Applying artificial intelligence (AI) in government and education has, until now, been regarded as “innovative.” Maturing technologies and changing stakeholder expectations mean that most organizations should now be planning implementation. Enhancing user interfaces, turbo-charging analytics, delivering individualized interaction pathways, and freeing staff from repetitious tasks should be high on the list of potential targets. When looking at any significant system purchase for delivering services to stakeholders or back-office administration, organizations should be asking the question “How has the vendor improved the underpinning value proposition of this system or service with AI?”
Three core technologies underpin the AI revolution:
Machine learning allows applications to improve their performance at certain tasks without human intervention. Rules are inferred from data and then applied to processing correlated cases. While earlier machine learning implementations required the development of complex rule sets or carefully prepared training data (increasing the risk of bias), more modern implementations can infer rules directly from the data. “Adaptive learning” is a term closely allied with machine learning.
Neural networks produce statistical analysis based on a theory of how the human brain functions. Information is processed by artificial neurons, which can generate identifying characteristics from the learning material provided. “Deep Learning,” another increasingly common AI term, is most often an implementation of neural networks (“deep” referring to the number of layers in the neural network).
Natural language processing (NLP) is the component that allows chatbots to communicate with people by reading, processing, interpreting, and generating human language. For AI to correctly respond to a single question is challenging; the ability to build context across a conversation is significantly more so. However, NLP capability is usually combined with one of the other techniques to provide an end-to-end service.
Specifically, there are three high-value applications of AI that are becoming common:
Chatbots that engage at the stakeholder’s convenience provide natural-language support – both technical and administrative – as well as initiating the processing of routine requests such as filing applications, reporting on progress of requests, and scheduling appointments. Many chatbots can automatically translate with high accuracy, allowing dialogue in the stakeholder’s native language.
Predictive insights, powered by AI are making it far easier to gain insights from the wealth of data now available to most organizations. Natural language queries are rapidly replacing complex user interfaces. The identification of at-risk stakeholders is now a mainstream capability.
Adaptive customer journeys, based on evidence-based analysis of the “most likely to be effective” intervention, can now be facilitated by AI – making a broad base of professional experience available to less experienced staff, and providing validation of the effectiveness of an individual’s preferred practices.
AI’s benefits only become available when three underpinning conditions are met: Firstly, data of appropriate quality must be available wherever it’s needed; all applications of AI rely on substantial quantities of fit-for-purpose data, so managing your data estate as a corporate resource will be a significant enabler. Secondly, the application ecosystem must have the capability to leverage that data. Finally, organizational culture needs to be supportive of evidence-based decision-making and of the fact that roles will inevitably shift to focusing on higher-value tasks as automation is increasingly applied to the mundane.
Straight Talk is a weekly briefing from the desk of the Chief Research Officer. To receive this newsletter by email, please contact us.