AI robot hovering above a phone

March 16, 2018 | By Tim Searle


Artificial Intelligence (AI) and Machine Learning (ML) are in use in many industries to automate processes, increase efficiency, predict outcomes and make better decisions. But what does this mean for local government and the current systems in use? The technical capability is available and accessible and there is a wealth of historical data, so how can local government take advantage of AI and ML.

How can AI be used for councils?

An exciting growth area is the use of virtual assistants to support and enhance customer service, which will save councils time and effort in the future. In the short term, combining a customer services officer or customer with an AI driven knowledge base and content can provide the efficiency and customer service required and will enhance your current workforce to be far more effective. As customer service representatives interact with the public and capture more information about questions asked and responses given, AI can provide the learning and prediction that can present the 'most likely' responses and filter out lesser used information.

AI is all about learning from variable inputs to be able to give consistent and progressively more accurate outputs. We just cannot physically process and learn from the volume of data we capture without the use of AI and ML. For example, if we want better predictability on processing an application within our service level agreements, it’s currently difficult to perform this real-time on every application received. We could produce reports and dashboards that give us the current state and analyse average processing times, etc. But to be able to predict an accurate outcome taking into account trends, busy periods, time of day, personnel assigned, when the variables are constantly changing becomes difficult. This is where machine learning helps out - the 'machine' can process vast amounts of historical information and understand the variables and in real-time, can let us know with a percentage of certainty, a prediction.

Machine Learning

There are some interesting examples of how we can apply machine learning to existing business processes.
  • Automate the approval of purchase orders. If a manager consistently (98%+) approves purchase orders for specific products, under a value, from certain suppliers the system can recognise this and ask the manager if they would like “us” to approve these in the future automatically. Only the exceptions then need to get visually approved, reducing the managers low risk approvals. At any time, the manager can review these preferences and take control again.
  • When an application is received into the organisation, the most sought-after information from the applicant is, when will this be completed by and will it be approved? We can apply machine learning to an application through its entire workflow process, with a percentage of certainty the expected completion date and an expected outcome. Based on what history tells us about the application, or an inspector performing any inspections - as more information is provided the predictions get more accurate.

Object recognition

AI can also be used alongside object recognition, which is available through cloud frameworks provided by the likes of Amazon, Google and Microsoft. Periodic snapshots of CCTV footage can be analysed by AI to produce information, such as pedestrian counts, animal recognition, free parking spaces, or traffic volumes. The result of the data received can then be fed into machine learning to provide predictive capability around automated park maintenance scheduling (don't mow the lawns at the busiest times!) guiding people to free car parks, and predicting traffic congestion.

What benefits would AI bring for councils?

There are immediate benefits that AI based technologies can bring to local government. Reducing backlogs, cutting costs, overcoming resource constraints by freeing people from mundane repetitive tasks, improving the accuracy of forecasts, injecting intelligence into processes and systems and by handling tasks that we (humans) cannot do easily on our own, such as sifting through millions of documents or sensor readings.

Backlogs of applications, requests, communications are a constant challenge in local government. By introducing robotic or cognitive automation, we can reduce pain points to these backlogs by increasing the speed of data entry, enhancing the reach to our communities and reducing the cost to capture and process information. Repetitive tasks can be identified and automated using Robotic Process Automation frameworks.

If we can become more efficient in the processes we perform today without looking to 'replace humans' we can provide better response times with our community and confidence in the decisions we make.
Datacom Group Limited | All content © Datacom 2018