Root Cause Analysis Agent – Root cause analysis for business processes: find out why cases are late, costly or faulty
Answers the question of why: traces dependencies in the causal process graph and explains problems with evidence.
What does the Root Cause Analysis Agent do?
The Root Cause Analysis Agent answers the question that follows every KPI: why? It takes a problem – late deliveries, high rework, long approvals – and looks for the factors that explain it. To do so it uses Noreja’s causal process graph, which represents the dependencies between activities, compares affected and unaffected cases by their attributes and draws on company knowledge from the Context Hub. The result is a traceable explanation: which factors are behind the problem, how strongly they act and what evidence supports them.
Questions the Root Cause Analysis Agent answers
“Why are so many orders in the order-to-cash process delivered late?”
“What distinguishes the slowest ten percent of cases from the rest?”
“Explain the selected case: why did it take so long?”
“What do invoices that go into clarification have in common?”
How the Root Cause Analysis Agent works
- 1
Specify the problem
The agent defines which cases count as affected – late, reworked or above a cost threshold – and asks if the question leaves this open.
- 2
Trace causal paths
In the process graph it traces which upstream steps and dependencies lead to the problem.
- 3
Compare influencing factors
It compares affected and unaffected cases by their attributes – supplier, product, region, team, document type – and weights the differences by frequency and impact.
- 4
Explain with context
What the data does not show – strikes, system migrations, new regulations – it adds from the Context Hub and cites the document it relies on.
KPIs and methods
- Share of affected cases
- How many cases show the problem – overall and per attribute.
- Deviation per attribute
- How much the problem rate for an attribute differs from the average.
- Contribution to total impact
- Which part of the added duration or cost is due to a factor.
- Upstream events
- Steps and deviations that precede affected cases in the process flow.
A practical example
In order-to-cash, a growing share of orders is delivered late. The agent compares late and on-time orders and finds two factors: subsequent changes to the delivery date and a credit check that costs new customers extra days. From the Context Hub it adds that the credit policy was tightened a few months ago – and so explains why the problem has occurred since exactly that point.
Matching end-to-end processes
What you get
A reasoned explanation of the problem with the main influencing factors, their weight and the evidence for them – affected cases, paths in the process graph and the documents used. As a report or presentation on request.
What data the agent needs
A perspective with as many attributes per case as possible. The more company knowledge there is in the Context Hub – policies, organizational changes, known issues – the better the agent can explain the why.
Root Cause Analysis Agent: frequently asked questions
What is root cause analysis in process management?
Root cause analysis looks not at the symptom but at the reason for a process problem. Instead of stating that deliveries are late, it asks which steps, conditions and decisions lead to that. Classic methods are the 5 Whys or the Ishikawa diagram; data-driven, it is done by comparing affected and unaffected cases.
Why does a causal process graph help with root cause analysis?
A frequency-based process model shows which steps often follow one another. A causal process graph shows which steps depend on one another. For root cause analysis that is the decisive difference: the agent can trace dependencies back instead of merely comparing time sequences.
Can the agent explain individual cases?
Yes. If you select a case in the Analyzer, the agent explains why that particular case took so long or deviates from the expected flow.
Part of Minerva 2.0
The Root Cause Analysis Agent is an agent in Minerva, the AI assistant in Noreja. It uses the same data sources, the same choice of model – local or cloud – and the same access rights: it only reads what you are allowed to see and changes nothing in Noreja.
More agents from the library
Structural Deviation Analysis Agent
Finds skipped steps, swapped sequences and paths that should not exist according to the target process – and shows what they cost.
View agentRework Analysis Agent
Calculates the rework rate, finds loops and repeated steps and connects them with their causes.
View agentOn-Time Performance Agent
Measures on-time delivery, OTIF and SLA compliance and shows at which step commitments are missed.
View agentSee what the Root Cause Analysis Agent finds in your process data
In a demo we show you what answers the agent delivers on one of your processes – with numbers, reasoning and sources.
Book a demo