Agentic AI for Mainframe Operations

Zero-touch abend managementfor mainframe batch.

Progull is the agentic AI layer that detects, diagnoses, remediates and closes mainframe batch incidents end-to-end — natively wired into z/OS, CICS, DB2, JCL and ServiceNow.

92%
Auto-resolved abends
<3m
Mean time to recovery
24/7
Hands-off operations
SDSF · LIVE · 02:14:09
$ progull watch --queue BATCH.PROD --lpar SYSA,SYSB
[00:00] ABEND S0C7 · JOB PAYRUN02 STEP0040 PROC RUNTAX
└─ SYSUDUMP 4,812 lines · PSW 078D1000_8000A4F2
[00:08] ingest → SYSOUT · SYSUDUMP · JCL · DB2 SQLCODE -911
[00:14] classify → data exception · packed-decimal · COBOL
confidence=0.94 signature=match#PAY-S0C7-014
[00:21] diff JCL vs baseline · 3 lines changed in CHGREL.2614
+ //STEP0040 EXEC PGM=PAYCALC,PARM='NEW'
- //STEP0040 EXEC PGM=PAYCALC,PARM='CLEAN'
[00:32] rca → field WS-GROSS-AMT redefined PIC S9(7)V99 → overflow
[00:41] evaluate playbook → pb.s0c7.parm-revert (v4)
policy.match=true · auth.scope=ops.batch · review=auto
[00:55] serviceNow → INC0098231 enriched · CI=PAY.BATCH
[01:08] remediation → resubmit STEP0040 PARM=CLEAN COND=(0,LE)
JES2 → JOB07412 queued · class=A · priority=9
[01:34] validate → STEP0040 RC=0000 · 412,189 rows posted
[01:58] DB2 → tablespace PAY.GROSS quiesced · rebind OK
[02:14] INC0098231 closed · MTTR=02:14 · trace attached → ServiceNow
$ _
Native toz/OSCICSIBM DB2JCLSDSFJES2/JES3ServiceNow
0%
Recurring abends auto-resolved
0%
Mean time to recovery reduction
$0+
Saved per avoided night-call
0/7
Hands-off batch operations
Corporate clients

The roster running on Progull.

Sector-wise view of the operators trusting Progull with their nightly batch. Click any logo to read the engagement story.

Click a client to read the engagement story in our Newsroom.

Live simulator

Pick an abend. Watch Progull resolve it.

Each tab streams a representative reasoning and remediation trace as Progull would handle it inside ServiceNow.

S0C7 · Data exceptionRUNNING
Incident lifecycle

What used to take hours, closed in minutes.

A representative S0C7 abend, handled end-to-end by Progull agents inside ServiceNow.

  1. 00:00
    1

    Abend detected

    S0C7 surfaced in JOB PAYRUN02 within seconds of the JES spool event.

  2. 00:15
    2

    Telemetry ingested

    SYSOUT, SYSUDUMP, JCL and DB2 SQLCODE pulled from SDSF — no human in the loop.

  3. 00:42
    3

    Root cause reasoned

    Correlated to a recent COBOL packed-decimal change and upstream feed file.

  4. 01:30
    4

    Remediation executed

    Approved playbook re-submits STEP0040 with corrected parm; ServiceNow notified.

  5. 02:14
    5

    Incident closed

    INC0098231 auto-resolved with the full reasoning trail attached.

Architecture

One ingestion plane. One reasoning plane. One ticketing plane.

No 3270 screen scrapers, no parallel ticketing — Progull fuses native mainframe telemetry into ServiceNow under your existing controls.

Plane 1

Mainframe sources

  • z/OS SYSLOG
  • SDSF · JES2/3 spool
  • JCL & PROC libraries
  • CICS region state
  • DB2 SQLCODE / dumps
Plane 2

Progull agents

  • Detection · classify abend
  • Reasoning · evidence fusion
  • Policy · playbook match
  • Action · gated execution
  • Closure · audit & KB
Plane 3

Enterprise systems

  • ServiceNow ITSM
  • Change & CMDB
  • SIEM / audit stream
  • Your IdP (SAML/OIDC)
  • On-call & paging
Status quo vs Progull

The same incident. Two operating models.

