AgentsStudioTemplatesPricingDocs
Sign inGet Started

Give your AI agent an automation engine.

Build

  • MCP server
  • kovca CLI
  • Kovca's agent
  • Studio
  • Templates

Run

  • Apps and events
  • Decision and Generative AI
  • Workflows API
  • Run history
  • Pricing

People in the loop

  • Approvals and waits
  • Tasks
  • Coding agents
  • Docs

Company

  • About
  • Security
  • Contact
  • Privacy Policy
  • Terms of Service
© Kovca. kovca.io
GuidesFinance & Procurement

Check a Gmail quote and log to Airtable

Build a Kovca workflow that starts from a new Gmail message, checks the customer quote in the body, logs the result to Airtable, and labels the message.

Quote Intelligence31 min readcURL
InputExtractLine countReady?Check quoteRecordRecord valuesLog quoteNext stepLabel loggedLabel reviewLabel missingResult
TemplateGmail quote logStart from a new Gmail message, check the customer quote the body states against your pricing rules, and log the result to Airtable.
13 nodesUp to 1 evaluation per run
Use this template

A deal desk gets draft customer quotes in a shared Gmail inbox. Someone reads the body, checks the prices and the discount against the rules, and types a row into a tracking table. A bad total or a discount over the limit is easy to miss, and the message sits there with no label.

Quote Intelligence does not read a supplier's email, and it does not pull a vendor name, a validity date, or terms out of free text. It checks a quote you already have as data: a quote id, a currency code, and at least one line with a product id, a name, a quantity, a list price, and a net price. It returns status (valid, issues_found, requires_approval, requires_review, or insufficient_information) and recommended_action (proceed, correct_quote, request_approval, or manual_review). The catalog describes it as quote validation for pricing errors, discount exceptions, configuration conflicts, and approval requirements. This guide is that check, for a quote a person pasted into an email, logged to Airtable.

This workflow starts when a new message arrives, copies the quote fields from the subject and the body when the email states them, and runs the check against pricing rules and an approval policy stored on the node. A quote that can be checked is logged to Airtable. proceed gets a logged label. Any other recommendation gets a review label. A message that does not yield a quote id, a currency code, and at least one line gets the same review label and no row.

The template already uses the Gmail New email trigger, Airtable Create records, and Gmail Add label three times. It does not store a Gmail or Airtable account. You choose the accounts when you activate. Until then the draft saves, and Studio tests simulate the Gmail and Airtable steps.

Log invoices that arrive in Gmail to a Google Sheet copies invoice header fields into a sheet. Check vendor invoices against the purchase order before payment matches a structured invoice to a purchase order. Neither checks a customer quote.

If the inbox only gets a few quotes a week, a person who reads the message and types the row is less machinery than this. A Gmail filter on subject:quote can label the mail. A filter does not compare the net price to the discount, and it does not compare the total to the lines. A quoting tool that checks the quote before anyone emails it is the other place to do this, when the quote never exists as text in the body.

What you'll build

ready and quote succeeded. quote is marked Decision AI. gap and logged are skipped. log and review are marked sample. The output shows the pricing discrepancy through the explanation: expected 850.00, actual 900.00.

StepNodeWhat it does
InputGmail: New email (gmail/new_email)Starts the workflow. The event fields are from, body_text, received_at, labels, message_id, thread_id, to, and subject.
ExtractGenerative AI (ai:extract@1)Copies quote_id, currency, total, and customer_requirements from the subject and the body when the email states them. lines is a list: one object per product, or an empty list when the email states no product.
Line countListCounts lines. An empty list counts as 0. A missing list counts as nothing.
Ready?ConditionTrue when quote_id and currency are present and the line count is at least 1.
Check quoteDecision AI (builtin:quotes/check@1)On the true branch. Checks the quote. Returns status, recommended_action, issues, approval_required, approval_reasons, computed_total, and missing_information.
RecordTextOne Airtable record as JSON text.
Record valuesConvertParses that JSON into the array Create records needs.
Log quoteAirtable Create records (action:airtable/create_records@1)Creates that record. Sent once. A retry can create another.
Next stepSwitchRoutes on recommended_action.
Label loggedGmail Add label (action:gmail/add_label@1)On proceed. Applies the logged label id. Safe to retry.
Label reviewGmail Add label (action:gmail/add_label@1)On any other recommendation. Applies the review label id. Safe to retry.
Label missingGmail Add label (action:gmail/add_label@1)On the false branch of Ready?. Applies the same review label id. Safe to retry.
OutputOutputThe card on the canvas says Result. Returns the quote fields, the check, and whether the row and the labels ran.

