Let your CSRs focus on what matters: customer relationships.
Karini AI agents read every order and quote request, check each line against the item master and apply your credit SOP. Your CSRs spend their time with customers, not chasing parts and credit holds.
- Written for
- VP Sales Operations
- Also in the room
- Credit director and order desk lead
- Governed by
- Order Processing Exception Management SOPO2C-OPS
Releasing an order takes four screens at once.
Before release, every order needs two answers. Can we build it: does each line match a real, stocked, correctly priced part? Can we credit it: does it stay within policy?
Answering takes the item master, AR aging, the contract and the SOP, so people approximate. And nothing records which rule they used or who accepted the risk.
The agents, in the order they run.
Six agents, one job each. They match the account, check every line, calculate exposure (open AR + unshipped orders + this order), then post the order or stage it for approval.
A cross-sell agent runs alongside the decision, never inside it, with up to three suggestions. What the approver sees is exactly what the ERP receives.
Quote requests run through the same item master check, so every quoted line is a real, correctly priced part before a CSR sends it, and an accepted quote converts to an order without being keyed twice.
See it in action.
An order arrives as a document, and one ambiguous line lists its candidates. The escalated order shows the exposure math, the SOP rule, and the release amount that brings exposure back to 100% of the limit.
The reviewer approves a partial release. The rest stays open, and the ERP receives exactly what was approved.
The rules come from your SOP.
These rules come from a reference SOP. They show the shape, not a policy Karini AI imposes. On your ERP, the thresholds and approval levels are yours.
Credit rules
Exposure = open AR + unshipped orders + this order, as a % of the credit limit. The SOP sets the formula, so every case shows the same math.
| Scenario | When it applies | Decision | Approval | SOP section |
|---|---|---|---|---|
| Exposure at or below 100% of limit, with no adverse signal | AUTO_RELEASE | No approver required | Scenario Dispatch, Credit | |
| Exposure 100% to 110%, clean aging, days to pay stable or improving | AUTO_RELEASE | No approver required | Scenario Dispatch, Credit | |
| Exposure 110% to 150%, no aging deterioration | ESCALATE | Level 2 | Scenario Dispatch, Credit | |
| Any balance in the 61 to 90 day bucket, or days to pay worsened by more than 25% over twelve months | ESCALATE | Level 2 | Scenario Dispatch, Credit | |
| Exposure 150% to 200% of limit | ESCALATE | Level 3 | Scenario Dispatch, Credit | |
| Exposure above 200% of limit, or an order above $100,000 | ESCALATE | Level 4 | Scenario Dispatch, Credit | |
| No credit limit on file, or an account less than 90 days old | CREDIT_APPLICATION | No approver required | Scenario Dispatch, Credit | |
| Any balance 90 days or over, active collections, a returned payment in the last 90 days, or a manual hold flag | HOLD | Level 3 | Scenario Dispatch, Credit |
AUTO_RELEASEReleased and posted by the agent.ESCALATESent to the approver at the level the SOP sets.HOLDNothing moves until the blocking condition clears.CREDIT_APPLICATIONA separate workflow starts to set up credit for the account.Item master rules
Two failure types, kept apart. A line that couldn't be read goes to someone who checks the document. A line that read cleanly but doesn't match is a product decision.
| Scenario | When it applies | Decision | SOP section |
|---|---|---|---|
| Resolved against the item master at confidence 0.90 or above, with no substitution | ACCEPT | Scenario Dispatch, Item Master | |
| The referenced part is discontinued and a direct replacement exists | ACCEPT | Scenario Dispatch, Item Master | |
| Two or more candidates resolve above 0.75 | NEEDS_REVIEW | Scenario Dispatch, Item Master | |
| Extraction confidence below 0.60 | NEEDS_REVIEW | Scenario Dispatch, Item Master | |
| Discontinued, with no replacement on file | NEEDS_REVIEW | Scenario Dispatch, Item Master | |
| The quoted price differs from the contract price list by more than 5% | FLAG | Scenario Dispatch, Item Master | |
| The ship from branch has no stock | FLAG | Scenario Dispatch, Item Master |
Customer master rules
The first check, and the one that saves the most rework. A duplicate PO caught here is an order never built twice.
Note: here CM_ means customer master. On the Credit Management page it means credit management. Each row's SOP trace shows the source.
| Scenario | When it applies | Decision | SOP section |
|---|---|---|---|
| A single account resolves at confidence 0.92 or above | ACCEPT | Scenario Dispatch, Customer Master | |
| Two or more candidate accounts resolve above 0.70 | NEEDS_REVIEW | Scenario Dispatch, Customer Master | |
| No candidate account resolves above 0.70 | CREDIT_APPLICATION | Scenario Dispatch, Customer Master | |
| The account matched, but the ship to is not on file | FLAG | Scenario Dispatch, Customer Master | |
| The customer purchase order matches an open order | NEEDS_REVIEW | Scenario Dispatch, Customer Master |
The same run also finds revenue.
Alongside the credit decision, agents suggest up to three add-ons, each backed by the account's history. Suggestions are advisory and never change the order or the credit decision.
Exception handling that also finds revenue. It needs real order history to work.
ATTACHMENT
Attachment gap
A part that usually ships with this order, but is missing.
UPGRADE
Upgrade path
A higher spec the account has bought before, or its segment is moving to.
REORDER
Reorder prompt
Replenishment due, based on the account's own reorder interval.
CONTRACT_LEAKAGE
Contract leakage
The account is buying at list price an item already on its contract price list.
How to get started.
Four steps. Only the first needs your process owner.
Review your SOP with a Karini AI engineer
Our forward deployed engineer works with your process owner and queue team to turn your SOP into a clear spec: rules, thresholds, approval levels and gaps.
The gaps matter most. Where your SOP is silent, the queue is being worked from memory.
You bring
The SOP that governs the process today, in whatever state it's in.
Connect the workflow and your ERP
Your SOP becomes the workflow the agents follow. We connect through APIs, governed screen access where there's no API, and email intake.
Your ERP stays the system of record. Nothing is migrated or replaced.
You bring
An environment to connect to, and the person who owns access.
Test against your own history
Every rule runs on your past exceptions first. You see which rule each case hit, and fix the SOP where it's wrong.
Fixing it here costs an SOP edit, not a reversal.
You bring
A sample of closed cases, including the awkward ones.
Deploy to production
Live in your environment, behind your approval levels, with the audit trail on from day one.
The next use case starts at step one with its own SOP, not a new project.
Bring your SOP. We'll run it against your queue.
Pick one use case. If the rules are written down, agents can follow them. If not, we write them together in the first session.
