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 →
Customer story

Just-in-time access

A vendor replaced standing Kubernetes credentials with audited, time-bounded access that expires automatically and logs twice: once by the platform, once by the cluster.

46 GBOne production service before right-sizingStanding vendor credentials into the clusterVendor engineerProduction cluster accessSource-verified accessKubernetes RBACStanding credentialsAlways-on accessgrantedEvery security reviewflagged standingvendor credentialsBefore: standing credentials, always onVendor engineersStanding credentialsCUSTOMER PRODUCTION KUBERNETES CLUSTERAgents monitor andremediateSelf-healing firstHuman escalationonly when agentcannotTime-boundedaccessLogged by platformand clusterAgents run inside the customer's own cluster; humans request inAgents inside; humans request when needed01IncidentdetectedAgent attemptsself-remediation02Agent escalationHuman access required03Vendor requestsJIT accessTime-bounded, perincidenta person approves04Access expiresautomaticallyLogged by platformand clusterCustomer approves each request; double audit trailJust-in-time access, approved per incidentStanding accessJIT audited access46 GB servicePermanent credentialsSecurity review flagsVendor credentials always live3 to 5 GB right-sizedZero standing credentialsLogged twiceFail-closed; adds minutes to incidentZero standing credentials; right-sized along the wayFail-closed. Audited twice.athenaintel.com
Fig. 02The story in 22 seconds, from the old way to “Fail-closed. Audited twice.”
Before · what we did · after
  1. Before

    A vendor with standing credentials into the production cluster. Every security review asked about it.

  2. What we did

    Replaced standing access with just-in-time, time-bounded access requested per incident, approved by the customer, expiring automatically, logged twice: once by the platform, once by the cluster.

  3. After

    Zero standing vendor credentials. Agents monitor and remediate inside the customer's own cluster; humans request access only when the agent cannot. Right-sizing along the way took one service from 46 GB to 3 to 5 GB.

Athena wants toAccess production Kubernetes cluster forincident responseTime-boundedLogged by platform and clusterAgent escalation: self-remediation failedWAITING FOR A PERSONAccess granted; expires automaticallyRequested by vendor · approved by customer · 14:23
Fig. 01Each incident triggers a fresh access request, approved by the customer
How long it takes · what to plan for

JIT adds minutes to an incident response. We think that trade is right for regulated environments.

Fail-closed. Audited twice.

Example workflow

From standing vendor credentials to JIT audited access

SURFACES INWHAT COMES INTHE PLATFORMWHAT GOES OUTDELIVERED BACKWebMobile · voiceText (SMS)EmailSlack · TeamsChromeOffice add-inMeeting botMeetingsWordPowerPointSheetsFiles · PDFsDataAppsLibrarySessionsSpacesStudiosMeetingsLakehouseSemantic modelsDatabaseWordPowerPointSheetsDashboardsData · APIsApps · portalsWebMobile · voiceText (SMS)EmailSlack · TeamsChromeOffice add-inMeeting botAgents1Workflows (AOPs)2Workflows (AOPs)3Dashboards41 RUNAgent monitors and remediatesinside the customer's owncluster; escalates whenself-remediation fails2 REQUESTVendor requeststime-bounded access perincident; platform logsthe request3 APPROVECustomer approves; accessgranted for fixed window,expires automatically4 CHECKAccess logged twice: onceby platform, once bycluster; zero standingcredentials remain
Fig. 03The route on the system map: Agents → Workflows (AOPs) → Dashboards. Connectors: Kubernetes cluster API, platform audit log, cluster audit log
What it runs on
Agents →Workflows →Permissions & identity →Attribution & reversibility →
What customers say
“He wrote the entire stuff. By itself. And I was there, and I was thinking, alright. If I had to write all this stuff, I mean, I would've spent days, days, days.”
Manufacturing analyst · A global manufacturer
“I send that to the programmer and he said, alright. I read your file. I copied the section of the code because he has the right interface, and it works. Wow. Just like magic.”
Manufacturing analyst · A global manufacturer
“Digital worker, it gets spun up, it gets given access to the toolkits that it needs. It gets given access to the data that it needs. Just exists for the period it needs to exist to perform the task. And then it gets torn back down again afterwards.”
Audit lead · A global professional services firm
“I like that direction, that concept of the agent is the agent. It becomes an entity in its own right with its own ownership and its own missioning, and you get rid of the gray area and the blurry lines.”
Audit lead · A global professional services firm
“A lot of people are telling us that, oh, I sold in this package because of this info... I secured this distribution because of the selling assistance help.”
Sales enablement lead · A global CPG provider
“I am not a programmer. I cannot read a line of program in SQL or anything... The only thing that I'm doing is just writing prompts.”
Manufacturing analyst · A global manufacturer
“I finally got a chance to play around with the computer asset and build out the chatbot you mentioned. It was really impressive.”
Marketing lead · An AmLaw 100 firm
“So we use the SDK, and we write that in Python, and it calls Athena majority by API. The results appear in the interface itself.”
Legal technology team · A global law firm
“You guys are functioning better because the harness is tested for a longer period of time against enterprise customers.”
Analytics lead · A Fortune 500 retailer
“It's like I'm running analysis that would have taken us, like, maybe a week, two. And thirty, forty minutes.”
Sales intelligence analyst · A global CPG provider

Verbatim from customer calls. Customers anonymized.

Related

Related products and stories

Permissions & identity →Your permissions, inherited live. Fail-closed.IT, Security & Identity →Dozens of AI vendors. One governed platform. Your permissions, inherited live.Internal AI assistant →A ten-person pilot became the company's own internal AI assistant, rolling to 10,000 employees, included in Athena services.

See this on your own work

Bring one recurring task, a sample input, and an example of work you would accept.

Get a demo