Stage
Today
With Progull
Detection
Operator paged after threshold breach
Agent observes JES spool in real time
Triage
Manual SDSF browsing across LPARs
Telemetry fused into one reasoning context
Diagnosis
Tribal knowledge in 2–3 SMEs
Explainable RCA against history + change events
Remediation
Free-form commands by ID with OPS authority
Policy-as-code playbooks under surrogate ID
Closure
Worknote written by exhausted on-call
Auto-closed with full reasoning trail attached
MTTR
47 min average · highly variable
< 3 min on the long tail of recurring abends
Knowledge
Stays in heads · lost on attrition
Compounds as approved playbooks & KB articles
ROI calculator

What does an abend cost you today?

Drag the sliders to your environment. Numbers refresh in real time using Progull's measured baselines (92% auto-resolution, 85% MTTR reduction).

200
202,000
45 min
10180
180 /hr
75400
Operator hours reclaimed / month
117 hrs
Direct labour savings / month
$21,114
Abends auto-resolved / month
184
Annualised savings
$253,368

Indicative only. Excludes downstream SLA recovery, deferred change, and avoided escalation cost.

Pricing

Two ways to bring Progull into your estate.

Start with an isolated sandbox to prove the model against your own abend corpus, then move to a usage-metered enterprise rollout once the playbooks are yours.

Provisioned environment

Experimental Sandbox

$48,000
per quarter · per LPAR profile

A pre-wired sandbox cloned to your runtime profile — JCL libraries, DB2 catalogs, abend corpus — so your team can stress-test Progull against representative batch loads in isolation.

  • Pre-staged z/OS simulator with your JCL/PROC shapes
  • Top 25 abend codes seeded from your last 12 months
  • Up to 10 named operators · ServiceNow dev instance
  • Shadow-mode runs against a captured workload
  • Quarterly read-out with the Progull engineering team
  • Convertible — fee credits against an enterprise rollout
Pay-as-you-go

Enterprise

Usage-metered
platform fee + per-resolved-incident

For production rollouts spanning multiple LPARs, sites and ServiceNow scopes. Priced against measured outcomes — auto-resolved incidents, autonomous remediations, and operator hours reclaimed.

  • Unlimited LPARs · job classes · ServiceNow assignment groups
  • Custom integrations (CMDB, change, SIEM, IdP, paging)
  • Air-gapped, BYOC and managed-tenant deployments
  • Policy-as-code authoring with your governance board
  • 24/7 named SRE pod · 15-minute P1 response
  • Quarterly business review tied to MTTR and toil targets
  • Volume discounts as autonomous remediation share grows

Pricing is indicative. Final terms reflect deployment topology, integration depth and audit scope.

What operators tell us

Trusted in the rooms where downtime is measured in basis points.

"We took a 3 a.m. payroll abend and watched Progull close it in 4 minutes — with a worknote our auditor was happy to sign off. That is the bar."
VP, Mainframe Operations
Global Tier-1 Bank
"Every other vendor wanted to put a 3270 scraper in front of z/OS. Progull is the first team that actually understood SDSF, JES2 and surrogate IDs on day one."
Director, Platform Engineering
Federal Services Agency
"92% of recurring abends now auto-resolve. Our SMEs spend their time on the long tail and on writing playbooks — not chasing S0C7s at midnight."
Head of Batch SRE
Insurance Carrier (Fortune 100)
Frequently asked

The questions teams ask in the first call.

Bring zero-touch operationsto your mainframe batch.

See a live walkthrough of Progull resolving a real abend — from JES spool to closed ServiceNow ticket — against your environment profile.

Request a demohello@progullsystems.com