Dispatches
FIG. 19 AI enterprise search → FIG. 18 Athena Python SDK → FIG. 17 Athena Skills → FIG. 16 Web Monitor → FIG. 15 Athena for Operations → FIG. 14 Athena for Analytics → FIG. 13 AI search for your DMS → FIG. 12 Multiplayer AI agents → FIG. 11 Athena Anywhere → FIG. 10 Hey Athena → FIG. 09 Athena for Finance → FIG. 08 What code review becomes → FIG. 07 GTM without a CRM → FIG. 06 The company brain → FIG. 05 Two identical stacks → FIG. 04 Finance builds its own APIs → FIG. 03 Model routing in production → FIG. 02 Knowledge workers don't have GitHub → FIG. 01 Pricing, read from transcripts →
Dispatch 2026-10-01 Solutions
№15

Athena for Operations: AI Agents for Supply Chain Teams

Make your operations team agent native with Athena. From a dashboard alert to an approved supplier email, Athena cleans the data, defines each metric once, answers in Teams with sources and drafts the fix for a person to approve.

Filed by Taylor Johnson — 11 min
Sepia engraved plate of an ancient harbor: a stone quay with a stack of clay amphorae, sacks of grain, a papyrus ledger with a wax tablet, stylus and olive branch, a galley with one sail at the pier, a lighthouse, and a hillside temple above a calm sea

Athena for Operations is a set of AI agents for supply chain teams, built by Athena Intelligence. Athena cleans the source files, defines each metric once in a semantic model, answers questions in Microsoft Teams with cited sources, and drafts the fix, such as a supplier email, for a person to approve.

A supply chain team already has plenty of data. The hours go to the step after the alert: working out why a number moved, finding who owns the supplier, and writing the email. Athena for Operations takes that work on, using the systems the team already runs.

This is the same platform Athena Intelligence runs for finance and other functions, pointed at the work operations teams repeat every week. Athena Intelligence, founded in 2022, builds an agent platform for regulated enterprises that can be self-hosted, where people and agents build agents, workflows and applications together. This dispatch follows one stockout from a dashboard alert to an approved supplier email. Figure 1 shows the whole idea on one page.

  • Athena for Operations covers the whole path from messy files to a sent fix: one clean table, shared metric definitions, a rebuilt dashboard, sourced answers in Teams and a drafted action.
  • The launch video follows one stockout from start to finish. Every name and number in it is demo data.
  • Every number links to its source, every change is attributed, every version can be recovered, and nothing goes out until a person approves it.
  • Athena connects to Snowflake and Databricks using each person's own access, and can be deployed in your VPC.
Contents
  1. 01The problem
  2. 02What Athena can do
  3. 03One example, start to finish
  4. 04Trusting the agents
  5. 05How this differs from other tools
  6. 06Security and deployment
  7. 07Getting started
  8. 08Questions teams ask
FIG. 01 / FROM MESSY FILES TO AN APPROVED FIXYour messy inputsFiles, mail, PDFs and warehousesSpreadsheetsSupplier emailsPDFs and packing listsWarehouse exportsSnowflakeDatabricksAthenaDaily stockout checkReads days of supply on a scheduleCarrier reportsWeekly files, your own scriptContract trackingTerms and renewals, citedSupplier follow-upDrafts the email for approvalYour next workflowA clean inventory tableOne SQL table, consistent unitsA rebuilt dashboardOn shared metric definitionsA cited answer in TeamsSources open from the replyA drafted supplier emailWaits for a person to sendACitedEvery number links to its sourceBReversibleEvery change can be undoneCYour permissionsAccess follows each personDHuman approvalNothing goes out unapproved FIG. 01 / MESSY FILES TO AN APPROVED FIXYour messy inputsSpreadsheetsSupplier emailsPDFs and packing listsWarehouse exportsSnowflakeDatabricksAthenaDaily stockout checkReads days of supply on a scheduleCarrier reportsWeekly files, your own scriptContract trackingTerms and renewals, citedSupplier follow-upDrafts the email for approvalYour next workflowA clean inventory tableOne SQL table, consistent unitsA rebuilt dashboardOn shared metric definitionsA cited answer in TeamsSources open from the replyA drafted supplier emailWaits for a person to sendACitedEvery number links to its sourceBReversibleEvery change can be undoneCYour permissionsAccess follows each personDHuman approvalNothing goes out unapproved
FIG. 01Athena for Operations takes messy inputs on the left, runs a workflow for each recurring job in the middle and returns finished outputs on the right. The same four protections (A to D) apply to every workflow.

