Hub for Portfolio Risk Trends

Portfolio managers see performance in aggregate and risk in aggregate, which is precisely the resolution at which a concentration is invisible. The segment driving the loss ratio is two levels below the line anyone reports on, and the capital allocated against it was set on last year's mix. Meanwhile the macro signal that should be triggering action — a sector story, a regulatory shift, a pattern three underwriters have each mentioned in a meeting — never reaches the portfolio view at all, because it is not the kind of thing that has a field in the system.

• Performance is reported at a level of aggregation that hides the concentration causing it.
• Capital allocation is compared against a portfolio shape that has already moved.
• External signal — news, industry reports, meeting notes — has no route into the risk view.
• By the time a macro event is reflected in the numbers, the window to reprice or re-cede has passed.

Hub gives portfolio management two views that usually live apart. The granular view analyses performance at every level of the portfolio and against capital allocation, so risk concentration and policy volume can be located rather than inferred — the segment, the peril, the geography, the cohort. The macro view enriches the same ontology with external public sources: trend information drawn from news articles and industry reports, and from internal material such as meeting notes that would otherwise stay unstructured. Because both views resolve against one model of the book, a macro observation can be traced directly to the exposure it affects, and a portfolio action can be triggered from it while there is still time to take one.

The external half of this is only tractable with a model. News articles, industry reports and meeting notes are prose written for humans, and the task — decide whether this development bears on our exposure, and where — is a judgment about relevance rather than a keyword match. Structuring that stream into something the portfolio ontology can hold is what makes macro context actionable instead of anecdotal. The granular half is conventional analytics; it simply had to be joined to the same model for either half to be useful.

GRANULAR VIEW

Performance at every portfolio level, against capital allocation

MACRO VIEW

External news, industry reports and internal notes, structured

ONTOLOGY

One model of the book behind both views

OUTPUT

Located risk and policy-volume concentrations, with trigger points

AUTONOMY

Surfaces and traces; the portfolio action stays with the manager

1. 01Performance analysed at every level of the portfolio, not only at the reported aggregate.
2. 02Risk and policy volume assessed directly against allocated capital.
3. 03External public data — news, industry reports — enriched into the risk ontology.
4. 04Internal meeting notes and other unstructured material brought into the same view.
5. 05Macro developments traced to the specific exposure they affect.
6. 06Portfolio actions triggered from contextualised events rather than from lagging numbers.

Every level

Portfolio resolution, against the reported aggregatedesign intent

Traceable

Macro signal to affected exposuredesign intent

• AGENTIC AIOpsResolve™→
• UNDERWRITINGHub Property Insurance Underwriting Assistant→
• RESTAURANT OPERATING SYSTEMPrime POS→