Prices, discounts, totals, required products, and excluded products are checked in code from the quote and from the rules you store on the node. A customer requirement, and a possible contradiction between two lines, are the parts that ask a question. This template's sample quote has one line and no customer requirements, so those questions are not asked.

Watch the test

Watch the Studio test (MP4). It is a recording of the production Studio at kovca.io, in light mode, 20 seconds. The frame is the page, with no browser chrome and no desktop around it. It opens on a clean Test panel: the widget quote is already in the input, and there is no earlier result. Nothing is sped up. No Gmail or Airtable account is connected, and Send real requests stays off. The clip scrolls the result until the end of the output is on screen, and holds there. Check quote succeeded. Log quote and Label review are marked sample. That is not a row in Airtable, and it is not a label applied in Gmail. Label logged and Label missing are skipped.

The output continues through the finding. expected 850.00, actual 900.00, computed_total 850.00, logged true, reviewed true, and gap null.

The recording follows these steps:

  1. The Test tab. The input is the widget quote: subject "Quote Q-2026-0151", from alex@example.com, to quotes@example.com, and a body that states quote Q-2026-0151, currency USD, total 900.00, and one WIDGET line. The input says Valid. Send real requests is off. The switch's own text says action steps return their sample result and nothing is sent. There is no result yet.
  2. Run test. The panel says Running, and that the previous result is cleared until this one finishes.
  3. The status is 200. The panel shows 7 credits used (1 Decision AI evaluation) and 1,340 Generative AI tokens.
  4. The clip scrolls to the steps and the output, then a little further, and holds on the end of the output. input, extract, count, and ready succeeded. Check quote (quote) succeeded and is marked Decision AI. Label missing (gap) is skipped. Record, Record values, and Log quote succeeded. Log quote is marked sample. Next step succeeded. Label logged is skipped. Label review succeeded and is marked sample. The output on screen at the end includes expected 850.00, actual 900.00, the explanation that the quote total does not equal the sum of the line amounts, computed_total 850.00, logged true, reviewed true, and gap null.

That is one live Studio test of that email. Extract returned quote_id Q-2026-0151, currency USD, total 900.00, and one line: product id WIDGET, name Industrial widget, quantity 10, list price 100.00, net price 85.00. Check quote ran. The JSON later in this guide is that run's output and steps. What Extract returns on another message depends on that message.

Build it in the Studio

1Start from the template

Click Use this template above, or open Gmail quote log in Workflows.

The draft opens on the canvas. The components library is on the left. Triggers, AI, Action in an app, Data transformation, Flow, Core, and Human review are the groups. This workflow uses Gmail and Airtable, so both are under Triggers and under Action in an app once the draft uses them. Apps you have not connected, and that this workflow does not use, stay behind Connect another app. The pictures below are the production Studio, in light mode, with no Gmail or Airtable account connected. The template's step titles are the ones in the table above. The last card on the canvas is Result.

Label missing. The card says Gmail: Add label and shows a warning, because no Gmail account is connected.

Triggers. Gmail and Airtable are in the list. Under the list: Showing the apps you've connected, and Connect another app. No account is connected on the steps.

Add label, Remove label, Mark as read, and Archive. The cards clip the longer descriptions. No Gmail account is connected.

Create records is the first Airtable action. The card clips the longer description. No Airtable account is connected.

