Most IT run contracts pay partners for effort. That one design choice explains why run costs climb with every server, user and market, and why a dashboard can be 99.9% green on a day the business is losing orders. Here is the operating and commercial model that fixes it.
The contradiction inside traditional run models
Resource-Unit and effort-based commercials penalise a service partner for being efficient. If the partner automates a process or eliminates a class of incidents, its revenue falls. The result is predictable: run costs grow in lockstep with the estate, technical debt stays where it is, and finance is left choosing between predictable spend with no efficiency gains, or innovation promises nobody can verify.
The same models govern by component: is the server up, did we respond within 15 minutes? That produces the Green SLA paradox. Compute, storage and network all hit target while a factory line stalls on an edge integration error, a batch job runs late and Order-to-Cash slips. Every component is “up”; the business is not running.
Three shifts, made together
The AIOps-to-NoOps model changes three things at once, because each one fails on its own. Commercials move from paying for effort to fees tied to verified AI maturity and shared savings. Governance moves from component SLAs to business Experience Level Agreements (XLAs). Operations move from reactive tickets worked by people to predictive, self-healing “no-touch” remediation.
A three-tier maturity and commercial framework
Rather than a vague automation promise, the roadmap is three contractually defined tiers of four months each, so the model proves itself inside the first year. Each tier changes how the estate is run, how the partner is paid and how progress is governed.
Tier 1 protects service during transition with familiar Resource-Unit pricing, while the data is cleaned and outcome metrics run in shadow mode, with no penalties, to create an audited baseline. Tier 2 is where the incentives flip: the partner absorbs operational inefficiency and its fees follow verified automation thresholds. Tier 3 is a deflationary gain-share in which run savings flow into an innovation reinvestment fund, so IT becomes self-funding.
From SLAs to XLAs
Governance shifts from “is the server up?” to “is the business running?” Targets are calibrated against the Tier 1 baseline; the figures below are typical starting points.
| Dimension | Traditional SLA | AIOps-to-NoOps XLA |
|---|---|---|
| Availability | Server or VM uptime, 99.9% | Business process continuity across Order-to-Cash, Procure-to-Pay and supply chain, 99.95% |
| Restoration | Incident response in 15 minutes | Mean Time to Restore Productivity (MTTRP): users and sites fully working in under 45 minutes |
| Quality | Tickets closed per FTE | No-touch rate: over 60% resolved with no human, under 2% reopened |
| Prevention | Root cause analysis after the event | Over 30% of critical incidents detected and averted before business impact |
How no-touch operations actually work
NoOps does not mean no people. It means people stop doing work machines can do safely, and spend their time on the exceptions, the root causes and the change. Every automated fix runs inside an authority model: routine, low-risk actions run alone, production changes need human approval, and every step leaves a log.
Governance without the black box
Finance leaders are right to be sceptical of outcome-based promises; most are impossible to audit. In this model every saving is proven before it is paid, through three pillars: auditable engine logs that verify each no-touch resolution before it counts commercially, a monthly TCO waterfall showing gross savings, automation efficiency and net run reduction, and a joint verification protocol that gives your IT commercial and finance teams a formal five-day window to validate the data before any gain-share is reconciled.
Where to start
The first step is a short working session to map three things: your baseline (run cost, ticket mix and CMDB readiness), the no-touch candidates where automation pays back first, and a draft glidepath with the XLAs and verification guardrails that fit your estate. IPDigital brings both halves of the model: 24/7 IOMC, NOC and SOC run discipline with AIOps layered on, and an AI Centre of Excellence whose products, from Jeeves for no-touch service to Telemetry for device-level prediction, drive the automation.
Tier durations, targets and gain-share splits are illustrative and are agreed per engagement against the Tier 1 baseline.
← Back to insights