Articles in this section

Use Celigo AI assist for Amazon - NetSuite inventory and settlements

Celigo AI assist is an add-on for the Amazon-NetSuite integration app that uses Celigo AI to explain inventory and settlement problems, without changing your inventory, listings, or financial records. It adds two flows to each Amazon store: one forecasts when your FBA and merchant-fulfilled (MFN) SKUs will run out, calls out SKUs that an upcoming sales event puts at risk, and flags FBA inventory exceptions, and the other investigates Amazon settlements that don't reconcile and explains the root cause and the steps to fix them. AI assist saves all results in NetSuite, where you review them on a Celigo AI assist dashboard.

This article explains how AI assist becomes available in your account, how each flow runs, how to turn on and run the flows, and how to read the results in NetSuite.

Contents

Use cases

Inventory and settlement problems often come to light only after they cost money: a SKU goes out of stock the week before Black Friday, or month-end close stalls on a settlement that's off by a few hundred dollars. Use AI assist to:

  • Plan replenishment around sales events: See which SKUs will run out before or during an event such as Prime Big Deal Days, Black Friday, or Cyber Monday, and which overstocked SKUs could sell down during it.
  • Catch FBA stock problems early: Get a NetSuite task and an inventory exception record when a SKU is out of stock, is at or below your fulfillable quantity threshold, has units inbound but none fulfillable, or holds unfulfillable units.
  • See FBA availability in NetSuite: Each run copies fulfillable, unfulfillable, reserved, inbound, and total quantities to NetSuite item fields, without changing NetSuite inventory.
  • Reconcile settlements faster: Get the root cause, the impacted transactions, an explanation of each fee from Amazon's documentation, and an ordered list of what to check, instead of reviewing every settlement line.
  • Spot recurring problems: An investigation whose cause also occurred in an earlier settlement is marked Recurring.

What AI assist includes

AI assist adds these two flows to the AI assist flow group of each Amazon store:

Flow What it does How it runs Default
Forecast Amazon inventory and triage FBA exceptions Forecasts stock for FBA and MFN SKUs, researches upcoming Amazon sales events, copies FBA availability to NetSuite item fields, and flags FBA inventory exceptions with a likely cause and recommended checks. Scheduled or on demand. It runs on the schedule you add, or when you select Run. Turned off
Investigate Amazon settlement variances Finds Amazon settlements that don't reconcile, diagnoses the root cause, explains each fee, and lists the steps to fix the settlement, in order. Scheduled or on demand. It runs on the schedule you add, or when you select Run. Turned off

AI assist also adds a Celigo AI assist tab to NetSuite with a dashboard and lists of the AI assist records. The results of both flows appear only in NetSuite. The integration app in integrator.io shows the flows and their run history, not the forecasts, exceptions, or investigations.

AI assist is advisory. It reads your Amazon inventory data and NetSuite items, sales, and settlement transactions, and writes only its own NetSuite records, tasks, and the FBA quantity fields on items. It never writes inventory quantities to NetSuite or Amazon, never changes listings, and never writes to financial or settlement records. It doesn't change how the integration app's order, inventory, or settlement flows behave. AI assist runs on your integration app's existing Amazon and NetSuite connections; you don't need a new connection.

AI assist uses Celigo AI credits: one AI action for each FBA SKU flagged as an exception, one for each settlement investigated, and one for sales-event research per store, per marketplace, per day for FBA, and one for MFN. Later runs on the same day reuse the researched events. FBA research can run again that day if the FBA Manage Inventory Health report includes a product group that wasn't researched yet. To monitor usage, see Track your AI agent actions and credits usage.

How AI assist uses Celigo AI

Rules in the flows decide whether something is a problem and calculate every number: detection, severity, root cause, forecast quantities, and stockout dates. The AI agent steps add the explanation and the research, so the results are repeatable and you can trace every number back to the source records.

AI agent step What it adds
Analyze FBA inventory exception A plain-language explanation of why a flagged FBA SKU needs attention, up to three recommended checks, and the reason for the severity.
Research upcoming Amazon sales events (FBA and MFN) Amazon and retail events in the next 90 days that affect your product groups, with dates, expected demand impact, and sources. It runs once per store per day; later runs that day reuse the results.
Investigate Amazon settlement variance An explanation of the settlement variance, an explanation of each fee involved, an ordered list of steps to fix it, evidence citations, and whether the cause has occurred before. It researches fees only in Amazon Seller Central and Amazon developer documentation. If no Amazon source describes a fee, it says so.

