Are Company Data Strategies Prepared For Artificial Intelligence | Peak Indicators

What is the most common problem ill-prepared companies come across when it comes to data to be integrated into AI?

I’m not sure I understand the question correctly, but there are many problems that companies have then integrating data into AI projects. Some of the issues will be around knowing what data exists and understanding that data; obtaining access to different data sources and ensuring appropriate security clearance and user consent (e.g. ensuring that data can be used in AI applications); having poor quality data; a lack of corporate memory and historical data (e.g. for predictive analytics). For me the greatest issue is having separate data and AI strategies that are not aligned (or maybe a data but no AI strategy).

How important is it to have the right people in place when correcting this? Who should these people be and what skills/training do they need?

You need to have appropriate data, analytics and AI people in place who are able to understand the data and data science lifecycles. However, on their own this is not enough and strategies must include stakeholders from the business to ensure the AI projects are supporting (and not leading) business priorities.

What are some examples of how rectifying the problem can create real world results?

Many organisastions recognise the value of data and AI activities — future success may rely on data and AI capabilities. Having data and AI strategies (that are aligned to each other and the business) can ensure data-driven business translation. For example, see: https://hdsr.mitpress.mit.edu/pub/4vlrf0x2/release/1

Whose responsibility is it for this data to be prepared?

Some companies will utilise a CDO or AI Strategist in order to prepare these strategies. Not sure again what you mean by ‘this data’. If you mean the source data being used by AI projects then this would normally by the responsibility of the BI teams or data/solutions architects. Typically this must involve IT who are likely to be implementing/managing the data systems.

What is the role of the CEO/C-Suite in this?

The CEO/C-Suite must be fully on-board with the data and AI strategies and be clear how data and AI will support future business prorities.

How do we get around data silos and data that is spread across a company and across servers worldwide?

There are various ways of managing this, such as developing a new, modern cloud-based infrastructure that could centralise the data and provision it for use by multiple stakeholders, such as data analysts and data scientists, through the use of centralised data

storage (e.g. data lake, data warehouse), data analytics and machine learning. All of this needs to be governed and fitted within existing business processes and systems.

What risks are there from dirty data that hasn’t been fully cleaned for duplicates and repetition?

There are multiple risks of not applying appropriate data transformations, such as weakening data quality; having to update the same information in multiple places, skewing reports and analytics (e.g. double counting duplicates); giving more work to the data science team (e.g. 80% of data science activity is often spent on cleaning up data and this is time that could be spent on more valuable tasks); and ultimately leading to inferior data-driven insights.

What HR and legal ramifications and concerns should be considered?

If the question is referring to data strategies and their preparedness for AI, then there are multiple issues. If the data provenance and data privacy issues are not made clear, then the AI applications could end up breaking data protection and privacy of regulations, such as GDPR. The use of AI must also include transparency and explainability — the AI processes must be transparent to consumers of AI services or products; and in many domains the internal workings of AI must be explainable, e.g. in financial or healthcare sectors; that is how the AI arrived at the output must be made clear.

If you would like to find out more we’d be happy to talk with you, simply give us a call or drop us an email to set it up!

Originally published at https://www.peakindicators.com on December 16, 2022.