Skip to content
KanchanFlow
AI credits

One credit pool, thirteen agents, published weights

Every AI operation in KanchanFlow has a printed price in credits. There is no per-agent licence and no per-seat AI upgrade — the workspace has one balance and everyone draws on it.

Diagram titled "How AI credits are metered", showing the seven RFQ-to-order stages from enquiry intake through to accepted order.

In short

KanchanFlow, the RFQ-to-Order CRM for job shops, meters AI with one pooled credit balance per workspace, shared across all thirteen agents. Starter includes 100 credits per user per month and Professional 500, with every operation's credit weight published so a shop can predict its usage before buying.

Starter
100 credits / user / month
At the 3-user minimum that is a pooled 300 credits a month.
Professional
500 credits / user / month
A 5-person shop pools 2,500 credits a month.
Enterprise
Committed pool
Committed pool + fair use

The rate card

What every operation costs

These weights are the same numbers the product bills against. If an operation is not on this list, it does not consume credits.

Credit weights per completed operation. Failed runs on our side are not charged.
OperationCredits
RFQ email extract → RFQ object2
Spec-sheet OCR (per page)3
PO OCR (per page)3
Similar-part vector search1
Quote Follow-Up WhatsApp draft1
RFQ Qualification score1
Quote Risk Agent scan2
Reorder Agent prediction1
Dealer Health Agent scan1
Sales Forecast Agent (per forecast)5
Meeting summarization (≤30 min)3
Meeting summarization (30–60 min)6
Cross-Sell Agent scan (per account)2
Deal Risk scoring (per opportunity)1
Quote Probability prediction (per opportunity)1
Churn Detection scan (per account)1
Voice Agent qualification call (Enterprise)5
Custom AI agent run (Enterprise)Variable

Everything outside this table is free of credits: writing a quote, sending it, revisions and diffs, approvals, e-signature, dashboards, dealer records, territory management, notifications, the mobile app and every report. Credits pay for model inference, not for using the CRM.

Worked example 1

A Starter shop with 3 users

Three seats at 100 credits each gives the workspace 300 credits a month. Starter runs two agents — the RFQ Agent and Quote Follow-Up — and their operations weigh 2 and 1 credits respectively, so the blended average is close to 2 credits per operation. That is roughly 150 AI operations a month.

In practice that looks like about 100 inbound RFQ emails extracted into structured records at 2 credits each, plus 100 follow-up drafts at 1 credit each, and you would still have credits left. For a shop quoting 20 to 30 jobs a week, that is comfortable.

If you also drop spec sheets in for OCR at 3 credits a page, the arithmetic changes quickly — thirty pages is ninety credits, nearly a third of the pool. That is the one line to watch on Starter.

A representative Starter month

ActivityOpsCredits
RFQ email extract (2 credits)80160
Quote Follow-Up draft (1 credit)9090
Spec-sheet OCR pages (3 credits)1545
Total185295 of 300

Worked example 2

A Professional shop with 5 users

Five seats at 500 credits each pools 2,500 credits a month. Professional opens nine agents, and the mix shifts: qualification scoring, quote risk scans, deal risk and churn detection are 1 to 2 credits, meeting summaries are 3 to 6, and a sales forecast is 5. The blended average still lands near 2 credits per operation, which is roughly 1,250 operations a month.

Spread across five people and twenty-two working days, that is about 11 AI operations per person per day. A busy estimator running every inbound RFQ through extraction, qualification and a risk scan uses three of those on one quote — so the pool supports a genuinely AI-assisted desk, not a demo.

The two line items that move the total are meeting summarization and forecasting. A shop recording every customer call should expect those to be the largest share.

A representative Professional month

ActivityOpsCredits
RFQ email extract (2 credits)300600
RFQ qualification score (1 credit)300300
Quote Risk scan (2 credits)180360
Quote Follow-Up draft (1 credit)400400
Meeting summary ≤30 min (3 credits)60180
Sales forecast (5 credits)840
Reorder + dealer + churn scans (1 credit)300300
Total1,5482,180 of 2,500

Pooling

Why the pool is not locked per seat

Work in a shop is not evenly distributed and pretending otherwise makes software worse. Your senior estimator handles the difficult RFQs and will burn ten times the AI of the person who mostly answers the phone. If credits were locked to a seat, you would end up buying a seat purely to harvest its unused allowance, which is a silly way to run a business and an even sillier way to sell software.

So the allowance is calculated per paid user and then pooled. Usage is still attributed per user in the admin view — you can see who is spending what, and which agent is spending it — but nobody is throttled by a personal balance.

  • Warned at 80% — an in-app banner shows the pool balance, the agents consuming most of it and the $20 / ₹1,500 top-up option
  • Blocked at 100% — AI operations stop; every non-AI function of the CRM keeps working normally
  • Attributed per user and per agent — so you can see whether the spend is going where you want it
  • Agents can be switched off individually — including a frequency setting on the scheduled scans
  • Top up in 100-credit packs — available on every tier, stacking without a stated cap
See the credit top-up add-on

Limits

What credits do not buy

Credits never

  • Send a quote, an email or a WhatsApp message without a person pressing send
  • Change a price, a margin or a lead time on a live quote
  • Sign or accept anything on your behalf
  • Interpret GD&T, tolerance stacks or drawing callouts
  • Train a shared model on your drawings, quotes or customer data

Credits do

  • Read an RFQ email or attachment and produce a structured, editable record
  • Score, rank and flag — qualification, quote risk, deal risk, churn
  • Predict — reorder dates, quote probability, forecast
  • Draft text a person reviews and edits before it goes out
  • Summarize a recorded meeting into commitments and tasks

Credit questions

The allowance is calculated per paid user, but the balance is pooled for the whole workspace. Three Starter seats produce one pool of 300 credits a month that anyone can draw on. An estimator who reads forty RFQs is not throttled because the office manager read none.

See a live quote draft built from a real RFQ

Fourteen days, no credit card, sample data pre-loaded. If it does not fit your shop, we will tell you in the first call.

  • Delaware LLC
  • SOC 2 Type II
  • USA Data Centers (AWS)