The AI agent steps are locked: you can't edit their instructions, model, or output format. For more about AI agents in integrator.io, see Learn about Celigo AI agents.

How AI assist becomes available

AI assist needs two updates: a NetSuite integration bundle update, which adds the Celigo AI assist tab and records to NetSuite, and an Amazon - NetSuite integration app update, which adds the AI assist add-on to integrator.io.

NetSuite integration bundle update

Note: On existing integration apps, the AI assist add-on comes pre-installed. If you don't want to use it, you can uninstall it and install it again later.

AI assist add-on in integrator.io

How you get the AI assist add-on depends on your integration app:

  • Existing integration apps: Celigo installs the AI assist add-on for you on every existing Amazon - NetSuite integration app, for every store in it. You don't need to install anything. The two AI assist flows appear in the AI assist flow group, and the Celigo AI assist card on the Add-ons tab shows Uninstall because the add-on is already installed.
  • A new store added to an existing integration app: Install the AI assist add-on yourself from the Add-ons tab for that store.
  • A new Amazon - NetSuite integration app tile: Install the AI assist add-on yourself from the Add-ons tab.
  • New Amazon - NetSuite integration app customers: Install the AI assist add-on yourself from the Add-ons tab.

Install the AI assist add-on

Prerequisites: The NetSuite integration bundle in the NetSuite account connected to the integration app includes the AI assist objects. See NetSuite integration bundle update.

AI assist is free and doesn't need an add-on license. It's installed per store, and each store gets its own copy of the flows. For general information about add-ons, see Understand integration app add-ons.

  1. Go to Home.
  2. Select the Amazon - NetSuite integration app tile.
  3. In the store list of the integration app, select your Amazon store.
  4. Go to Admin > Add-ons.
  5. On the Celigo AI assist card, select Install.

After installation, the two AI assist flows appear on the Flows tab under AI assist.

Find the AI assist flows in integrator.io

The AI assist flow group lists the two AI assist flows for the selected store, where you turn them on, schedule them, and run them.

  1. Go to Home.
  2. Select the Amazon - NetSuite integration app tile.
  3. In the store list of the integration app, select your Amazon store.
  4. Select the Flows tab.
  5. In the Flows list, select AI assist.

The AI assist flows page lists the two flows for the selected store. The store name appears in brackets at the end of each flow name. The Schedule column shows a schedule icon for both flows.

To open a flow in the flow builder, select its name. The flow builder shows the flow steps, the Enabled toggle, Run, Schedule, and the Run history tab.

Turn on the flows

Both flows are turned off by default. Review the AI assist settings first, and then turn on each flow you want to use: on the AI assist flows page, select the toggle in the Enabled column for that flow.

Run the flows

Run a flow in either of these ways:

  • On a schedule: On the AI assist flows page, select the schedule icon in the Schedule column, or open the flow and select Schedule. Set how often the flow runs, and save the schedule. A daily schedule suits most stores: the sales-event research runs once per store, per marketplace, per day, and Amazon refreshes the FBA Manage Inventory Health report about once a day.
  • On demand: On the AI assist flows page, select the Run icon, or open the flow and select Run.

To follow a run, open the flow. The Run console shows each step's status and its Success, Ignored, Errors, and Auto-resolved counts. To review earlier runs, select the Run history tab.

Configure the AI assist settings

The AI assist settings control how the forecast flags SKUs, how the flow finds NetSuite items, and which settlements the settlement flow investigates. The settings apply to the store selected in the integration app.

  1. Go to Home.
  2. Select the Amazon - NetSuite integration app tile.
  3. In the store list of the integration app, select your Amazon store.
  4. Select the Settings tab.
  5. In the Settings list, select AI assist.
  6. Select Save after you change a setting.

Inventory forecast

Setting Default Description
Scenario range for forecast (%)

10

A percentage from 0 to 100. The forecast's best-case and worst-case band assumes the SKU sells this much slower and faster than its expected sales rate, so the band widens the further out the forecast goes. Upcoming sales events and Amazon seasons still pull the worst case in. Applies to both FBA and MFN SKUs.
Fulfillable quantity threshold (units)

0

SKUs whose FBA fulfillable quantity is at or below this value are flagged as FBA inventory exceptions. SKUs with unfulfillable units are always flagged. The default of 0 flags only stockouts.
NetSuite item field ID for Amazon SKU

itemid

The internal ID of the NetSuite item field that stores the Amazon SKU. For example, the ID of the Item Name/Number field is itemid. The flow uses this field to find which NetSuite item gets the FBA quantity fields.