01What problem does Athena solve for supply chain teams?

Athena closes the manual gap between seeing a supply problem and acting on it. Teams wait on refreshes, reconcile exports and hunt for owners, and the stockout often arrives before the answer does.

The video opens on that week. On Monday morning a VP of supply chain asks why the team is short on one SKU at a distribution center. The inventory dashboard is refreshing, with 14 minutes left. The next stop is tab 7 of an 11-tab warehouse export, full of mismatched names and #VALUE! errors. By early afternoon the thread is a chain of forwards asking who owns the supplier account. On Thursday at 7:10 AM the SKU is out of stock and three customer orders are on hold. Figure 2 sets that week beside the same week with Athena.

FIG. 02 / THE SAME WEEK, TWO WAYSTHE MANUAL WAYMON AMDashboard is refreshing14 minutes leftMON 10:30 AMDig through the exportsTab 7 of 11, #VALUE! errorsMON PMWho owns the supplier?A chain of forwardsTHU 7:10 AMVK-220 is out of stock3 customer orders on holdWITH ATHENADAILY CHECKShortfall flaggedDays of supply read on a scheduleMON 9:40 AMAsk why, in TeamsA sourced answer in the channelMON AMFix draftedSupplier email waits for approvalAPPROVEDA person clicks SendSupplier asked to ship 400 unitsDemo scenario. All names and numbers are demo data. FIG. 02 / THE SAME WEEK, TWO WAYSTHE MANUAL WAYMON AMDashboard is refreshing14 minutes leftMON 10:30 AMDig through the exportsTab 7 of 11, #VALUE! errorsMON PMWho owns the supplier?A chain of forwardsTHU 7:10 AMVK-220 is out of stock3 customer orders on holdWITH ATHENADAILY CHECKShortfall flaggedDays of supply read on a scheduleMON 9:40 AMAsk why, in TeamsA sourced answer in the channelMON AMFix draftedSupplier email waits for approvalAPPROVEDA person clicks SendSupplier asked to ship 400 unitsDemo scenario. All names and numbers are demo data.
FIG. 02The same shortfall handled two ways. The manual path in the video ends with an out-of-stock notice on Thursday at 7:10 AM. The Athena path ends with a drafted supplier email waiting for approval. All names and numbers are demo data.

Two surveys describe the gap in the wider industry. In ABI Research's 2025 supply chain survey, 71% of respondents ranked the lack of clearly defined standard operating procedures among their top three blockers to acting on real-time supply chain data. An agent operating procedure (AOP) is that procedure written down: Athena follows it on a schedule or when something triggers it, the way a new hire follows an SOP. The AOPs page covers how they work.

Impinj surveyed 1,000 US supply chain professionals about visibility, and 91% believed their organization is equipped to drive accurate supply chain visibility, while 33% consistently obtain accurate, real-time inventory data.

02What can Athena do for operations teams?

Athena runs the recurring work of an operations team: inventory watches, carrier reports, supplier follow-ups, contract tracking and the dashboards behind them. The examples below are grouped by sub-team, and Figure 3 lays them out as a grid. The supply chain and operations solution page covers more.

Inventory and planning

  • A daily check reads days of supply by SKU and distribution center, finds the cause of any shortfall and drafts the fix.
  • Inventory positions from spreadsheets, supplier emails, packing lists and warehouse exports land in one table with consistent columns and units, the job in the CSV to SQL database use case.
  • Planners ask questions of an inventory analysis app built on Snowflake data, and each answer shows its source.

Logistics and carriers

  • Weekly carrier files arrive in a shared inbox, the team's own Python script runs on them unchanged, and the output template fills in.
  • Control tower views are built from the same cleaned data.
  • Athena replaces legacy RPA that broke when a vendor moved a button. Athena pauses on exceptions and asks before writing anything back.

Procurement and suppliers

  • Each new contract is read, and the parties, term, value and renewal date are extracted with a citation to the page. Friday's deck builds from the results.
  • NDA and approval routing tracks the status of every signer.
  • Purchase order steps, specs and pricing tables are pulled out of vendor PDFs.
  • When a ship date moves, Athena drafts the follow-up email to the supplier.

