Procurement fails quietly. Nothing errors when a requisition sits for nine days, when an order slips past its promised date, or when a supplier's price drifts up twenty per cent across four orders. Each of those is a thing that did not happen, and things that do not happen do not appear on a dashboard unless somebody deliberately looks for their absence.
Concentration risk
One supplier carries 41% of spend across three categories.
This is what makes procurement different from the other three verticals. Accounting and sales mostly measure things that happened. Procurement mostly measures things that were supposed to happen and did not, which requires knowing what was expected and when.
An RFQ that has sat without movement for longer than a number of days you set — seven by default. Nothing is broken about a stale request, which is exactly why it stays stale: no queue shows it, no owner is blocked, and the person who raised it assumes procurement has it in hand.
Orders past the date they were promised for. Read alongside the on-time rate, this separates "this supplier is late" from "we are late with everything", which are different conversations with different people.
A unit cost that has moved beyond a multiple of what you have been paying — 1.2× by default, so a twenty per cent rise is the trigger. Price creep across several orders is the classic invisible loss: each individual increase is arguable, and the cumulative one is not.
How much of your spend depends on how few suppliers. A high concentration is not automatically wrong — it buys leverage and simplicity — but it should be a decision somebody made rather than a position you drifted into over two years of convenient reordering.
How dependably each supplier delivers when they said they would, built from your own receiving records rather than from their claims. This is the number that makes a difficult supplier conversation short, because it is your data about your orders.
How long orders really take, as opposed to the lead time somebody typed into a supplier record once. The gap between the two is what quietly breaks inventory planning, since the stockout calculation next door is using a number that may never have been true.
We could have left that in the settings screen for you to find. It belongs here, because the failure mode is specific and expensive: a buyer sees “analysed” next to a set of quotations, reasonably assumes the analysis considered delivery, and awards to a supplier the analysis never assessed on delivery at all.
Carries the whole comparison as shipped. It is the one factor that is unambiguous, comparable across suppliers and present on every quotation, which is why it is the safe default rather than the right answer.
Switched off until you turn it on. For most organisations this is the weight worth setting first — a cheaper quotation that arrives three weeks later is not cheaper, it is a stockout with a discount attached.
Also off. It matters intensely for bulky or perishable goods and not at all for a software licence, so a default either way would be wrong for half our customers.
Why not ship sensible non-zero defaults? Because a weighting is a statement about what your organisation values, and a weighting you did not choose is one you cannot defend in a procurement review. Zero is visibly a decision waiting to be made. A quietly reasonable 0.3 looks like an answer and is a guess about a business we have never seen.
An insights card tells you something is true. It does not stop anything happening, and it is worth being clear about which parts of procurement do stop things, because a reader who conflates the two ends up believing they have a control they do not have.
Payment is genuinely gated: a quantity discrepancy between what was ordered and what was received blocks release across every payment route, and paying past it needs its own permission and a written reason. Approving a purchase order is gated by permission — whether a person may approve at all.
Those are described on Accounting controls and Governance & roles.
Surfaces the things nothing refuses. A request going stale breaks no rule. A supplier creeping up on price breaks no rule. Drifting into depending on one supplier for two-fifths of your spend breaks no rule. These are exactly the problems a control system cannot catch, which is the argument for measuring them.
Each finding carries a link to where the work is done, so the gap between noticing and acting is one click rather than one meeting.
This is the single most important thing to understand about Procurement Insights, and it is the thing most “AI analytics” pages are careful not to say. Every figure on the page — every count, every total, every flagged record — is computed by an ordinary database query against your own data, with thresholds you control. A language model is used for exactly one job: turning that computed result into a paragraph a person can read quickly.
Counts, totals, ages and comparisons, run against your workspace under the same tenant scoping as every other screen. This step is deterministic: run it twice on unchanged data and you get the same answer twice. If the engine says nine invoices are overdue, nine rows matched.
What makes a request stale, a cost a spike, an entry large or a balance overdue is a number in your settings, not a judgement in a model. The defaults are listed below. Change one and the flags change with it, immediately and predictably.
Only then is a short plain-language summary generated from those facts. It is written on a background job so the page never waits for it, and it fails soft: if no provider is configured or the request fails, the summary is simply not there and every number on the page is unaffected.
The summary is stored with a fingerprint of the exact figures it was generated from. When those figures move, the fingerprint no longer matches and the summary is regenerated rather than served stale. This closes the failure that makes cached AI commentary dangerous: a confident paragraph describing last week, sitting directly above this week's numbers, with nothing on the screen indicating the two disagree.
Results are cached per workspace for a window you set — fifteen minutes by default — so the page is fast without being out of date, and a refresh action clears it when you want the current picture immediately.
These are the actual settings behind the flags on this page. They are per workspace, they are yours to change, and the defaults are chosen to be defensible rather than dramatic — an engine tuned to flag everything on day one is an engine nobody reads by day three.
| Setting | Ships as | What it decides |
|---|---|---|
| procurement_stale_rfq_days | 7 | Days without movement before a request is flagged as stale. Shorten it where sourcing is fast; lengthen it for tenders that legitimately sit. |
| procurement_cost_spike_multiplier | 1.2 | How far a unit cost has to move above what you have been paying before it is flagged — a twenty per cent rise, as shipped. |
| procurement_quote_auto_analyze | true | Whether quotations are analysed as they arrive rather than on request. |
| procurement_quote_price_weight | 1.0 | How much price counts in that analysis. It carries the entire comparison until you change the other two. |
| procurement_quote_delivery_weight | 0.0 | How much promised delivery counts. Zero as shipped — usually the first weight worth setting. |
| procurement_quote_distance_weight | 0.0 | How much supplier distance counts. Zero as shipped, because it is decisive for bulky goods and irrelevant for others. |
Every flag carries the way to act on it. A finding is not a number on a dashboard you then go hunting for — it is rendered with a label and a link to the screen where the work is done. An insight you cannot act on from where you are reading it is a report, and you already have reports.
What AWRA OpsHub does today
More we can add to your workspace
Where we point you to a specialist
Value-based approval routing is the most asked for row here, and the engine already stores value thresholds — the routing that consumes them is the piece we would build. If that is your control requirement, it is a well-scoped piece of work.
Cash flow split three ways, entries that are large for you, and anomalies.
Revenue momentum, overdue money with no hidden grace period, days to pay.
Stockout risk on a horizon you set, and an engine that declines to guess.
All four modules on one card, and the workspace-level narrative across them.
The module underneath the stale-request and quote-analysis signals.
The reliability figures in detail, and what suppliers see of them.
Help Center
Run inventory, procurement, assets, sales, and field work with approved AWRA guidance for setup, migration, integrations, security, pricing, and support.