2The workflow starts from a new email

Select the input step. It is Gmail: New email, not an API request. The account line asks you to connect Gmail. Leave it empty while you test.

Only emails matching is to:quotes@example.com subject:quote, fully visible. In label says Connect Gmail to choose a label. Apply settings is under that.

SettingTemplate value
Only emails matching (query)to:quotes@example.com subject:quote
In label (label_ids)unset

to:quotes@example.com subject:quote is a stand-in. Replace it with the search you actually want. The field's description is "Advanced Gmail search using the same syntax as Gmail's search box." A live message that does not match that search does not start a run. The Test tab and an API call send the event you provide, so they still run when the search would not have matched. The following-up example in the template has the subject "Following up on the draft", which this search would not start.

In label filters by a single label id. Until an account is connected, the field says to connect Gmail to choose a label. The template leaves it unset.

The trigger description says it fires when a new message is received in Gmail. The same message does not start two runs: the event id is message_id. A later message is a new event and runs again.

Event fields later steps read:

FieldUsed for
input.subjectExtract's instructions, and the Airtable record
input.body_textExtract's instructions
input.fromExtract's instructions, and the Airtable record
input.message_idThe three Add label steps, and the Airtable record

The event also includes received_at, labels, thread_id, and to. labels is a list of label ids already on the message. The trigger's sample uses INBOX and UNREAD. This workflow does not read that list. from is a string, the sender as Gmail reports it.

body_text accepts at most 8,000 characters. Extract's instructions, with the subject, the sender, and the body filled in, can be at most 60,000 characters of workflow data. A body at the event's own limit stays under that.

The body is what this workflow can read. The New email event carries body_text and does not carry attachments. A quote that exists only as a PDF is not read by this trigger, and this workflow does not fetch it. The prices have to be in the email body.

3Copy the fields the email states

Select Extract. It is the Extract data node. The instructions name the quote fields. The returns format is Fields (structured data).

quote_id, currency, and total are not marked required. lines is a list and is marked required. currency max length is 3. The description boxes clip the longer sentences.

The instructions say to copy a value that is written in the subject or the message, and to leave a field out only when it is not written. They say not to invent a product, a price, a currency, or a total. They say this is a quote the sender is offering a customer, not a request for a quote and not a supplier invoice. They ask for currency as the 3-letter code written in the message, and for prices and quantities as digits and a decimal point. lines is one object per product, with product_id, name, quantity, list_price, and net_price. The subject, the sender, and the body are inserted as {{input.subject}}, {{input.from}}, and {{input.body_text}}.

FieldWhat to put there
quote_idThe quote number the subject or the message states. At most 64 characters. Not required.
currencyThe 3-letter code the message states, such as USD. At most 3 characters. Not required.
totalThe quote total the message states, as digits and a decimal point, with no currency symbol. Not required.
linesOne object per product the message quotes. Required. An empty list when the message states no product. Each object needs product_id, name, quantity, list_price, and net_price.
customer_requirementsThe customer requirements the email states, one sentence each. Leave it out when the email states none. Not required.

quote_id, currency, total, and customer_requirements are not required on this step. A field the email does not state is left out. lines is required. A message with no product comes back as an empty list, which is a count of 0, so Ready? is false and Label missing runs. On each line, product_id, name, quantity, list_price, and net_price are required. If Extract returns a line that is missing one of those, the step fails and the run stops. Label missing does not run in that case. The template does not ask for discount_percent or amount on a line. The field list in the Studio keeps a nested list as it was defined and points you at the JSON view to change that shape. The description boxes clip the longer sentences. customer_requirements is the fifth field. Its key is wider than the key box in that list, so the box clips the end of the name. The description under it is "The customer requirements the email states, one sentence each."

The template does not set a tier, a creativity, or a reasoning level, so those stay at the Extract data defaults: Fast, which that menu marks recommended, Precise, and Standard. It does set the maximum length to Long (4096 tokens). A quote with line objects is longer than the Extract default of Medium, and a cut-off answer fails the step.