Data and analytics for operations

FIG. 03 / WORKFLOWS FOR AN OPERATIONS TEAMINVENTORY AND PLANNINGDaily stockout checkInventory cleanup into onetablePlanner Q&A onSnowflake dataLOGISTICS AND CARRIERSWeekly carrier file runControl tower viewsRPA replacement withapprovalsPROCUREMENT AND SUPPLIERSContract term extractionNDA and approval routingPO and spec extractionShip-date follow-up emailDATA AND ANALYTICSDashboard on a sharedmodelAsk why, in TeamsWarehouse permissionskeptYour next workflow, built by your team FIG. 03 / WORKFLOWS FOR OPERATIONSINVENTORY AND PLANNINGDaily stockout checkInventory cleanup into one tablePlanner Q&A on Snowflake dataLOGISTICS AND CARRIERSWeekly carrier file runControl tower viewsRPA replacement with approvalsPROCUREMENT AND SUPPLIERSContract term extractionNDA and approval routingPO and spec extractionShip-date follow-up emailDATA AND ANALYTICSDashboard on a shared modelAsk why, in TeamsWarehouse permissions keptYour next workflow, built by your team
FIG. 03Four groups of workflows for an operations team: inventory and planning, logistics and carriers, procurement and suppliers, and data and analytics. A dashed tile marks the next workflow a team builds for itself.

03How does Athena for Operations work? One example, start to finish

The video walks through six steps: fix the data, connect the warehouse, define the metrics, rebuild the dashboard, answer questions in Teams and draft the fix. Figure 4 shows the flow, with a checkpoint before the last step where a person approves. The video runs about 48 seconds with music and no voiceover, so the on-screen text carries the message.

VIDEOThe launch video (48 seconds): a dashboard flags a shortfall, then Athena cleans the data, defines the metrics once, rebuilds the dashboard, answers why in Teams with sources and drafts the supplier email for a person to send. The video has music and on-screen text and no narration. All names and numbers are demo data.

In the demo, a dashboard flags a shortfall: days of supply for VK-220 at DC East reads 6.1, and a supplier notice says PO 4471 has moved to Oct 6. Athena cleans the data behind it, defines the metrics once, rebuilds the dashboard, answers why in Teams with sources, and drafts the supplier email for a person to send. All names and numbers are demo data.

  1. Fix the data. An inventory cleanup chat takes a warehouse spreadsheet, a supplier email, a packing list PDF and a warehouse export, runs Python, and writes one SQL table of 1,284 rows with SKU, distribution center, on-hand quantity and unit.
  2. Connect what you have. Snowflake and Databricks connect with a setting that requires each user's own connection, so people see what their own access allows.
  3. Define your metrics once. A semantic model drafted from a prompt holds on-hand units, days of supply and the join to daily demand. In the demo, the data team changes the demand window from a 28-day average to a 56-day average and clicks Deploy. The semantic model generation use case describes the pattern.
  4. Rebuild the dashboard. The inventory health dashboard recalculates on those metrics, and days of supply for the demo SKU updates on screen beside a bar chart across distribution centers.
  5. Ask anywhere. The VP of supply chain asks Athena in Teams why days of supply dropped to about six. Athena answers in the channel and cites the semantic model measures behind the answer: the supplier moved a purchase order ship date from Sep 24 to Oct 6, and demand at the distribution center is up 18% over four weeks.
  6. Act on it. A daily check finds the cause and drafts an email to the supplier asking whether 400 units can ship by Oct 1. A person clicks Send Email.
FIG. 04 / THE VIDEO, STEP BY STEP01Fix the dataAn inventory cleanupchat writes one cleanSQL tableONE TABLE02Connect whatyou haveSnowflake andDatabricks, eachperson on their ownaccessYOUR WAREHOUSE03Define yourmetricsA semantic modelholds days of supply,defined onceONE DEFINITION04Rebuild thedashboardThe dashboardrecalculates on thosemetricsLIVE DASHBOARD05Ask anywhereAsk why in Teamsand get a sourcedanswerCITED ANSWER06Act on itA daily check draftsthe supplier emailDRAFT EMAILA PERSON APPROVESAthena does the collecting, defining, rebuilding and drafting.A person decides, and nothing goes out until they click Send. FIG. 04 / THE VIDEO, STEP BY STEP01Fix the dataAn inventory cleanup chat writes one clean SQL tableONE TABLE02Connect what you haveSnowflake and Databricks, each person on their own accessYOUR WAREHOUSE03Define your metricsA semantic model holds days of supply, defined onceONE DEFINITION04Rebuild the dashboardThe dashboard recalculates on those metricsLIVE DASHBOARD05Ask anywhereAsk why in Teams and get a sourced answerCITED ANSWERA PERSON APPROVES06Act on itA daily check drafts the supplier emailDRAFT EMAILAthena does the collecting, defining, rebuilding and drafting.A person decides, and nothing goes out until they click Send.
FIG. 04The six steps from the video, with a human checkpoint before the email is sent. Athena does the collecting, defining, rebuilding and drafting, and a person approves the action.