Settlement variance investigation

Setting Default Description
Ignore reconciliation variances at or below (amount)

0

The variance tolerance in the settlement currency. Settlement transactions whose absolute reconciliation variance is at or below this amount aren't investigated. Transactions with a reconciliation status other than Applied are always investigated. The default of 0 investigates every non-zero variance.
Investigate settlements that ended in the last (days)

0

A number of days from 0 to 90. The flow investigates only settlements whose end date falls between today and that many days ago. All transactions of a settlement are picked up together, and settlement transactions without a settlement summary are skipped. The default of 0 investigates every flagged settlement, whatever its date.

Forecast inventory and triage FBA exceptions

The Forecast Amazon inventory and triage FBA exceptions flow reads three sources: the FBA inventory summaries and the FBA Manage Inventory Health report from Amazon, and the store's NetSuite inventory for MFN items. The Route by inventory source router sends each record to every branch whose filter it matches.

Branch What it processes Result in NetSuite
FBA inventory snapshot Every SKU from the FBA inventory summaries. Writes the fulfillable, unfulfillable, reserved, inbound, and total quantities to the item's FBA fields, and adds a replenishment snapshot.
FBA inventory exceptions SKUs whose fulfillable quantity is at or below the Fulfillable quantity threshold (units) setting, or that have unfulfillable units. The AI agent explains each exception. The flow creates a NetSuite task and adds or updates an inventory exception record.
FBA triage run One summary per page of FBA inventory summaries. Logs how many SKUs were scanned and flagged.
Inventory forecast The FBA Manage Inventory Health report. Reuses the sales events researched earlier that day, or researches them if none exist. Writes a forecast for each SKU and a forecast run.
MFN inventory forecast The store's Amazon-linked NetSuite items, with the units each item sold from every sales channel at the store's inventory locations. Researches events in the same way, and writes MFN forecasts and a forecast run.

The FBA quantity fields on the NetSuite item are the only fields AI assist writes on standard NetSuite records. The flow updates existing items only: if an Amazon SKU has no matching NetSuite item, the run shows an error for that SKU. See: A SKU has no matching NetSuite item.

FBA exception causes and severity

Each flagged SKU gets one likely cause. The flow checks the rules in this order and uses the first one that matches:

Likely cause When it applies
inbound_pipeline_stall No fulfillable units are left, but units are inbound.
unfulfillable_units The SKU has unfulfillable units that can't be sold or shipped.
reserved_spike No fulfillable units are left, but units are reserved.
stockout_risk The fulfillable quantity is at or below your threshold.
Severity Meaning
High Out of stock, with or without units inbound.
Medium At or below the fulfillable quantity threshold.
Low Only unfulfillable units.

When the AI agent isn't confident about the likely cause, the exception is marked Needs human review.

For each exception, the flow creates a NetSuite task titled, for example, FBA inventory exception: SKU-123 (stockout_risk). The task lists the SKU's quantities, the likely cause, severity, confidence, explanation, recommended checks, and severity rationale. If the AI agent isn't available for a run, the task recommends standard checks instead: review replenishment plans and inbound shipments when the fulfillable quantity is 0, and review item dispositions and removals in Seller Central when unfulfillable units exist.

How the forecast works

The flow calculates these values for each FBA and MFN SKU:

  • Sales rate and demand trend: A weighted daily sales rate (Units a day), and a demand trend of Rising, Steady, Falling, or No sales.
  • Days of cover and runout date: How long the stock lasts at the current sales rate. Inbound units count as sellable only after a 14-day inbound lead time.
  • Best and worst case: A range based on the Scenario range for forecast (%) setting.
  • Stock status: Out of stock, Stockout risk (under 30 days of cover), Healthy, Excess, or No sales.
  • Event risk: High, Medium, Opportunity, or None, with a short explanation (callout). Opportunity means an overstocked SKU with falling demand could sell down during the event. FBA SKUs also show Amazon's recommended action and ship-in quantity and date.

