Most dashboards are chart galleries. Every panel is technically accurate and none of them answers "so what?" The executive who opens one has to do the analysis themselves, and usually doesn't.

This week's release of the Sonar AI Prompt Library added a group of prompts aimed at the other kind of view: the one that briefs. Where the cash is and where it's going. What's concentrated enough to hurt. Which triggers have gone amber. What the board should be asked to decide. There are seven of them, and I want to walk through what each one does and the rule that makes it trustworthy.

If you're new to this, Sonar AI is an AI agent that runs inside NetSuite. Every prompt in the library is a playbook that I engineered and tested against live NetSuite data, and you run it inside your own account, against your own records.

Cash Position Pulse Dashboard

This is a 13-week cash forecast rendered the way a hurricane track is rendered. The base case is a bold line. The conservative, stress, and best cases are dashed tracks, and the envelope between them is shaded. Around it are weekly flow bars, a scenario table, liquidity metrics, and a weekly detail table with a confidence label on every week: high for the first four, medium for the next four, medium-low for the rest, with a note on what would upgrade it.

The rules are the point. Starting cash must come from the live bank-balance query, never estimated and never stale. Collection rates are assumptions, not system data, and the report says so. The forecast is computed in code, not mental arithmetic, because four scenarios across thirteen compounding weeks is exactly where hand math breaks. And if no scenario breaches the minimum threshold, the prompt says that explicitly and then redirects the summary to the real cash story, which is usually aged receivables or one concentrated outflow week. A clean forecast still owes the reader an insight.

Executive Command Center

A single-screen, dark-mode dashboard of the kind you'd put on an office TV: cash position, receivables aging against the payables schedule, revenue trend, a cash-coverage gauge, the pipeline funnel, top metros, and top customers. It's headlined by an anomaly ticker that scrolls the findings an executive should act on.

The ticker is the product. It has to carry specific, quantified findings, with names, dollars, and days, not "some receivables are aging." It grades severity and includes positives, because an all-red ticker reads as noise. And the numbers on the KPI strip have to reconcile to the panel-level detail beneath them.

Board Meeting Executive Summary

This is the pre-read for a board meeting, and brevity is the deliverable. A board pre-read that needs its own summary has failed. So the prompt caps itself: at most five KPIs, three risks, and four recommended decisions. Every recommended decision has to be specific and voteable, such as approving a collections program targeting a stated dollar amount of 90-plus-day receivables, never "improve collections."

It uses NetSuite's standard reports as the source for every financial statement, verifies that each report's period header matches what was requested, and treats two benchmark runs with identical headers as a filter failure rather than a coincidence. It's the compressed, decision-oriented sibling of the Financial Performance Overview prompt. Where that report explains, this one decides.

Risk Exposure Heatmap

The ERP equivalent of a portfolio concentration heatmap. Instead of sectors and instruments, it measures concentration across customers, vendors, products, and geography, and it flags any single name above a threshold, ten percent by default, with a proposed mitigation.

Two of its rules matter more than the rest. Receivables concentration and revenue concentration are different risks, and it reports both, because a customer can be five percent of revenue and forty percent of receivables. And vendor risk includes single-source dependency, meaning vendors supplying items no other vendor has supplied in twelve months, which is exposure even when the spend share is small. It also checks whether address coverage is good enough before publishing a geography axis at all, and reports "insufficient data" when it isn't.

Threshold Alert Scorecard

This is the honest ERP version of real-time alerting. A one-shot agent can't push alerts, but a scorecard with stable trigger definitions, re-run weekly or monthly, delivers the same alarm logic with an audit trail. The panel covers the cash floor, receivables concentration, days sales outstanding, inventory days, the approval backlog, and overdue payables.

Every trigger has exactly the same shape: metric, thresholds, current value, a green, amber, or red state, the 30-day direction, and a pre-agreed response protocol. A red state without a protocol is an alarm without a fire drill. The definitions are versioned, and any change between runs is listed, because the value of the scorecard is comparability from one run to the next.

Depreciation Expense Forecast

For accounts running the Fixed Assets Management SuiteApp, this prompt projects depreciation expense forward, month by month, from each asset's actual cost, method, life, convention, and current book value. It reconciles the projection against what has already been posted per asset before it projects anything, because a model that can't reproduce the past has no business predicting the future.

It forecasts only what the configuration supports. Straight-line methods project exactly. Declining-balance and MACRS variants project with stated formulas. An asset on a method it can't faithfully model is excluded, listed, and quantified rather than silently straight-lined. And it's explicit that this is a projection of the current register, with no future acquisitions or disposals, because that's the number one way readers misuse the result.

Money Flow Sankey

This one is coming in the next release rather than this one, and it'll be free. It renders the whole P&L as a river: revenue streams enter on the left, split into cost of sales and gross margin, and gross margin splits into the operating expense branches and the net income that survives. One glance answers "where does the money go?" better than any statement layout.

The rule is that the river has to reconcile. Every node's inflow equals its outflow, and the net income node on the right has to equal the income statement's net income for the same period, to the dollar, with the tie-out displayed in the report. A pretty diagram that doesn't tie is a liability. The account groupings come from the real chart of accounts, never invented categories.

Wrapping Up

What ties these together is a refusal to be decorative. Each one has a rule that forces it to say something: the redirect to the real cash story, the ticker that has to be specific, the decisions that have to be voteable, the protocol behind every red state. I've found that the rule matters more than the visual, because the visual is what gets the report opened and the rule is what gets it acted on.

Six of these are in the paid tier of the Sonar AI Prompt Library now. Look for Cash Position Pulse Dashboard Generator, Executive Command Center Generator, Board Meeting Executive Summary, Risk Exposure Heatmap, Threshold Alert Scorecard, and Depreciation Expense Forecast (Fixed Assets Management). Money Flow Sankey follows in the next release.