«Shift-left» is probably the most overused expression in ITSM, and it refers to the tendency to resolve problems as far upstream as possible in the support chain, with the goal of containing costs. The idea is seductive and, on paper, unimpeachable. The problem is that many organizations treat shift-left as an end in itself — a project to launch or a tool to deploy — rather than as the consequence of well-established operational prerequisites.

Transferring tickets toward the less expensive levels of the support chain — the «leftward» movement from which the practice takes its name — produces no cost reduction if the following are not simultaneously in place: a quality knowledge base, self-service design, and mature service desk automation.

The following analysis aims to restore shift-left to its correct logical position. For ITSM organizations, it must represent a destination, made possible by solid operational foundations, rather than a technological solution to be implemented at the start of the journey.

The math that makes shift-left indispensable

Interest in shift-left rests on a precise cost figure: the cost per ticket increases significantly at each step toward more specialized support levels, from the first-level service desk all the way to external vendor support.

According to MetricNet benchmarks, in North America a ticket handled by the Level 1 service desk costs approximately $22, while the same issue resolved through self-service at Level 0 costs approximately $2. Moving up, desktop support (Level 2) costs around $70, Level 3 IT approximately $100, field support around $220, and vendor support up to $600 per ticket. MetricNet also identifies a «Level -1», the furthest left point of all: incident prevention through problem management, which eliminates tickets upstream before they are ever generated.

In practice, the more tickets we can shift to the left — for example, from a password reset handled by a human agent costing $22 to an automated one costing $2 — the more we reduce the total cost. From this cost scale arises the promise of cost reduction associated with shift-left: every request handled at a lower level costs less. The logic holds, however, under one condition only: that the request is actually resolved at that level, and not simply transferred.

When shift-left increases costs instead of reducing them

The savings per ticket presuppose actual resolution, not mere transfer. A self-service request that bounces back to Level 1—whether due to an inaccurate KB article, a clunky portal, or broken automation—doesn’t cost $2. It costs $2 plus the $22 Level 1 fee, plus the cost of any further escalation. A premature shift, therefore, multiplies the overall cost of the problem: it reduces the share of requests resolved at first contact and increases repeat contacts for the same issue.

This is why a shift-left initiative driven solely by the objective of reducing expenditure often produces the opposite effect. A self-service designed exclusively to contain costs sees low adoption and fails to generate the expected benefits. When users cannot find an answer, they tend to turn to solutions outside the official channels.

This is confirmed by TeamViewer’s research on digital friction: 54% of employees say they do not feel capable of resolving technical problems on their own — a signal that a poorly maintained self-service remains unused and forces users to turn to support regardless. The result is the paradox that every ITSM manager should fear: volume metrics look better on paper, while actual experience and total cost per issue quietly deteriorate.

The prerequisites checklist

For the redistribution of tickets across support levels to become a viable strategy, three conditions must already be in place — the true enabling elements of shift-left.

1. Accurate and accessible knowledge

The knowledge base is the number one asset in the ITSM value chain. However, simply «capturing» knowledge is not enough: value comes from its use, more than from mere storage. Established methodologies such as Knowledge-Centered Service (KCS) exist precisely to keep articles updated, relevant, and effective—making issues truly ‘Level Zero Solvable,’ meaning autonomously resolved.

2. Effective management at Tier 0 and Tier 1

Level 0 (self-service) and Level 1 must be designed around the real needs of users, not around software features. An easy-to-use portal with effective search functionality and well-crafted FAQs, and a first level equipped with the tools and authority to close the ticket without escalation: this is where the success or failure of shift-left is determined.

3. Closed feedback loops between levels

Every time a higher support tier resolves a new issue, the solution must be documented and passed down—enabling lower levels to resolve it autonomously next time. The process is structured in four phases: document, search, correct, and resolve. In the absence of this flow for transmitting the knowledge produced, the knowledge base progressively loses currency, undermining the effectiveness of shift-left.

Organizational change management

Upstream of the three prerequisites we have identified is a fourth, largely overlooked one: organizational change management. Shift-left entails a change in working patterns and requires a corresponding revision of personnel evaluation and incentive criteria. Without this, knowledge management, self-service, and automation remain tools that no one truly adopts. And it is precisely on these prerequisites that IT managers should measure themselves, even before looking at ticket volumes. First-contact resolution, the percentage of ‘Level Zero Solvable’ issues, and ticket reopening rates reveal far more about shift-left success than any simple count of redirected tickets.

AI and shift-left: real benefits and hidden costs

Artificial intelligence, and in particular large language models (LLMs), are rewriting the economics of shift-left, but not in the way many hope. Used correctly, AI concretely increases the throughput capacity of Level 1. Gartner identifies agent enablement among the most effective use cases: automatic ticket summaries, quick responses, and next-best-action suggestions that save the agent time without sacrificing accuracy. Similarly, an LLM can generate draft knowledge base articles from resolved tickets, rapidly expanding the knowledge that feeds Tier 0 and Tier 1: it is a powerful accelerator of service desk automation.

The invisible costs of automation without solid foundations

The flip side is poorly designed automation that generates hidden downstream costs. A conversational assistant based on a low-quality knowledge base does not resolve the problem: it merely redirects the request, often incorrectly, and returns inaccurate answers with a misleadingly high degree of confidence. This undermines user trust and increases repeat contacts. It’s no coincidence that in TeamViewer’s findings, 32% of respondents state that the AI solutions they tested failed.

The potential nonetheless remains enormous: Gartner estimates that by 2029 agentic AI will autonomously resolve 80% of the most common service problems, with an operational cost reduction of around 30%. These results, however, are achievable only where Level 0 and Level 1 already rest on solid foundations. Otherwise, service desk automation merely reproduces existing errors at scale.

Shift-left as a measure of maturity

For IT leaders tasked with proving business impact—rather than mere SLA compliance—shift-left cannot be a starting point. It must be built as the deliberate outcome of solid operational foundations. It is necessary to address a widespread practice that tends to redistribute the support load trusting that quality will assert itself on its own. First must come the building of prerequisites (knowledge, self-service, automation, and processes for transmitting and returning knowledge), then, as a natural consequence, the progressive redistribution of tickets toward lower levels, and finally cost reduction, sustainable because it is grounded in actual resolutions and not mere transfers.

Read in these terms, shift-left truly becomes a valuable indicator of the operational maturity of an ITSM organization.

FAQs

1. What is shift-left in ITSM? 

It is the practice of resolving problems and requests as close as possible to the user and the point at which they arise, moving tickets from expensive support levels to more economical ones, ultimately to Level 0 self-service.

2. Why does shift-left sometimes increase costs instead of reducing them? 

The savings presuppose resolution, not mere transfer. A ticket that is sent back to higher levels accumulates the costs of both steps, lowers first contact resolution, and increases the number of repeat contacts.

3. What prerequisites are necessary before redistributing tickets toward lower levels? 

The following are required: an accurate and accessible knowledge base, effective management of Tier 0 and Tier 1 designed around the real needs of users, and structured processes that return to lower levels the solutions identified at higher levels.

4. Does AI allow the prerequisites of shift-left to be bypassed? 

No, AI amplifies existing conditions: on a solid knowledge base it accelerates service desk automation; on a low-quality one it produces incorrect answers, repeat contacts, and loss of trust.