5 Principles to Know

7 mins

Automation is the use of technology to reduce the human input required in a task, process or system, especially those involving manual or repetitive actions. In any automation, there will be an input (the data entered into the system) and an output (the system outcome). In feedback control systems (Principle 3), the input (also known in these systems as the ‘set point’ or ‘reference value’) can be understood as the desired state the system is to achieve (e.g. in the case of a thermostat the desired room temperature) and the output as the current state (or system variable) that might need modification (e.g. the actual room temperature).

01 TA01 0

Automation systems are programmed to determine the output based on the input (Principle 2) and feedback controls can achieve a specified output by self-adjusting to environmental feedback (Principle 3). These simple concepts can be scaled beyond one discrete task across a whole process, using technology within a self-governing system to replace the control and operation of the system itself.

This flexibility has seen automation employed in many different ways across a wide array of sectors. From robotic process automation that uses rule-based applications to simply replicate human action, to more complex AI-based automation tools that replicate human judgment and decision-making. This flexibility allows users to tailor the use of automation to meet their specific needs, resource availability and technical expertise, and decide whether it replaces or simply augments existing processes or systems.


An automated system relies on machine or computer programming to determine the sequence of steps to be taken, which can vary significantly in their complexity. In a simple system, this could be gathering revenue figures across different departments and consolidating these into a central report. However, a programme can also define the conditions that must be met before certain actions are performed e.g. producing a report on the last day of each month or only flagging a report for the CFO’s attention if total revenue is below a certain amount.

01 TA02 01

Low-code or no-code platforms allow users with little or no programming expertise to implement automation systems. These platforms typically involve the presentation of logical connectors (e.g. ‘if’ both A ‘and’ B happen, ‘then’ do C) in a format that enables users to create a process flow. For example, for each company named in an agreement, if it is labelled as a UK company, perform a search on Companies House to verify its details. Alternatively, if it is labelled as a foreign company, try to identify its country of incorporation, then if the country of incorporation can be identified, perform a search on that country’s companies register, if not, flag it to a human operator’s attention.

Once agreed, a process flow can then be deployed in an automated system. For example, as a self-serve tool or chatbot where users answer a series of questions to reach a solution, with each question determined by the user’s previous responses e.g. if the answer to Q1 is ‘no’, then skip Q2 and ask the user Q3 instead. This allows the user to obtain tailored responses or advice. These process flows can also be deployed to build tools that automatically gather and filter for relevant data within the business, supported by data visualisation techniques for ease of interpretation e.g. building dashboards that keep track of key performance indicators.


Automated systems can be ‘open-loop’ or ‘closed-loop’, with the key difference being the use of feedback controls. An open-loop (or non-feedback) system simply acts on a user’s input or command. That is, it executes the command exactly as specified and does not adjust based on feedback to achieve the desired output. For example, a clothes dryer programmed for a specified period of time (input) will continue for that period of time and then stop, regardless of whether the clothes are dry.

In contrast, a closed-loop (or feedback control) system achieves the desired output by adjusting its processes in response to environmental feedback. In particular, a negative feedback control system monitors the difference between the input and the output (with any variation generating an error message) and adjusts accordingly. In a closed-loop system, the clothes dryer would only stop working when the clothes were dry e.g. when it detects that there is no difference between the input (requirement for dry clothes) and output (whether the clothes are in fact dry).

01 TA03 01

Both open and closed-loop systems offer benefits. Open-loop systems are less complex and therefore offer a relatively cost-effective way of realising some advantages of automation. However, they likely require ongoing human oversight to achieve the desired output as they do not have an automated monitoring system. In contrast, closed-loop systems offer greater automation and an ability to respond to environmental changes by relying on feedback loops to determine when the desired output has been achieved. As a result, they generally involve more complexity.


The goal of most automation systems is to utilise data in a manner that enables the system to correctly calculate and generate the corresponding output or action. It stands to reason therefore that the quality of the system’s output depends on the quality of the data that it relies upon. The dimensions of data quality are discussed in the data module here, but incorporate factors such as timeliness, completeness and format.

01 TA04 01

The quality of a system’s data is particularly important in feedback control (or closed-loop) systems (Principle 3), as here the resulting outputs will be fed back into the system. If an output is based on inaccurate data, this may result in a feedback loop whereby errors are either repeated or amplified. In these circumstances, it is important that the system incorporates a mechanism whereby humans can intervene if needed to correct the system error in a timely manner.

Inconsistently formatted input data or poorly designed systems require human effort to manually correct the outputs or actions of the system. This effectively diminishes the value of the automation system in reducing time costs and minimising repetitive human input. After a system is deployed, its performance should be monitored and its system kept up-to-date with the workflow so as to avoid an inconspicuous buildup of human correction efforts required.


Most legal processes include repetitive and manual tasks which are ripe for automation. Even the more complex processes may consist of many smaller sub-tasks where automation can be implemented; the objective is not to achieve 100% automation, but rather to achieve efficiency gains where possible. Contract automation is a common application of automation in law, with more sophisticated systems able to include certain clauses automatically, if predefined conditions are met (e.g. including a provision related to a particular benchmark rate only if the loan interest rate references that particular benchmark).

Another common application being deployed by in-house teams is the provision of automated 'self-help' advice. By creating an automated question-and-answer decision flow through low-code or no-code platforms (Principle 2), a business user can get answers to common legal questions. Where more nuanced advice is needed, beyond the capability of the system, these tools can also direct the users to a human operator. Here, the system still provides a value as it can automatically generate these referrals to human advisers, attaching all background facts that have already been provided by the user.

NEW 01 TA05 01

Various aspects of client relationship management can also benefit from automation. These include cataloguing clients’ interactions with key contacts within the firm, customising and keeping track of the marketing communications that clients receive, and gauging their interests in webinars and knowledge materials that the firm publishes online. The data generated from automating these elements of the relationship can also be analysed alongside revenue and billing figures, helping to inform marketing strategy.


Thank you for signing up for LawtechUK news and updates

You will now be notified of all our latest updates.

Thank you for contacting LawtechUK

We will get back to you as soon as we can.

Thank you for your interest in the Open Legal Data project.

The LawtechUK team will be in touch with you shortly.

Thank you for signing up for LawtechUK news and updates

You will now be notified of all our latest updates.