Examples of event risk callouts:

  • High: At 0.27 units a day, this SKU runs out around 2026-10-09, before the event starts.
  • High: About 40 units will be left when it starts, but the expected lift needs about 90.
  • Medium: Stock should last through the event but runs low within two weeks after it.
  • Opportunity: Demand for this SKU is falling, and it has excess stock, so the event is a chance to clear units with a deal or coupon.

Forecast flow steps

The flow builder and Run console show these steps, among others:

  • Get FBA inventory summaries from Amazon
  • Get FBA manage inventory health report from Amazon
  • Get MFN inventory from NetSuite
  • Post FBA inventory snapshot to NetSuite items
  • Analyze FBA inventory exception (AI agent)
  • Post FBA inventory exception summary to NetSuite tasks
  • Post FBA inventory exception to Celigo AI assist records
  • Post FBA replenishment snapshot to Celigo AI assist records
  • Post FBA triage run to Celigo Amazon AI assist record
  • Get today's Amazon forecast events from NetSuite
  • Research upcoming Amazon sales events (AI agent)
  • Post Amazon inventory forecast by SKU to Celigo Amazon AI assist record
  • Post Amazon inventory forecast run to Celigo Amazon AI assist record
  • Get MFN item sales from NetSuite
  • Get today's Amazon MFN forecast events from NetSuite
  • Research upcoming Amazon sales events for MFN items (AI agent)
  • Post Amazon MFN inventory forecast by SKU to Celigo Amazon AI assist record
  • Post Amazon MFN inventory forecast run to Celigo Amazon AI assist record

Investigate settlement variances

The Investigate Amazon settlement variances flow works on the settlement transactions that the integration app's settlement flows bring into NetSuite (see Understand the Amazon settlement reports and Sync your Amazon settlement report (manually or automatically)). It investigates Amazon settlements that don't reconcile and saves one settlement investigation record for each settlement and root cause.

Each run works through these steps:

  1. Find flagged transactions: Settlement transactions with a reconciliation status other than Applied, or with a variance above the Ignore reconciliation variances at or below (amount) setting, grouped by settlement.
  2. Read the whole settlement: The flow reads every flagged transaction of the settlement, even when the settlement continues on the next page.
  3. Diagnose: The flow totals the settlement and assigns a root cause (the cause with the largest unmatched amount), the share of the variance it explains, the impacted transactions, and the fee lines.
  4. Check earlier investigations: Earlier investigations for the same store show whether the cause has occurred before, so recurring causes are flagged.
  5. Explain: The AI agent writes the explanation, explains each fee from Amazon's documentation, and lists the steps to check, in order.
  6. Save: The flow adds or updates the settlement investigation record. The Outcome you set on a record is kept on later runs.

Root causes

Root cause When it applies
marketplace_or_currency_unidentified The settlement's marketplace couldn't be identified.
refund_unmatched_order_item A refund couldn't be matched to an order item.
fee_or_promotion_misrouted A fee or promotion line failed because of its account mapping.
blocked_by_flow_error A reconciliation error from an integration flow blocked the transaction.
order_not_imported The order behind the transaction wasn't imported.
amazon_reserve_carryover_adjustment An Amazon reserve, carryover, or adjustment line has no linked NetSuite transaction.
unknown_fee_misclassified A miscellaneous or unknown fee.
partial_payment_application The payment was only partially applied.
missing_upstream_transaction No linked NetSuite transaction was found.
month_split_boundary No line matched, and the posted dates fall in different months more than one day apart.
posted_date_timezone_drift No line matched, and the posted dates fall in different months but at most one day apart.
unknown None of the rules matched.

When the root cause explains less than 70% of the variance, the dashboard marks the investigation as having mixed causes, and the explanation also covers the secondary causes.

Settlement flow steps

The flow builder and Run console show these steps:

  • Get flagged Amazon settlement transactions from NetSuite
  • Get every flagged transaction of the Amazon settlement from NetSuite
  • Get prior Amazon settlement investigations from NetSuite
  • Investigate Amazon settlement variance (AI agent)
  • Post Amazon settlement investigation to NetSuite

Find AI assist in NetSuite

After the bundle is updated, NetSuite shows a Celigo AI assist tab. The tab is shared with other Celigo integration apps, such as Shopify - NetSuite; the Amazon category holds the Amazon dashboard and records. If your NetSuite account has many tabs, the Celigo AI assist tab is under the more options (...) menu in the NetSuite navigation bar.

