Lending · Pricing scenarios
DashboardLendingLoan pricing scenarios your pricing committee can actually read
Pricing decisions often rest on a spreadsheet only one analyst fully understands. Here is a scenario model built on a dashboard, where every assumption is visible and every scenario sits next to the others.
Valnora & Elric team5 min read
Sensitivity of net margin to each input
| Input | Low case | High case | Margin swing |
|---|---|---|---|
| Funding cost | −50 bp | +50 bp | ±0.37 pts |
| Expected loss rate | 0.6% | 1.2% | ±0.29 pts |
| Cost to serve | 0.45% | 0.65% | ±0.10 pts |
| Early repayment rate | 8% | 14% | ±0.06 pts |
3 days → 2 hrs
Time to prepare a pricing pack
Scenarios are saved on the dashboard, not rebuilt each cycle.
1
Definition of net margin
Every panel and alert reads the same governed metric.
4 → 9
Scenarios reviewed per committee
Side-by-side compare makes more options practical.
Illustrative example. Figures show how the workflow is measured; they are not customer results.
Loan pricing is a set of trade-offs: funding cost against market rate, expected loss against volume, cost to serve against product design. Most lenders model those trade-offs in a spreadsheet that grows a tab per question. By the time a pricing committee meets, nobody is sure which assumptions sit behind which number.
Agent Serve does not include a loan pricing product. What it does include is a scenario workbench inside Dashboards, and that is enough to build a pricing scenario model the committee can read, question and reuse. The scenario, input and form panels are not in the dashboard's Add panel menu yet, so today your Agent Serve team sets them up with you; once they are on the dashboard, analysts work with them directly. This post walks through one, panel by panel. All figures shown are illustrative.
Start from definitions, not formulas
The model is only as trustworthy as the definitions under it. Before building panels, set up three things.
- The data. Connect the warehouse or database that holds the loan book, rate card, funding cost curve and loss assumptions through Data sources.
- The semantic model. Describe loans, products, segments and risk grades once in the semantic model, including how they join. Panels then ask for "net margin by risk grade" rather than repeating joins.
- Governed metrics. Define net margin, expected loss rate and risk-adjusted return as named metrics. Weighted metrics handle balance-weighted rates, so a large loan counts for more than a small one.
When the committee asks how a number was worked out, the answer is one metric definition, not a chain of cell references.
The input and form panels
Input panels hold the assumptions a scenario changes: funding cost, expected loss rate by grade, cost to serve, early repayment rate and the customer rate itself. Each one is a visible, labelled control rather than a cell somewhere on a hidden tab.
A form panel suits single-deal questions. A relationship manager or pricing analyst enters the amount, term, grade and proposed rate for one deal and sees its margin against the floor straight away.
The scenario panel and side-by-side compare
The scenario panel saves a set of input values under a name, such as "base case", "funding +50 bp" or "grade 1 to 2 reprice". Compare shows saved scenarios side by side, so the committee sees the base case and each alternative in one view instead of flicking between tabs.
The KPI tiles at the top of the dashboard follow the selected scenario, and each one shows the gap to the base case. In the example, a 50 basis point rise in funding cost takes net margin from 2.31% to 1.94%, just above a 1.90% floor.
Sensitivity tornado and goal seek
Two analysis tools answer the questions committees ask most.
- Sensitivity tornado. It moves each input between a low and a high value and ranks inputs by how far they move the result. In the example, funding cost and expected loss explain most of the swing in margin. That tells the committee where to spend its debate time.
- Goal seek. Pick a target, such as a 12% risk-adjusted return, and the input to solve for, such as the customer rate. The workbench finds the value that hits the target, here 7.62%.
Neither tool makes the pricing decision. They make the trade-offs explicit, so the committee can decide and record why.
The rule and monitor behind the margin floor
Scenarios look forward. A rule and a monitor watch what actually gets booked.
- Below-floor rule. Flags any newly booked loan whose net margin is under the floor in your pricing policy. Flagged rows can be highlighted in the book table on the dashboard.
- Rule monitor. Fires when the rule matches more than a set number of loans in a period, and alerts the pricing lead in the app, by email or through a webhook to Slack or Teams.
A rising count of below-floor deals is often the first sign that the rate card and the market have drifted apart.
What the solution produces
Committee note from the scenario panel (illustrative)
MemoQuestion: should the SME term loan rate card move if funding costs rise 50 basis points?
Base case net margin is 2.31%. With funding up 50 bp and no rate change it falls to 1.94%, just above the 1.90% floor.
Grades 1 and 2 fall below the floor in the stress case. Higher grades stay above it.
Goal seek: a customer rate of 7.62% restores the 12% target return on the blended book.
Funding cost and expected loss explain most of the swing. Cost to serve matters less.
For decision: hold, reprice grades 1 and 2 only, or reprice the full card.
Sharing with the committee
Schedule a monthly report of the dashboard as a PDF to committee members, or share a password-protected link that expires after the meeting. Row-level access policies can limit regional teams to their own books while the central pricing team sees everything. Annotations let the pricing lead mark the week a rate change took effect, so later readers can see cause next to effect.
How to measure it
3 days → 2 hrs
Time to prepare a pricing pack
Scenarios are saved on the dashboard, not rebuilt each cycle.
1
Definition of net margin
Every panel and alert reads the same governed metric.
4 → 9
Scenarios reviewed per committee
Side-by-side compare makes more options practical.
Illustrative example. Figures show how the workflow is measured; they are not customer results.
Controls and governance
Row-level access
Row-level policies narrow what each person sees in tables, dashboards, exports and scheduled reports.
Roles and permission groups
Admin, write and read roles per workspace, plus permission groups that limit which models, blocks and integrations a team can use.
Audit trail
Proposals, approvals, declines and executed actions are written to the workspace audit log, which admins can export.
SSO and SCIM
Sign in through your identity provider with SAML 2.0 or OIDC, and provision people with SCIM.
Self-hosting
Run Agent Serve in your own cloud or data centre, next to the systems it reads.
Bring your own keys
Use your organisation’s own model provider keys at organisation or workspace level.
More on how Agent Serve handles access, data and audit on the security page.
Pricing frameworks are also a supervisory topic. In the EU, the EBA guidelines on loan origination and monitoring set expectations for how pricing reflects risk and cost. In the UK, Consumer Duty asks firms to show fair value. A model whose assumptions are visible makes that evidence easier to produce. For the credit side of the same deals, see credit memo drafting.
From pricing inputs to a committee pack
Data · Pricing inputs
- 01Warehouse tablesBook, rate card, funding cost and loss assumption tables, read from Snowflake, BigQuery or a database.
- 02Semantic modelDefines loans, products and segments once, with the joins between them.
- 03Governed metricsNet margin, expected loss rate and risk-adjusted return, defined once and reused.
Dashboard · Scenario workbench
- 04Input and form panelsAssumptions such as funding cost, loss rate and cost to serve, plus a form for one proposed deal.
- 05Scenario panelSaves named scenarios and shows them side by side against the base case.
- 06Tornado and goal seekShows which input moves margin most, and the rate needed to hit a target.
- 07Margin floor monitorAlerts the pricing lead when booked deals fall below the floor in policy.
- 08Pricing committeeReviews the scenarios and sets rates. The dashboard informs the decision.
What you need
- Book, rate card, funding cost and loss assumption data in a warehouse or database you can connect
- Agreed definitions for net margin, expected loss and risk-adjusted return
- Your pricing policy, including any margin floor and target return
- A short list of scenarios the committee asks for every cycle
- A pricing lead who owns the model and its assumptions
- Row-level access rules if regional teams should only see their own books