Table 1 compares the manual week with the Athena week, task by task.

TaskThe manual wayWith Athena
Find out there is a problemA dashboard number someone happens to see, after a refreshA daily check reads days of supply on a schedule and flags the shortfall
Get one clean viewTab 7 of an 11-tab export with mismatched names and #VALUE! errorsFiles, emails and PDFs land in one SQL table with consistent columns
Agree on the definitionEach report calculates days of supply its own wayA semantic model defines the metric once, and every dashboard reads from it
Work out whyA chain of forwards asking who owns the accountAsk in Teams, and Athena answers with cited sources
Fix itSomeone writes the supplier email when they find timeAthena drafts the email, and a person approves and sends it
Show the work laterRebuild the story from an inboxEvery number traceable, every change attributed, every version recoverable
TABLE 1The same six tasks done by hand and with Athena. The right-hand column describes what happens in the demo.

04How do people trust what the agents do?

Every number Athena shows links to its source, every change is attributed to whoever or whatever made it, every version can be recovered, and nothing goes out until a person approves it. Figure 5 shows how that looks in the product.

The Teams answer in the video carries its sources with it, so the person who asked can open the measure behind each claim. Human approval can be required on any tool, which is why the supplier email waits for a click. A long run of agent actions can be rolled back together, and Athena flags the actions that cannot be reversed. Athena does not store data from connected sources, and revoking someone's access takes effect in real time. The audit trail and rollback use case goes into detail.

FIG. 05 / TRUST, SHOWN IN THE PRODUCTCITED ANSWERDays of supply fell because thesupplier moved the PO shipdate.days_of_supplySOURCESemantic modeldays_of_supply = on_hand/ daily_demandACTIVITYDana edited demand windowDeployedRoll backAthena rebuilt the dashboardAttributedRoll backAthena drafted the emailWaitingRoll backEvery version can be recoveredPERMISSIONSDana's own accessInventoryReadPurchase ordersReadSupplier contractsReadPayrollNo accessRevoked access ends in real timeAPPROVALSend this email?To: Keller ComponentsCan 400 units of VK-220 ship byOct 1?EditApprove and sendPanels redrawn from the demo. All names and numbers are demo data. FIG. 05 / TRUST, SHOWN IN THE PRODUCTCITED ANSWERDays of supply fell because the supplier moved the PO ship date.days_of_supplyopen_purchase_ordersSOURCESemantic modeldays_of_supply = on_hand / daily_demandACTIVITYDana edited demand windowDeployedRoll backAthena rebuilt the dashboardAttributedRoll backAthena drafted the emailWaitingRoll backEvery version can be recoveredPERMISSIONSDana's own accessInventoryReadPurchase ordersReadSupplier contractsReadPayrollNo accessRevoked access ends in real timeAPPROVALSend this email?To: Keller ComponentsCan 400 units of VK-220 ship by Oct 1?EditApprove and sendPanels redrawn from the demo. All names and numbers are demo data.
FIG. 05Four views of trust in the product: a source popover on a cited answer, an activity list with a rollback button beside each change, the permissions a person's own access grants, and an approval prompt before a send.

05How is Athena different from other AI tools for supply chain teams?

Athena is one platform for data cleanup, metric definitions, dashboards, questions and actions, so the answer in Teams and the number on the dashboard come from the same definition. Many AI tools for inventory and logistics sit inside a single system and watch one set of numbers.

Athena works on the systems the team already runs. Athena reads a shared inbox, runs the team's existing Python script unchanged, opens spreadsheets, and connects to more than 150 data connectors, including Snowflake and Databricks. The connectors page lists the systems Athena reaches.