Page What it shows
Dashboard The Celigo AI assist — Amazon dashboard, with the Settlements and Inventory forecast views.
Inventory exceptions The list of inventory exception records, one per flagged SKU.
Replenishment snapshots The list of replenishment snapshot and forecast records.
Triage runs The list of triage and forecast run records.

Use the Celigo AI assist dashboard

The Celigo AI assist - Amazon dashboard is where you review forecasts, inventory exceptions, and settlement investigations. Select Settlements or Inventory forecast at the top to switch views.

Dashboard controls

  • All stores: Shows results for all stores, or select one Amazon store.
  • Time window: Last 7 days, Last 30 days, Last 90 days, or Last year.
  • Refresh: Reloads saved results. Refresh doesn't run either flow.
  • Auto-refresh every 5 min: Reloads saved results every five minutes when selected.
  • Back to NetSuite: Returns to the NetSuite home page.

A note at the top right shows when the results were last created, for example, Last forecast: Oct 12, 11:20 AM. Refresh reloads saved results. When results are older (24 hours for settlement investigations, 7 days for forecasts), the note tells you to run the flow in the integration app. Times use your browser's time zone.

The dashboard displays up to 1,000 exceptions, 500 snapshots, 60 runs, 500 investigations, 2,000 forecasts, and 60 forecast runs. When it shows a limited set of records, counts reflect only the loaded records; narrow the time window or use NetSuite saved searches for the complete list.

Inventory forecast view

The Inventory forecast view shows which SKUs will run out, which upcoming sales events put them at risk, and which FBA SKUs need attention now. It has five tabs:

Tab What it shows
Event risk by SKU The latest forecast for each SKU, with the High event risk, Medium event risk, Stockout risk, Excess stock, Demand rising, and Last forecast metrics. Filter by High event risk, Medium event risk, Sell-down opportunity, Stockout risk, Excess stock, or Rising demand. The Upcoming events panel shows the events the AI agent found for the next 90 days, with sources.
Forecast runs Each forecast run, with SKU counts by risk, the AI summary, and the upcoming events.
Inventory exceptions FBA inventory exceptions with severity, likely cause, status, quantities, marketplace, and last seen time. Filter by Open, Needs review, Resolved, All, High, Medium, Low, and cause.
Replenishment snapshots The FBA quantities captured on each run: fulfillable, unfulfillable, inbound (working, shipped, receiving), reserved, and total.
Triage runs How many SKUs each run scanned and flagged.

Expand a SKU on the Event risk by SKU tab to see the event risk callout, the stock forecast chart (actual stock, forecast, best and worst case, upcoming events, and the runout date), and details such as the sales rate, days of cover, runout date, and Amazon's recommendation. Select NetSuite record to open the forecast record.

On the Inventory exceptions tab, the metrics change to Open exceptions, High severity, Needs human review, Unfulfillable units, Resolved in period, and Last triage run, and the charts show Open exceptions by likely cause and Open exceptions by marketplace. Expand an exception to see the AI explanation, recommended checks, severity rationale, and details, with the Open in Seller Central and Open in NetSuite links.

Settlements view

The Settlements view, titled Settlement operations, shows which Amazon settlements didn't reconcile, why, and the order to fix them in. The metrics are Settlement investigations, Settlements affected, Variance flagged, Recurring causes, Mixed causes, and Last diagnosis.

The Settlement investigations tab lists each investigation with the Settlement, Root cause, AI explanation, Variance, Account, and Outcome columns. A Recurring badge marks a cause that also occurred in an earlier settlement. Search by settlement, cause, account, or order, or filter by settlement.

Expand an investigation to see:

  • Where to fix: The steps to fix the settlement, with links to the right place in the integration app, for example, Import the missing Amazon orders on demand, Review the settlement flows and their errors, Review the settlement settings, Check marketplace and currency settings, and Review the settlement summary in NetSuite.
  • AI explanation: Why the settlement doesn't reconcile.
  • Amazon fees explained: For each fee, what it is, why Amazon charges it, what to do, and the sources.
  • Resolution sequence: The steps to check, in order.
  • Details: Settlement, Marketplace, Diagnosed, Evidence citations, Recurring, and Share of variance explained.
  • Impacted transactions: Each settlement transaction with its Amazon order, type, reconciliation status, error, variance, the amount explained by the root cause, and the NetSuite transaction.

