Sample output from the Explain the ERP to the Board: One-Pager prompt in the Sonar AI Prompt Library, run against a NetSuite test account. Every name and number here is test data. Back to the post · The library
Board Briefing  ·  Information Systems  ·  August 2026

The ERP, Explained

A plain-language account of what our NetSuite system does on its own, what still requires people, where the operational risk sits, and what the system costs in management attention.


4,364
Transactions processed, trailing 12 months (~17 per business day)
+55%
Volume growth, Apr–Jul 2026 vs the prior eight months — absorbed with no added headcount or systems
98.5%
Of transaction volume flows through structured, automated cycles — not manual journal entries
~1,170
Active automation programs, the large majority vendor-built and vendor-maintained

1The system is absorbing real growth

483 280
Aug 2025Nov 2025Feb 2026Jul 2026
483 / mo Monthly transactions reached a twelve-month high in July — up 72% from last August — while the count of manual journal entries stayed flat. Throughput is scaling; hand-work is not.

2What runs itself — two closed loops

The two commercial engines of the business operate as complete, closed loops inside the system. Each box below is a step the system links automatically; the counts are actual documents processed in the last twelve months. Nothing re-keyed, nothing carried between systems by hand.

Order-to-cash
(sell & collect)
Sales order
438
Shipment
403
Invoice / register sale
411 + 553
Cash collected
398 receipts
Procure-to-pay
(buy & settle)
Purchase order
424
Goods received
411
Vendor bill
558
Payment issued
544
The near-matching counts along each chain are the point: orders placed become goods moved, bills booked, and cash settled at almost one-to-one ratios. Manual journal entries — the accounting equivalent of hand-typing — numbered just 64 all year (1.5% of volume). Fixed-asset depreciation, tax calculation, electronic bank payments, and warehouse shipping run as installed vendor modules that post to the books automatically.

3Where the money actually comes from

Two very different businesses share the same automated rails. Wholesale/distribution produced $1.06M — 88% of trailing-twelve-month revenue — from just 17% of sales transactions; the retail stores produced the remaining 12% from 83% of transactions. The system's economics work in both directions: it handles the stores' high-frequency small tickets and the distribution centers' large orders with the same process and the same people.

Wholesale / distribution — $1,056K (88%) · 161 transactions · avg order ≈ $6,600
Retail — $147K · 799 tickets · avg ≈ $184
ChannelRevenue (T12M)TransactionsAvg. size
Los Angeles Distribution Center$636,183114$5,580
Miami (wholesale)$420,20147$8,940
San Francisco Store$91,639463$198
New York Store$55,008336$164

4What still needs hands

  • Pre-sales. Only 22 opportunities and 10 quotes were recorded in the system all year against 438 sales orders. The pipeline exists — it lives outside the ERP, in inboxes and spreadsheets — so forecasting starts from anecdote, not data.
  • Deployment & configuration changes. New automation is installed and configured by hand, deliberately — a control point, but one that depends on specific people.
  • Judgment work. Month-end close decisions, credit terms, and exception handling remain human calls. The system prepares the data; it does not make these decisions — by design.

The positive reading: very little of the accounting is hand-typed, and the growth of the last four months required no new manual process.


5Where the operational risk sits

ExposurePlain-language descriptionPosture
Vendor dependence Roughly 95% of the system's automation and custom structure is vendor-supplied (asset management, bank payments, tax, shipping, retail suite). We benefit from their maintenance — and inherit their upgrade timing and defects. Accepted; monitored at upgrade windows
Customer concentration 88% of revenue flows through two wholesale channels and 161 transactions. This is a commercial exposure the system makes visible, not one it creates. Visible; commercial matter
Key-person concentration Administration and change deployment are concentrated in a small number of individuals. Continuity depends on documentation, which is actively maintained. Partially mitigated
Credential hygiene One in-house integration was found storing a supplier credential in plain text. Identified in internal review, August 2026; remediation flagged to management. Flagged; remediation open
Small in-house footprint Genuinely custom, in-house-built components are few. This limits both maintenance burden and the risk of undocumented logic. Strength, not exposure
Note on control Every change made through the system's administrative tooling is written to a permanent audit trail. Deletions in production require explicit, recorded human confirmation. The system is configured to make quiet changes impossible.

6What it costs in attention — and the one thing to fund

The system's attention cost is modest and concentrated where it should be: reviewing exceptions the system surfaces, timing vendor module upgrades, and maintaining documentation against key-person risk. Volume growth of 55% arrived without a corresponding rise in manual work — the strongest evidence the automation is real.

The funding ask: bring the sales pipeline into the system. It is the one commercial process the ERP cannot currently see, and — given that 88% of revenue rides on a small number of wholesale relationships — it is also the process where early visibility matters most. Everything downstream of a signed order is already automated; the gap is everything upstream of it.