4Decide whether the quote can be checked

Line count counts nodes.extract.output.lines. If that value is missing or is not a list, the count is empty.

Select Ready?. A condition continues on true when every rule matches. This one has three rules.

Continue on true when all rules match. nodes.extract.output.quote_id is present. nodes.extract.output.currency is present. nodes.count.output.result is at least 1. The three references are fully visible.

RuleValue
Continue on true when all rules matchnodes.extract.output.quote_id is present
nodes.extract.output.currency is present
nodes.count.output.result ≥ 1
CaseWhenGoes to
trueQuote id, currency, and at least one lineCheck quote
falseAny of those is missingLabel missing

An empty lines list has a count of 0, so it takes the false branch. A line that is missing a required price fails Extract before Ready? runs, and the run stops on Extract. It does not apply a label.

5Check the quote

Check quote runs only on the true branch. It is Quote Intelligence.

Every reference is fully visible, including nodes.extract.output.customer_requirements. pricing_rules is max_discount_percent 25 and price_tolerance 0.01. approval_policy is total_over 50000 and discount_percent_over 15.

InputValue
quote_idnodes.extract.output.quote_id
currencynodes.extract.output.currency
linesnodes.extract.output.lines
totalnodes.extract.output.total
customer_requirementsnodes.extract.output.customer_requirements
pricing_rulesFixed value, below
approval_policyFixed value, below
JSON
{
  "max_discount_percent": 25,
  "price_tolerance": "0.01"
}
JSON
{
  "discount_percent_over": 15,
  "total_over": "50000"
}

max_discount_percent is the highest discount a line may take. A larger discount is a discount_exception. price_tolerance is how far a computed price may be from the stated one before it is a pricing_discrepancy. The template uses 0.01. discount_percent_over and total_over are the approval policy. A discount above 15, or a computed total above 50000, adds an approval reason. These are the template's stand-in rules. Replace them with yours. Changing them is a new publish.

The node can also take pricing_rules.requires (a product that must be quoted with another) and pricing_rules.excludes (two products that cannot be quoted together). This template does not set them. A line can also carry attributes. This template does not ask Extract for attributes.

What comes back:

statusrecommended_actionWhen
issues_foundcorrect_quoteA pricing discrepancy, a discount over the maximum, a missing required product, an excluded pair, a requirement treated as not met, or a configuration conflict treated as clear.
requires_reviewmanual_reviewA requirement that is not clearly met, or a configuration conflict that is only possible.
requires_approvalrequest_approvalNo issue, and a discount or the total is over the approval policy.
insufficient_informationmanual_reviewPricing rules, the approval policy, and customer requirements are all missing. This template sets the first two, so this status does not come back unless you clear both.
validproceedNone of the above.

missing_information lists which of pricing_rules, approval_policy, and customer_requirements you did not supply. proceed can still list customer_requirements there when the email stated none. That is not a failed check. The requirements were not asked. approval_required is true or false when you supplied a policy, and null when you did not. approval_reasons says which threshold was crossed. An issue is decided before approval, so a quote can be issues_found and still have approval reasons.

computed_total is the sum of the line amounts, as a decimal string. A stated line amount is used when the line has one. Otherwise the line amount is quantity times net price.

A quote with two or more lines is also asked whether any two lines contradict each other. Each customer requirement is asked on its own. This template's widget example has one line and no requirements, so neither question is asked. The template does not copy a stated discount percent or a stated line amount. The line amount used in the check is quantity times net price. For the widget example that is 10 times 85.00, which is 850.00. The stated total 900.00 does not match, so the one issue in the JSON below is that pricing discrepancy. The discount implied by the list price and the net price is 15 percent. That is not above the 25 percent maximum, and it is not above the 15 percent approval threshold.

6Route on the recommendation

Select Next step. A switch reads one value and continues on the branch of the first case that lists it, or on default.

Route on nodes.quote.output.recommended_action, fully visible. The case is named proceed and matches proceed. Anything else is default.