To import missing orders, select Copy order IDs, then select Open Order (FBA) settings or Open Order (MFN) settings. On the Orders tab, paste the IDs into Amazon order Id(s), select Save, and then run the settlement flow again. The on-demand import takes 50 IDs at a time, so only the first 50 IDs are copied. For more about the order flows, see Sync FBA orders and customers from Amazon to NetSuite and Sync MFN and SFP orders to NetSuite as sales orders.

Review AI assist records in NetSuite

AI assist saves its results as NetSuite records. Open them from the dashboard links or from the Celigo AI assist tab. Status is Open or Resolved, and Severity is High, Medium, or Low.

Inventory exception record

The Celigo Amazon AI exception record holds one record per flagged SKU, and the flow updates it on later runs. It includes the SKU, ASIN, Marketplace, Status, Severity, Likely cause, Confidence, AI explanation, Recommended checks, Severity rationale, the Fulfillable, Unfulfillable, Inbound, Reserved, and Total quantities, the Seller Central listing URL, and First seen, Last seen, and Resolved on dates.

To mark an exception resolved:

  1. Open the Celigo Amazon AI exception record in NetSuite.
  2. Select Edit.
  3. Change Status to Resolved.
  4. Select Save.

Replenishment snapshot record

The replenishment snapshot record holds the FBA quantities captured on each run, plus the forecast for each FBA and MFN SKU: product group, units sold (30 days), units per day with slow and fast cases, days of cover, demand trend, stock status, stockout date with worst and best cases, event, event start, event risk, callout, and, for FBA SKUs, excess units and Amazon's recommended action, ship-in quantity, and ship-in date.

Digest run record

The Celigo Amazon AI digest run record holds one record per triage run or forecast run. A forecast run includes the number of SKUs scanned, counts by event risk and stock status, the top event, the AI summary, and the researched events.

Settlement investigation record

The settlement investigation record holds one record per settlement and root cause, for example, SVI 1234567 order_not_imported. It includes the Settlement ID, Settlement summary, Root cause, AI explanation, Fee explanations, Impacted transactions (the top 50 by amount), Resolution steps, Confidence (the share of the variance the root cause explains), Citations, Variance, Recurrence, Amazon account, Marketplace, Store, Integration app link, and Outcome.

After you review an investigation, set its Outcome to Accepted or Rejected. The flow keeps your Outcome on later runs. When a later run finds the same root cause for an investigation you rejected, the AI explanation notes that the earlier diagnosis was rejected.

AI assist limits

These limits apply to both flows and the dashboard.

  • AI assist results are advisory. Verify the source records before you retry flows or make changes.
  • Sales-event research runs once per store, per marketplace, per day for FBA and once for MFN. Later runs that day reuse the results, unless the FBA report includes a product group that wasn't researched yet.
  • Each flagged FBA SKU and each investigated settlement uses one AI action per run.
  • The FBA quantity fields update existing NetSuite items only. AI assist doesn't create items.
  • Exceptions cover FBA SKUs only. Forecasts cover FBA SKUs and the store's Amazon-linked NetSuite items for MFN.
  • The settlement flow investigates settlements only within the variance tolerance and lookback period you set, and skips settlement transactions without a settlement summary.
  • A settlement investigation lists up to 50 impacted transactions, by amount.
  • The settlement agent researches fees only in Amazon Seller Central and Amazon developer documentation.
  • The dashboard shows a limited set of records. Counts reflect only the loaded records.
  • The AI agent steps' instructions, model, and output format can't be edited.

Troubleshoot AI assist

Use these causes and resolutions when AI assist doesn't install, a setting doesn't save, a flow returns no results, or the dashboard doesn't load.

Celigo AI assist tab missing in NetSuite

Symptom: The Celigo AI assist tab doesn't appear in the NetSuite navigation bar.

Cause: The NetSuite integration bundle isn't updated, or the tab is under the more options menu.

Resolution: Check the more options (...) menu in the NetSuite navigation bar. In a sandbox account, update the NetSuite integration bundle. In a production account, contact Celigo Support if the bundle hasn't been updated. See NetSuite integration bundle update.

AI assist flows don't appear

Symptom: AI assist isn't listed on the Flows tab of the selected store.

Cause: The AI assist add-on isn't installed for the selected store, for example, a store you added to an existing integration app.

Resolution: Install the add-on from the Add-ons tab. See Install the AI assist add-on.

