AI and Technology: A symbiotic relationship

Posted: 8th September 2023 Lauren Trevelyan, Principal Consultant

Artificial Intelligence (AI) as a disruptive technology continues to be a hot topic, whether its stories of heavily prejudice social media content generated by AI, fictitious case law being cited in court, advances in research that generate medical breakthroughs, or as a tool to increase employee productivity and engagement.  There is still much to understand about this evolving technology and its application.  One thing is certain, eyes will turn to the Technology decision makers to provide a view on where to deploy the benefits and the risks involving AI.

Here at Altair, we have been looking at the impact of AI on organisations in the sector. This article focuses on deployment of AI through the lens of Technology, providing 5 key areas to support Technology Leaders and answer that crucial question, “Are we ready for AI?”

Understanding the risks

It is crucial for organisations to educate all members of staff on the different types of AI tools that are available, and to highlight the risks they pose to operational performance. As an example, using simple AI applications, such as Robotic Process Automation (RPA), to move specific data from an electronic document into specified fields within an IT system would carry less risk than actioning a customer request based on pre-defined parameters and logic, as used in machine learning (ML).

The application of more complex Generative AI (Gen AI) to automatically create and send written responses to customer complaints based on case notes, emails, and generically researched customer service data carries a much higher risk.

The rise of shadow IT and/or citizen development is widely recognised. To address this, establishing design principles and guidelines from the outset, while creating and heavily promoting a ‘work out loud’, open, inclusive, and collaborative AI community will generate innovative ideas and activity that is in line with IT and Data governance frameworks.  Enabling the reputational, financial, cyber and data handling risks to be assessed, mitigated, documented, and monitored.

Getting the foundations in place

AI makes processing decisions based on the initial data it is provided with, so it is vital to understand the data, from where it is data mastered, and all the integrations. Will AI generate new data? Does the output need to be stored securely?  Is the quality of the data itself high enough to use as a prompt for AI? If the data is low quality, i.e., has inconsistencies, gaps or is incorrect, then there is a higher likelihood of AI making ‘wrong’ decisions. So, having well established data quality routines with fit for purpose data governance is essential.

It is essential to seek assurances that there is a robust understanding of the complete spectrum of business processes, including all the variations, workarounds, and any off-system activities. Equally significant is the need to measure the efficiency savings that are achievable through AI automation. It is imperative that the desired outcomes can be empirically measured and align with Key Performance Indicators (KPIs).  Although these tools are intentionally open source and available via platforms like Azure OpenAI, establishing the necessary groundwork, up skilling teams, and running AI projects requires investments. Therefore, evidencing success to the Board is key.

Determine the right use case

AI thought leaders such as Microsoft promote the use of Gen AI as being supplementary to human interactions, an example of this is Microsoft Co-pilot. So, it remains imperative to understand the benefits, the limitations, and how the varying types of AI can meet desired outcomes, whilst also maintaining a personal service.

So, what are the steps?

Assess the risk, the data quality, the level of transparency and level of consistency in the business process.  Recommend that operational teams seek out processes that are deemed high volume, low risk, with good data quality, and all system based with predictable customer outcomes.

Team capability

Digital democratisation and the availability of AI code means that the role of the IT professional is changing. There is a requirement to have a deeper understanding of business outcomes, be versed in user-centred development with data driven design and have the skills to foster a culture of collaboration which requires Technology leaders to think laterally and consider those with transferable skills while capitalising on widely available training materials.

In conclusion, assess team capabilities, start small and keep it simple.  Validate the use case, evidence the benefits, gain insights from the experience and refine the approach as needed.  if the chosen use case where AI has been deployed isn’t deriving benefit, quickly understand why, apply the knowledge and be open to seeking out alternative use cases to trial the benefits of AI.

If you want to know more on how to prepare your technology function and organisation for AI, please contact:

Ian Lever
Director of Technology.

 

 

 

 

 

Latest News

See all news