The bar Progull is being built to.
Progull is pre-revenue. The figures on this page are design-goal targets for the Year-1 build, not measured production outcomes. Real deployment numbers will replace these once design-partner pilots complete.
Target: recurring abends auto-resolved without human touch
Target: reduction in mean time to recovery on overnight batch
Target: fewer Sev-2 escalations to senior mainframe SMEs
Non-negotiable: audit-grade trace of every agent decision in ServiceNow
Same batch window. Different team.
Human-driven, reactive
- Detection
- Operator notices, pages SME
- Triage
- Manual SDSF / JES spool hunt
- Diagnosis
- Tribal knowledge, war-room
- Remediation
- Hand-typed JCL resubmit
- Closure
- Hours later, sparse worknote
- MTTR
- Hours, sometimes a full shift
Agent-driven, governed
- Detection
- Sub-second from JES event
- Triage
- Telemetry auto-attached
- Diagnosis
- Explainable agent narrative
- Remediation
- Policy-approved playbook
- Closure
- Auto-closed with full trace
- MTTR
- Minutes — typically <3
Downstream of the agent, the whole org benefits.
Operations
Fewer overnight pages. SLA breaches drop. Run-books become living code.
Engineering
Senior SMEs stop being the human dispatch system. Their expertise compounds into playbooks.
Business
Payroll runs on time. GL closes. Cut-offs are met. The mainframe stops being the bottleneck.
Where the lift is expected to show up.
Design-goal auto-resolution rates on the families that dominate overnight batch. These are the targets Progull is being built against — not measured customer deployments.
What a design-partner pilot is scoped to look like.
Progull is pre-revenue. The arc below describes the intended shape of a design-partner pilot — Advisory tier only during the pilot; higher tiers unlock post-pilot for confirmed customers.
- Weeks 1–2
Advisory on top abends
Agents observe the top recurring abends on a scoped, non-production subset. Operators read reasoning traces; zero action on z/OS.
- Weeks 3–12
Recommendation review
Agents produce full remediation proposals inside ServiceNow. Weekly review calls sharpen the playbook library. Still no autonomous z/OS actions.
- Weeks 13–24
Playbook hardening
Approved patterns roll into the shared library. Design-partner sees the full evidence pack for every recommendation before any promotion out of Advisory.
The org chart gets quieter.
Progull retires entire categories of standing calls — the war-rooms, the post-mortems on repeat causes, the 7am hand-offs.
- The nightly batch stand-up reading run-book status to a roomful of people.
- The Friday RCA on the same recurring S0C7 that everyone already knows.
- The Sev-2 escalation that wakes up the senior SME for a 90-second fix.
- The hand-off email at 7am explaining what the on-call did and why.
- The quarterly capacity review that surfaces issues already fixed.
- The audit prep sprint that re-derives evidence the agent already attached.
Model the savings on your batch.
Six inputs. A printable, shareable report. No login required — email yourself the PDF if you'd like a copy.
Your inputs
Six numbers. Adjust to match your batch.
Your ROI report
| Operator-time savings | $353,600 / mo |
| On-call relief | $1,440 / mo |
| SLA-breach avoidance | $226,667 / mo |
| MTTR delta | 52 → 3 min |
Assumptions: new MTTR set to 3 min on auto-resolved; SLA-breach probability at 4%; on-call relief at 25% of standby hours. Tunable in a deeper engagement.