Planning cycle

Underwriting sends plan assumptions, then revisions, then “one more scenario”. Each request queues behind the next model run, and each answer lands in another spreadsheet. With STEER, every assumption is versioned with an owner. Indicative answers keep the discussion moving, and governed runs confirm what actually goes into the plan.

SCR submission

Run records, reconciliations, review notes and approvals are captured as the work happens. The evidence behind an SCR submission (or, for a Lloyd’s syndicate, the capital return) is assembled as you go, not reconstructed in the final fortnight.

Reconciliation

“Why is this run different from last quarter’s?” Every run records exactly what changed: inputs, assumptions and model version. Compare any two identified runs by class, risk type and risk driver, and see where the capital moved without rebuilding the analysis by hand.

Explaining the number

The movement analysis lives next to the figure it explains. When Cyber growth adds less capital than expected, the explanation (extra stop-loss recoveries, a shift in diversification) is attached to the result, ready for the committee.

What Lloyd's asks for

How STEER helps

Pre-application form by end of January

Test the planned change early and see its likely SCRu impact, so you know whether it is major before the deadline.

A hypothetical LCR with every application, and a resubmission if uSCR moves more than 10%

Run the hypothetical next to the approved basis. The movement is visible before you commit to a resubmission.

AoCs on 1Yr and ultimate SCR, with risk-specific impacts isolated

Reconcile any two runs. The bridge between them is broken down by risk type, so catastrophe or market risk changes can be shown on their own.

Independent validation of the change and its output

Inputs and results are versioned and repeatable, including multiple seed runs, so the second line can re-run the change rather than take it on trust.

Minor changes that add up to a major one

Each change is logged with its impact, so the running total is always in view and does not surface at year end.

Board approval, evidenced through papers and minutes

Board material comes from the same runs, and the challenge raised is recorded against the figure it was about.

Workflow walkthrough

Plan assumptions in → explore indicatively → request a governed run → reconcile against the last run → review and sign off → submit with the trail attached.

Cycle time

The time from plan assumptions received to a signed-off capital result.

Scenarios tested

The number of plan alternatives evaluated per cycle, indicative and governed.

Audit findings

Validation and audit points raised on evidence, documentation or traceability.

Traditional capital MI

• New scenarios queue for the next model cycle
• Results re-keyed by hand into spreadsheets and slides
• Movements explained weeks after the fact
• Documentation rebuilt for validation and audit
• Approvals recorded in an email, if at all

• Indicative answers in the meeting, governed runs on request
• Results flow straight from the model, untouched
• Any two runs compared and explained by driver
• Evidence captured as you decide
• Approvals recorded against the exact supporting run

Appetite setting

Capital is only useful once it becomes limits the business can act on: premium caps, aggregate limits and line sizes. STEER ties each limit to the capital view behind it, so a change in appetite and a change in capital can be traced to each other.

Portfolio mix

Which lines earn their capital? STEER shows return on allocated capital alongside marginal capital, the extra capital the next pound of premium would need. Because of diversification, a line can look expensive on its own and efficient within the portfolio, or the reverse. STEER shows both views side by side, from the same governed run.

In-year steering

When the book drifts from plan through rate movements, mix shifts or a large loss, test the response straight away with an indicative view. Then confirm it with a governed run before committing to it.

Board narrative

One story, consistent with the capital team’s numbers, because both teams read the same record. The board sees one version of the plan, not two competing decks.

Portfolio view

Expected profit against marginal capital by class, with appetite headroom marked.

Decision latency

The time from a business question being asked to a recorded, evidenced decision.

Appetite breaches

Limits exceeded before anyone saw them coming.

Plan variance

The gap between planned and actual mix, and the capital it consumes.

Steering today

• Capital view arrives once per cycle, after the plan is set
• Underwriting and capital each hold their own version of the numbers
• Mix decisions made on loss ratio alone
• In-year changes wait for the next model run

• Capital impact tested while the plan is still being shaped
• One governed record shared by both teams
• Mix judged on return and marginal capital together
• In-year responses explored instantly, confirmed by a governed run