Optimize and automate

Every organization has limited resources: people, money, time and technology. The seventh guiding principle asks organizations to make the most of them. It has two parts that must be taken in order. First, optimize: make work as effective and useful as it needs to be. Then, automate: use technology to perform work that is frequent, repetitive and well understood, freeing people for work that needs human judgment. Automating before optimizing simply makes poor work happen faster. This lesson explains how to optimize, how to decide what to automate, how to estimate the value of automation, and what risks and human factors to keep in mind.

What the principle asks

Optimize and automate means: resources of all types, particularly human resources, should be used to their best effect. Eliminate anything that is truly wasteful and use technology to achieve whatever it is capable of. Human intervention should only happen where it really contributes value.

Applying the principle involves:

• Simplify and optimize before automating: understand and improve the work first.
• Define your metrics: decide how success will be measured before starting, linked to outcomes.
• Use the other guiding principles when applying this one: start where you are, progress iteratively, keep it simple, and so on.
• Recognize that automation is not always the answer; sometimes the best optimization is a simpler process or better skills.

Optimization

Optimization means making something as effective and useful as it needs to be. Note the words "as it needs to be": optimization is not perfection. Making a process twice as fast is pointless if consumers would not notice the difference and the effort could be better spent elsewhere. Optimization should be directed by value (lesson 2).

ITIL describes a typical path to optimization:

• Understand and agree the context: the organization's vision and objectives, and how the work in question supports them.
• Assess the current state of the work to be optimized (lesson 3).
• Agree what the future state and priorities should be, focusing on simplification and value.
• Ensure the optimization has the appropriate level of stakeholder engagement and commitment.
• Execute the improvements in an iterative way (lesson 4).
• Continually monitor the impact of optimization, to identify further opportunities to improve.

Deciding what to automate

Good candidates for automation share several features. The work is frequent, so savings accumulate. It is repetitive and rule-based, so it can be described precisely. It is stable, so the automation will not need constant rework. And errors in manual work are costly or common, so automation improves quality as well as speed.

Typical service management examples include password resets, account provisioning for new staff, routing incidents by category, deploying standard changes through a pipeline, running health checks, gathering diagnostic data when an alert fires and producing routine reports. Work that involves negotiation, complex judgment, empathy with an upset user or novel problems is usually better left to people, perhaps supported by tools.

Estimating the value of automation

Automation costs money to build and maintain, so its value should be estimated before starting. A simple calculation compares the time saved with the cost of building and running the automation.

The general formula is \( \text{payback period} = \dfrac{\text{implementation cost}}{\text{monthly saving} - \text{monthly running cost}} \).

Optimizing people's time

The principle says human intervention should happen only where it really contributes value. This is sometimes misread as "replace people with machines". The intent is the opposite: free people from repetitive, low-value work so they can do work that needs human skills, such as solving novel problems, improving services, building relationships and helping consumers with difficult situations. Organizations that use automation only to cut headcount often lose the knowledge and goodwill needed to improve further.

Risks of automation

Automation has risks that should be managed:

• Automating a flawed process makes errors happen faster and at larger scale.
• Automated systems can fail silently, so monitoring is essential.
• Skills can fade when people no longer do a task manually, making recovery harder when automation fails.
• Automation can embed bias or outdated rules that nobody reviews, a concern that grows with AI.
• Dependencies on tools and suppliers increase, linking to holistic thinking (lesson 6).

Good practice includes keeping documentation of what automated processes do, monitoring their results, reviewing rules regularly and keeping manual fallback procedures for critical work.

Automation and AI

Modern automation increasingly uses artificial intelligence, for example to classify incidents, suggest knowledge articles or predict failures. ITIL (Version 5) keeps optimize and automate as a principle but adds guidance that AI use should be subject to appropriate governance, ethical and compliance controls. That means AI-based automation needs clear ownership, transparency about how decisions are made, monitoring for errors and bias, and human oversight proportionate to the impact of its decisions. ITIL-150 covers AI governance in depth.

Links to DevOps and other principles

DevOps relies heavily on this principle, using automated build, test and deployment pipelines to deliver changes quickly and reliably, and automated monitoring to give rapid feedback (lesson 4). Lean contributes the idea of removing waste before automating. And the principle depends on keep it simple and practical (lesson 7), since simple processes are much easier and cheaper to automate, and on focus on value, which decides what is worth optimizing in the first place.

Optimizing before automating: a worked case

The order of simplify, optimize and automate is easiest to understand through an example. Consider an account provisioning process for new employees. Before any change, a request passes through seven steps: the hiring manager emails HR, HR fills in a form, the form is printed and signed, scanned and emailed to IT, IT re-keys the details into the ticketing tool, a technician creates accounts in four separate systems, and a confirmation email is sent to the manager. Errors from re-keying cause about one request in eight to need correction.

Simplifying removes the printing, signing and scanning, which add no value, and replaces the email chain with a single online form that HR submits directly. Optimizing standardizes the access each role needs, so technicians no longer have to ask managers what to grant, and adds validation to the form so incomplete requests cannot be submitted. Only then does automation make sense: the form triggers account creation in all four systems using the standard role profiles, and a confirmation is sent automatically.

Had the organization automated the original seven-step process, it would have built automation around printing, scanning and re-keying, carried forward the error rate and spent far more on a fragile solution. This is why ITIL insists on the order.

Optimizing beyond technology

Optimization is not always about tools. Sometimes the best optimization is better skills, clearer roles or improved scheduling. A service desk that staffs evenly across the day while most calls arrive in the morning can cut waiting times simply by matching staff rosters to demand. A team whose members each know only one system can reduce delays by cross-training. Such optimizations are often cheaper than technology and can be implemented quickly, in keeping with progress iteratively with feedback.

Key point to remember

Automation multiplies whatever it is applied to. Applied to a simple, well-designed process, it multiplies value; applied to a wasteful one, it multiplies waste. Simplify, then optimize, then automate, and keep people in the loop where their judgment adds value.