CaseMatches recommended_actionGoes to
proceedproceedLabel logged
defaultcorrect_quote, request_approval, manual_reviewLabel review

Log quote runs before the switch, on every quote that was checked. Both labels are after the row is created. This template does not open an approval wait. request_approval is labeled for a person and logged. It is not sent to an inbox by itself.

7Log the row

Record builds one JSON record. The keys inside fields are stand-in column names.

The template is one record. The placeholders for quote_id, currency, status, recommended_action, computed_total, approval_required, from, subject, and message_id are visible.

[{"fields":{"quote_id":{{json nodes.extract.output.quote_id}},"currency":{{json nodes.extract.output.currency}},"status":{{json nodes.quote.output.status}},"recommended_action":{{json nodes.quote.output.recommended_action}},"computed_total":{{json nodes.quote.output.computed_total}},"approval_required":{{nodes.quote.output.approval_required}},"from":{{json input.from}},"subject":{{json input.subject}},"message_id":{{json input.message_id}}}}]

Create records describes records as an array of record objects. The tool this action calls says each object must contain a fields property, whose keys are the column names. The template follows that. Each text value uses {{json …}}, which writes a JSON string. A quote, a backslash, or a line break in the quote number, the currency, the status, the recommendation, the total, the sender, the subject, or the message id stays in the cell. approval_required is a boolean, so it is inserted as true or false. The issues list stays on the workflow output. It is not a column on this row.

Record values parses that text into JSON.

Log quote is Airtable's Create records action.

If it fails is Stop the run. Base ID is appQuoteLogStandIn01, fully visible. Table is Quotes. Records is nodes.rows.output.result. Typecast is true.

Create records fieldValue
Base ID (base_id)appQuoteLogStandIn01
Table (table_id_or_name)Quotes
Records (records)nodes.rows.output.result
Typecast (typecast)true

The base id and the table name are stand-ins. Replace them with your base id (it starts with app) and your table id or table name. The column names in fields have to be columns on that table. Typecast asks Airtable to convert string values to the column type. The field's description is "If true, Airtable will perform best-effort automatic data conversion from string values to the appropriate cell type." The action's description says it creates up to 10 records. This template sends one.

If it fails is Stop the run. The action is not marked safe to retry: a second attempt can create another record. If the create does not succeed, the run stops and neither label runs. Choose the Airtable account on this step before you activate. In a Studio test, with Send real requests off, the step returns its sample and is marked sample. The sample is ok: true and result: {}. That is not a record id, and it is not a row in a base.

8Label the message

Label logged runs on proceed. Label review runs on default. Label missing runs when Ready? is false. Review and missing send the same label id, so one Gmail label covers both. Logged is a different id.

If it fails is Continue (output.ok is false). Message ID is input.message_id. Label IDs is the fixed value Label_review. Both values are fully visible.

Add label fieldValue
Message ID (message_id)input.message_id
Label IDs (add_label_ids)Label_logged on Label logged. Label_review on Label review and on Label missing.

The field's description says these are label ids, not display names. Label_logged and Label_review are stand-ins. Replace them with the ids Gmail lists once the account is connected.

If it fails is Continue (output.ok is false). Add label is safe to retry: repeating it does not apply the label twice in a way that changes the message further. The step can be tried again within the run. A failed label does not stop the run. In a Studio test, with Send real requests off, the label that runs returns its sample and is marked sample. The sample is ok: true and result: {}. That is not a label applied in Gmail.

9Test it

Open the Test tab. The input is the trigger's event. Sample new email fills the trigger's sample. Replace it with a message you want to check. Send real requests stays off until you mean to create a row and apply a label.

The input says Valid. Sample new email and Recent from Gmail are the trigger's buttons. The subject is Quote Q-2026-0151. The body states currency USD and one WIDGET line.