Bundle is missing AI assist objects

Symptom: The installation stops with, for example, Celigo AI assist needs the latest Celigo Amazon - NetSuite bundle (ID …). These NetSuite objects are missing: … Update the bundle in NetSuite, then install the add-on again.

Resolution: Update the Celigo Amazon - NetSuite bundle in NetSuite, and then install the add-on again. See Install the AI assist add-on.

The installation can't check the bundle

Symptom: The message reads, for example, Could not check the Celigo Amazon - NetSuite bundle…

Cause: The NetSuite connection couldn't be reached, or the check failed.

Resolution: Test the NetSuite connection, and then install the add-on again. If the bundle check still fails, see Resolve error when verifying Amazon bundle.

A setting doesn't save

Symptom: Saving the AI assist settings shows one of these messages:

  • The Scenario range for forecast (%) value isn't valid. Enter a number from 0 to 100 in Settings > AI assist.
  • The fulfillable quantity threshold isn't valid. Enter a number of 0 or greater in Settings > AI assist.
  • The NetSuite item field ID for Amazon SKU isn't valid. Enter a valid NetSuite item field ID, such as itemid, in Settings > AI assist.
  • The Ignore reconciliation variances at or below value isn't valid. Enter a number of 0 or greater in Settings > AI assist.
  • The Investigate settlements that ended in the last (days) value isn't valid. Enter a whole number from 0 to 90 in Settings > AI assist.

Resolution: Correct the value as the message describes, and then select Save.

A setting can't be updated

Symptom: The message reads Missing hook. Contact Celigo Support

Cause: The export that the setting updates was changed.

Resolution: Contact Celigo Support.

A SKU has no matching NetSuite item

Symptom: The forecast flow's run shows the NetSuite error cannot "update" record without internalid for some SKUs.

Cause: The Amazon SKU has no matching NetSuite item. The flow updates existing items only, so this error is expected for SKUs that aren't in NetSuite.

Resolution: If the SKU should exist in NetSuite, create the item, or check that NetSuite item field ID for Amazon SKU points to the item field that holds the Amazon SKU. Otherwise, you can ignore the error.

Item FBA quantity fields aren't updated

Symptom: The FBA quantity fields on NetSuite items stay empty or don't change after a run.

Cause: NetSuite item field ID for Amazon SKU doesn't point to the item field that holds the Amazon SKU, or the forecast flow is turned off.

Resolution: Correct the setting in Settings > AI assist, make sure the forecast flow is turned on and scheduled, and then run it.

Too many or too few FBA exceptions

Symptom: The Inventory exceptions tab lists more or fewer SKUs than you expect.

Cause: The Fulfillable quantity threshold (units) setting is too high or too low. The default of 0 flags only stockouts.

Resolution: Adjust the threshold in Settings > AI assist.

The settlement flow finds no settlements

Symptom: The Settlements view shows No settlement investigations in this window. Run the Amazon settlement variance flow from the integration app.

Cause: The flow is turned off or hasn't run, the variance tolerance is too high, or the lookback period is too short to include older settlements.

Resolution: Turn on and run the settlement flow, check Ignore reconciliation variances at or below (amount) and Investigate settlements that ended in the last (days), and widen the dashboard time window.

The forecast has no upcoming events

Symptom: The forecast shows The agent found no upcoming events for this forecast.

Cause: The AI agent found no events in the next 90 days for your product groups, or the research step failed.

Resolution: Check the Research upcoming Amazon sales events step in the flow's run history for errors. Events are researched once per store per day.

Dashboard shows no results

Symptom: The dashboard shows, for example, No inventory forecast in this window. Run the Amazon inventory forecast flow from the integration app.

Cause: The flows haven't run, or no results exist in the selected time window or store.

Resolution: Make sure both flows are turned on and scheduled and have run, widen the time window or select All stores, and then select Refresh.

The dashboard couldn't refresh or load

Symptom: The dashboard shows Couldn't refresh data. Showing the previous <n>-day snapshot. Sign in to NetSuite again if your session expired., Couldn't load the <tab name> tab, or The Celigo AI assist dashboard couldn't load (<error name>).

Resolution: Sign in to NetSuite again if your session expired, and then select Refresh. If a tab keeps failing, check that your NetSuite role can view the AI assist records. If the dashboard can't load, check the script execution log for details.

Related articles