagedAccountability is documented, audited and reviewedOptimized

Capability 2 of 8

0 / 2 questions

Planning

Turning goals into prioritized, coordinated steps.

Complex goals are broken into clear steps and routed to the right owners.

Work is handled ad hoc, without a planInitialPlans exist but steps are unclearDevelopingWe document plans for major initiativesDefinedGoals are decomposed and routed systematicallyManagedDecomposition and routing are automated and optimizedOptimized

Work is prioritized and assigned based on capacity and importance.

First-come, first-served onlyInitialPriorities set informallyDevelopingPriorities documented per projectDefinedCapacity-aware prioritization in placeManagedDynamic prioritization adjusts in real timeOptimized

Capability 3 of 8

0 / 3 questions

Memory

Shared, reliable context, decisions, and knowledge that can be reused.

Decisions, context, and lessons are captured for reuse.

Knowledge lives only in people's headsInitialSome notes exist but are scatteredDevelopingWe document key decisions in a shared placeDefinedContext and lessons are captured systematicallyManagedMemory is structured, searchable and reused automaticallyOptimized

Past work informs new work so we rarely start from scratch.

Every effort starts from zeroInitialWe occasionally reuse old materialDevelopingTemplates and references existDefinedSystems surface relevant past workManagedLearning compounds loop over loopOptimized

There is a reliable source of truth for context and data.

Data is fragmented across toolsInitialA few shared docs, often outdatedDevelopingA central store exists but is incompleteDefinedA maintained source of truth for most contextManagedUnified, current and trusted across the orgOptimized

Capability 4 of 8

0 / 3 questions

Routing

Directing work and handoffs to the right people, systems, or tools.

We maintain a registry of who (or what) can perform which tasks.

No registry of capabilities or ownersInitialTribal knowledge onlyDevelopingA partial list exists somewhereDefinedA maintained registry maps tasks to ownersManagedRegistry drives automated routing and load balancingOptimized

Hand-offs between people and systems are coordinated, not chaotic.

Hand-offs are manual and error-proneInitialHand-offs work but rely on individualsDevelopingDefined hand-off processes for key flowsDefinedA control plane coordinates most hand-offsManagedCoordination is centralized, observable and adaptiveOptimized

Deliverables reach the right people or systems without friction.

Hand-offs are manual and get lostInitialHand-offs work but are slowDevelopingDefined delivery channels existDefinedDelivery is reliable and traceableManagedDelivery is automated with confirmation and trackingOptimized

Capability 5 of 8

0 / 6 questions

Execution

Reliable tools, suitable roles, scalable work, and usable, checked outputs.

Work is executed with reliable tools and integrations, not manual effort.

Mostly manual, repetitive workInitialSome tools, poorly integratedDevelopingKey tasks are tool-assistedDefinedIntegrated tools handle most executionManagedExecution is automated end to end with fallbacksOptimized

Tasks are handled by specialized roles or systems suited to them.

Generalists handle everythingInitialLimited specializationDevelopingDefined roles for core functionsDefinedSpecialized roles or agents for most workManagedSpecialized, composable units orchestrated togetherOptimized

Automated work runs reliably and handles errors gracefully.

Frequent breakages, no recoveryInitialFailures require manual rescueDevelopingBasic error handling in placeDefinedRetries and fallbacks for key flowsManagedSelf-healing execution with monitoringOptimized

We can scale execution without proportional human effort.

Scaling means hiring more peopleInitialScaling is slow and costlyDevelopingSome processes scale with automationDefinedMost execution scales independently of headcountManagedExecution scales elastically on demandOptimized

Finished work is packaged into clear, usable deliverables.

Outputs are raw and inconsistentInitialFormatting depends on the individualDevelopingStandard formats for key deliverablesDefinedConsistent, polished deliverables by defaultManagedDeliverables auto-formatted to audience and channelOptimized

We verify deliverable quality before it goes out.

No review before deliveryInitialAd hoc spot checksDevelopingReview for important deliverablesDefinedConsistent QA on most outputsManagedAutomated quality gates plus human sign-offOptimized

Capability 6 of 8

0 / 3 questions

Observation

Reviewable evidence of execution, failures, costs, and outcomes.

We track cost, failures, and runaway loops across our AI systems.

No tracking of cost or failuresInitialWe notice problems only when they escalateDevelopingWe review cost and errors periodicallyDefinedDashboards track cost, errors and loops in near real timeManagedBudgets and limits enforced automatically with alertingOptimized

We know whether deliverables actually landed and were useful.

No visibility after hand-offInitialWe hear only when something is wrongDevelopingWe follow up on major deliverablesDefinedWe track receipt and usefulnessManagedClosed-loop measurement informs the next cycleOptimized

How consistently can people review a record of what happened during agent-assisted work?

No reviewable record of the work existsInitialScattered notes or logs cover only some workDevelopingReviewable records exist for important work, but links or details are inconsistentDefinedPeople can consistently review linked records of decisions, tool use, errors, costs, and outcomesManagedLinked records are consistently reviewable, and their completeness is routinely checkedOptimized

Capability 7 of 8

0 / 2 questions

Learning

Using feedback and outcome checks to improve subsequent work.

We capture feedback and feed it back into how work is done.

No feedback loop existsInitialFeedback gathered occasionallyDevelopingRetrospectives for major projectsDefinedFeedback routinely improves processesManagedContinuous, automated learning loopsOptimized

How consistently does the organization check whether changes made from feedback improve subsequent outcomes?

We do not check whether feedback-driven changes helpInitialWe rely on anecdotes about whether changes workedDevelopingWe periodically review outcomes after important changesDefinedWe regularly compare outcomes before and after changesManagedWe systematically compare outcomes and retain or revise changes based on what we findOptimized

Capability 8 of 8

0 / 4 questions

Simulation

Testing options, scenarios, and risks before committing resources.

We model or test approaches before committing resources.

We act first and find out laterInitialOccasional informal estimatesDevelopingWe plan scenarios for big betsDefinedWe simulate options before major decisionsManagedSimulation routinely guides strategy selectionOptimized

We consider multiple scenarios and pick the strongest path.

Single plan, no alternativesInitialWe sometimes weigh a couple of optionsDevelopingStructured options analysis for key choicesDefinedMultiple scenarios compared systematicallyManagedContinuous scenario modeling at system scaleOptimized

We anticipate failure modes before they happen.

We react to problems as they ariseInitialWe know the obvious risksDevelopingRisk reviews for major initiativesDefinedProactive risk modeling in planningManagedPredictive modeling flags risks earlyOptimized

Decisions are informed by models and data, not just intuition.

Decisions are mostly gut feelInitialSome data used inconsistentlyDevelopingData informs major decisionsDefinedModels support most decisionsManagedDecisions are continuously optimized with dataOptimized