The widget example:

FieldValue
fromalex@example.com
toquotes@example.com
subjectQuote Q-2026-0151
body_textQuote Q-2026-0151, currency USD, total 900.00, one line WIDGET, quantity 10, list price 100.00, net price 85.00, discount percent 10, amount 850.00
message_id18f0b1c2d3e4f5a1

The unit test does not read the email. It supplies an empty lines list and no other fields. line_count is 0, Ready? is false, and Label missing is the only action that runs. Check quote, Record, Record values, Log quote, Next step, Label logged, and Label review are skipped.

A second unit test supplies quote fields so Check quote can run. That supplied object is not a live read of the email. The live Studio test above is the read. On that run Extract returned the quote id, USD, the total, and one complete line, and Check quote returned issues_found and correct_quote. The stated total 900.00 is not the line amount, which is quantity times net price, 850.00. Log quote and Label review return their samples. Label logged and Label missing are skipped. That Studio run is the JSON in the next section.

The following-up example has the subject "Following up on the draft" and no prices. The search would not start it. In the unit test, lines is an empty list and Label missing runs. A Studio test of that note also returned an empty lines list, with quote_id, currency, and total left out, ready false, and Label missing as the action that runs.

Before you activate, run messages you have actually seen, including a complete quote and a note that is not a quote. When the routes look right, turn on Send real requests once, with both accounts chosen and the stand-in ids replaced, on a message where an extra row and an extra label are harmless.

Publish, call it, and activate it

The same workflow can be called over the API. The body uses the trigger's field names. Create an API key under API keys and keep it on your backend.

cURL
curl https://kovca.io/api/v1/workflows/wf_.../execute \
  -H "Authorization: Bearer $KOVCA_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: 18f0b1c2d3e4f5a1" \
  -d '{
    "input": {
      "from": "alex@example.com",
      "to": "quotes@example.com",
      "subject": "Quote Q-2026-0151",
      "body_text": "Quote Q-2026-0151\nCurrency: USD\nTotal: 900.00\n\nproduct_id: WIDGET\nname: Industrial widget\nquantity: 10\nlist_price: 100.00\nnet_price: 85.00\ndiscount_percent: 10\namount: 850.00",
      "message_id": "18f0b1c2d3e4f5a1",
      "thread_id": "18f0b1c2d3e4f5a1",
      "received_at": "2026-03-12T15:04:00Z",
      "labels": ["INBOX", "UNREAD"]
    }
  }'

That curl calls a published workflow. Extract reads the body. If the fields come back complete, Check quote runs, Log quote creates a record, and one of the two labels runs, for real. If the fields do not come back, Label missing runs for real. The steps need the Airtable account on Log quote and the Gmail account on the label that runs. If an account is missing, the action fails. With If it fails set to stop on Log quote, a failed create stops the run and the label does not run. A published call does not return the sample you see with Send real requests off.

The JSON below is the output of the Studio test in the recording, with Send real requests off. Check quote finds the total discrepancy. Log quote and Label review are simulated. Label logged and Label missing are skipped. This is not a response from a published API call. The unit test that supplies quote fields checks the same route. It is not this response.

JSON
{
  "request_id": "req_...",
  "run_id": "run_...",
  "workflow": { "id": "wf_...", "version": 1 },
  "status": "succeeded",
  "output": {
    "quote_id": "Q-2026-0151",
    "currency": "USD",
    "total": "900.00",
    "line_count": 1,
    "ready": true,
    "status": "issues_found",
    "recommended_action": "correct_quote",
    "issues": [
      {
        "type": "pricing_discrepancy",
        "method": "deterministic",
        "expected": "850.00",
        "actual": "900.00",
        "explanation": "The quote total doesn't equal the sum of the line amounts.",
        "evidence": ["total", "lines"]
      }
    ],
    "approval_required": false,
    "approval_reasons": [],
    "computed_total": "850.00",
    "missing_information": ["customer_requirements"],
    "route": "default",
    "record": "[{\"fields\":{\"quote_id\":\"Q-2026-0151\",\"currency\":\"USD\",\"status\":\"issues_found\",\"recommended_action\":\"correct_quote\",\"computed_total\":\"850.00\",\"approval_required\":false,\"from\":\"alex@example.com\",\"subject\":\"Quote Q-2026-0151\",\"message_id\":\"18f0b1c2d3e4f5a1\"}}]",
    "logged": true,
    "labeled": null,
    "reviewed": true,
    "gap": null
  },
  "steps": [
    { "id": "input", "status": "succeeded" },
    { "id": "extract", "status": "succeeded" },
    { "id": "count", "status": "succeeded" },
    { "id": "ready", "status": "succeeded" },
    { "id": "quote", "status": "succeeded" },
    { "id": "gap", "status": "skipped", "reason": "branch_not_taken" },
    { "id": "line", "status": "succeeded" },
    { "id": "rows", "status": "succeeded" },
    { "id": "log", "status": "succeeded" },
    { "id": "route", "status": "succeeded" },
    { "id": "logged", "status": "skipped", "reason": "branch_not_taken" },
    { "id": "review", "status": "succeeded" },
    { "id": "output", "status": "succeeded" }
  ],
  "evaluations": 1
}