Every answer is cited and every change can be undone, which matters when a supply chain decision ends in a purchase order or a customer promise. Athena runs in the customer's environment, in your VPC, on-premises or air-gapped, and works with more than 60 models from every frontier provider.

The team builds the next workflow itself. Athena drafts a semantic model from a prompt, and the data team refines it. An operations lead who already has a written procedure can hand it over as an AOP.

06Is Athena secure enough for regulated companies?

Yes. Athena Intelligence's security program includes SOC 2 Type II and HIPAA, third-party penetration tests and zero data retention agreements with Anthropic, OpenAI and Gemini, and Athena does not train on customer data. Figure 6 shows where the boundary sits.

Athena deploys in a managed cloud on AWS, in your own VPC on AWS, Azure or GCP, on-premises, or air-gapped, including US government air-gapped networks and AWS GovCloud. A global consulting firm has run Athena in production for over a year, with over 1,000 users across five localities. The deploy anywhere and permissions and identity pages go into detail.

FIG. 06 / THE SECURITY AND DEPLOYMENT BOUNDARYYOUR VPCON-PREMISESAIR-GAPPEDYour dataStays where it livesAthena does not store data fromconnected sourcesAthenaAgents, workflows, appsRuns inside the same boundaryModelsMore than 60, any providerZero data retention agreementsSOC 2 Type IIAudited controlsHIPAACompliantZero data retentionWith model providersPenetration testsBy third parties FIG. 06 / SECURITY AND DEPLOYMENTYOUR VPCON-PREMISESAIR-GAPPEDYour dataStays where it livesAthena does not store data from connected sourcesAthenaAgents, workflows, appsRuns inside the same boundaryModelsMore than 60, any providerZero data retention agreementsSOC 2 Type IIAudited controlsHIPAACompliantZero data retentionWith model providersPenetration testsBy third parties
FIG. 06Data, Athena and the models sit inside the customer's boundary, whether that is a VPC, an on-premises data center or an air-gapped network, with compliance controls shown as a band underneath.

07How do we get started?

Start with the alert your team already chases every week, such as a days-of-supply drop on one SKU. Connect one warehouse, define the metric once, and ask why in Teams. A check that finds the cause and drafts the supplier email is a good first agent operating procedure to save.

Curious how it would work on your data? We're happy to show you.

08Questions teams ask

What is an AI agent for supply chain operations?

An AI agent for supply chain operations is software that does a recurring piece of operations work instead of only reporting on it. It reads inventory, order and supplier data, works out why a number moved, and drafts the next step, such as a supplier email. In Athena, a person approves anything that goes out.

What is days of supply?

Days of supply is the number of days current inventory will last at the recent rate of demand. A planner divides on-hand units by average daily demand, so the result depends on which demand window the team uses. A semantic model in Athena defines that window once, so every dashboard reports the same figure.

What is a semantic model?

A semantic model is a shared, written definition of the metrics and dimensions a business uses, such as on-hand units, days of supply, SKU and distribution center. Dashboards and questions read from the model instead of calculating their own versions. Athena drafts one from a prompt, and the data team refines it.

Does Athena replace our dashboards or BI tools?

Athena can run next to the tools a team already uses, and rebuild a dashboard on a shared semantic model when the team is ready. Teams keep their existing reports while they compare results. Consolidating onto Athena is a choice the team makes over time.

Does Athena connect to Snowflake and Databricks?

Yes. Athena connects to Snowflake and Databricks, and an administrator can require each user to connect with their own credentials, so people see only what their own access allows. Athena also reads files, PDFs, emails and existing spreadsheet or Python work from the systems where they already live.

Who approves what goes out?

A person does. Athena drafts the supplier email, the report or the change, and a team member reviews it and clicks send. Human approval can be required on any tool, every action is attributed, and Athena flags actions that cannot be reversed. Nothing reaches a supplier or customer until someone on your team approves it.

Where does Athena run?

Athena runs in a managed cloud, in your own VPC on AWS, Azure or GCP, on-premises, or air-gapped, including US government air-gapped networks and AWS GovCloud. Athena is model agnostic, with more than 60 models from every frontier provider and zero data retention agreements with Anthropic, OpenAI and Gemini.

Athena Library

Run this on a schedule

Run supply chain briefs like these on a schedule, delivered as a spreadsheet, deck and document after every run.

№14 Athena for Analytics Ask in plain English in Athena or Teams and get answers from your semantic model, with the query behind every one. next dispatch

All dispatches → Work on this →