The JSON above leaves out credits. A succeeded execute response includes that field. This example is the simulated run, so Log quote and Label review are not charged. credits for an API call of this shape is 6 plus Extract data: 5 for Check quote and 1 for the API run. Extract data is at least 1 credit more, by tokens. The Studio test does not charge the run fee.

Look up the run later. This returns status, timing, and steps. It does not return the output. Keep the execute response if you need the fields or the status. Runs are kept for 30 days.

cURL
curl https://kovca.io/api/v1/runs/run_... \
  -H "Authorization: Bearer $KOVCA_API_KEY"

18f0b1c2d3e4f5a1 is a key you choose. The message id is a natural one. Keys are kept for 24 hours. A retry with the same key and the same request returns the original run instead of creating another record. A new key can create another record. Add label can be repeated without a second effect. Create records cannot: a later run that checks the quote creates another row. Kovca does not store the message or the output.

To run it from Gmail, open Activate. The checklist is what's left: connect Gmail, connect Airtable, and, once an account is connected, choose that account on the trigger, on Log quote, on Label logged, on Label review, and on Label missing. It says up to 19 credits per event, plus Generative AI by tokens. The activate row is Publish and activate: run on every new email. The button on that row says Publish & activate. That publishes the draft and starts it. A live automation runs that pinned version. Later edits don't reach it until you publish them to the automation.

Connect Gmail and Connect Airtable are still to do. Up to 19 credits per event, plus Generative AI by tokens. Publish and activate: run on every new email. The button says Publish & activate.

Connect Gmail and Airtable from the checklist, or from the steps, when the routes in the Test tab look right. Until then, leave the accounts empty and keep Send real requests off.

A new message that passes the trigger settings starts a run when Gmail sends the event. Follow runs on the workflow's Executions page (Studio → Usage → Executions). While an automation is paused, events that arrive are skipped. Kovca does not keep them to run later. A failed run is not retried on its own, because the row may already have been created.

If someone already reads the inbox

You don't need this if a person can read the shared inbox and the volume is small. Gmail can filter on subject:quote, and a table is a fine place to type the quote by hand. A filter does not read the body, and it does not compare the total to the lines.

What the workflow adds is the same copy of the quote fields on every new message that matches the search you set, a check of those fields against the rules on the node, a row when the quote could be checked, a logged label when the recommendation is proceed, and a review label otherwise. You still choose the accounts, replace the stand-in search, base id, table, column names, and label ids, and read a sample of results before you activate it.

Make it yours

  • Point the search at the real inbox. to:quotes@example.com subject:quote is a stand-in. Use the search your mailbox actually needs. Set In label once Gmail is connected if the mail already lands under one label.
  • Replace the label ids. Label_logged and Label_review are not ids from a mailbox. Use the ids Gmail lists once the account is connected. Label review and Label missing share Label_review. Use two ids if you want missing quotes and checked quotes in different labels.
  • Point the base and the table at the real table. appQuoteLogStandIn01 and Quotes are stand-ins. The keys inside fields have to be columns that already exist. Typecast converts strings. It does not create columns.
  • Replace the pricing rules and the approval policy. 25 percent, a tolerance of 0.01, 15 percent, and a total of 50000 are stand-ins. Add requires or excludes when a product has to be quoted with another, or cannot be quoted with another.
  • Leave a partial quote off the table. That is already what the false branch does. A missing quote id, currency, or line gets Label missing and no row.
  • Do not expect a PDF attachment to be read. The trigger's event has body_text and no attachment fields. A quote that exists only as a PDF is not in the event.
  • Do not treat proceed as a check of customer requirements. If the email stated none, missing_information contains customer_requirements and the status can still be valid.
  • Change it without surprising a live mailbox. Edit the draft and publish the changes to the automation. The running version stays pinned until you do.

What it costs

Credits, and only for steps that succeed. Billing is the list.

A Studio test skips the run fee. It does not skip Extract data or Check quote. Extract data is Generative AI, billed by tokens, at least 1 credit, including a test with Send real requests off. Check quote is 5 credits when it succeeds, including that test. A simulated action is not charged, and an action that fails is not charged. A skipped action is not charged. With Send real requests off, the action that runs returns its sample and adds nothing.

  • A Gmail event: 6 credits for the run.
  • A published API call: 1 credit for the run.
  • Extract data: Generative AI, by tokens, at least 1 credit, on every run.
  • Check quote: 5 credits when it succeeds. Only the true branch is charged.
  • Log quote: 2 credits when it runs for real and succeeds. Only the true branch is charged.
  • Label logged: 2 credits when it runs for real and succeeds. Only the proceed branch is charged.
  • Label review: 2 credits when it runs for real and succeeds. Only the default branch is charged.
  • Label missing: 2 credits when it runs for real and succeeds. Only the false branch is charged.

Line count, the condition, the switch, the text step, and the convert use no credits. A branch that doesn't run isn't charged. The Activate panel adds every step together, including the branches that do not run on the same event, so its number is an upper bound for a live event, and it adds Generative AI by tokens on top. With this template that upper bound is 19 credits per event, plus Extract data: the 6 credit event fee, Check quote, Log quote, and all three labels. A live event that logs and labels logged is 15 plus Extract data when those steps succeed. A live event that logs and labels review is 15 plus Extract data. A live event that only applies the review label, because the quote could not be checked, is 8 plus Extract data. A published API call is 10 plus Extract data when it logs and labels, and 3 plus Extract data when it only applies the missing label. The response's credits field is the run's total. See the workflows API, HTTP actions, and automations.

On this page

  • What you'll build
  • Watch the test
  • Build it in the Studio
  • Publish, call it, and activate it
  • If someone already reads the inbox
  • Make it yours
  • What it costs

Built on Quote Intelligence: Validate quotes for pricing errors, discount exceptions, configuration conflicts, and approval requirements.

Keep going

Workflows API
InputInvoice Intel…Next stepApprove in APNote to vendorOutput
Invoice Intelligence8 min read

Check vendor invoices against the purchase order before payment

Build a Kovca workflow that matches each vendor invoice to its purchase order, approves clean invoices in your AP system, and drafts a note to the vendor when prices or quantities don't match. Step by step in the Studio, with a ready-made template.

Read guide
InputLead textLead Intellig…Follow up?Search contac…New contact?Create contactOutput
Lead Intelligence24 min read

Add qualified sales emails from Outlook to HubSpot

Build a Kovca workflow that starts from a new Outlook email, classifies it, and creates a HubSpot contact only when no matching contact is found.

Read guide
InputShipment Inte…Needs action?Ticket subjectOpen support …Output
Shipment Intelligence8 min read

Catch delivery exceptions from carrier tracking events

Build a Kovca workflow that reads every carrier tracking update, spots delays, failed deliveries, address problems, and disputes, and opens a support ticket when someone needs to act. Runs from the carrier's webhook. Step by step in the Studio, with a ready-made template.

Read guide
Sign inGet Started