From Good Data to Great AI: Strategies for Organisational Success

Over the last year the buzz around AI has reached fever pitch. The release of ChatGPT and other Generative AI models has captured the imagination (and fear) of every industry. For some organisations and industries, this introduces a huge unknown on how this will disrupt their business models. For others already on the pathway, the question is how they embrace this technology to get ahead of their competition. Across everyone, 2 things hold true.

1. We are all on this journey.
2. Good AI relies on good data management.

Recently I was lucky enough to attend and chair a session at the 8th Investment Data and Tech (IDT) Summit on Advanced Analytics in Practice with Kathryn Kerner (Head of Data Analytics, NZ Super) and Matt Plant (Head of Technology, Strategy and Planning, Australian Super). This session helped reinforce some fundamental concepts that you should be focused on when thinking about transforming how to use and manage data to unlock it’s potential.

 

Here are 4 things to think about to set yourself up for success and build a strong foundation to support things like GenAI.

Start with the Business Objectives

You can build the best data platform in the world with the best people, but if it doesn’t align to what your business needs then it is worthless. Too often I have seen technology solutions designed and built seemingly for the purpose of having new technology.

If you hope to be successful, you must start with the “Why”. That is, start with your business strategy and objectives and make sure you are supporting that.
This is also why good business sponsorship is key. It makes sure you are solving the problems that need to be solved. Not building cool looking things for the sake of it.

Investing in People as well as Tech

Over the last 12 months there has been a significant investment in data. It was noted at the IDT summit that the industry as a whole had made major strides since the previous conference. No matter who I spoke with, they were all going through a data transformation journey focused around the latest technologies and people.

One of the biggest technology moves happening across the industry as a whole is the move from traditional relational data stores (like Microsoft SQL Server that could store your structured data) to the modern data lake architectures (like Snowflake and Azure Synapse that can co-mingle structured and unstructured data).

But that’s not the end of the story. Key to having these technologies work effectively were the people who support, maintain, expand and use these platforms. Those who are successful in transforming their organisation invest in data teams, governance teams and analytics teams to really unlock the power of their data.

When developing your future data architecture and operating model, it’s the teams and people you invest in that make it come to life.

Centralised vs Decentralised Data Teams

“To Centralise or Not to Centralised. That is the Question”. Every organisation is faced with this conundrum. Do they centralise all data activity to ensure a well organised, structured set of data assets or do they federate this across the organisation to empower their people and deliver quickly on the problems they need to solve.

As in most things in life, there is no right answer. You have to think about what best suits your organisation at that moment.

I have seen firsthand the benefits and pitfalls of both models. Too much centralisation and things move way to slow. Too much decentralisation and you are left with something akin to spaghetti.

Whatever model you land on, remember that whatever is built today, must be supported tomorrow.

The Importance of Good Governance

Like JANA, everyone wants to have better, cleaner, more trusted data than they have today. The focus and investment from all organisations on data governance is coming through as a key building block of data transformation journeys. This is vital if you want a trustworthy, secure and reusable data. Below is a summary of some of the core practical steps organisations should be making.

• Data Ownership: Ensure you have business owners that are accountable for the data.
• Data Quality Standards: Ensure you are embedding checks on data when bringing it into your organisation.
• Data Cataloguing: Ensure you are capturing metadata (information about the data) on your data, and this is searchable. Providing people with a catalogue where they can search for available data sets helps unlock the power of your data. Too often you see the same data repeating over and over because no one knows whether it is already available.

Finally, when implementing governance, you must make sure that whatever governance is put in place, is the right size for your organisation.

Conclusion

It can be difficult to keep pace with the emerging AI technology and truly understand what it means for your business. One thing that is certain though is that you must get your fundamental data and structures right first.

Ask yourself:

• What is the business strategy and objective and how does data support that.
• What investment into people will you need to help unlock the power of your data in line you’re your business strategy.
• How should you organise your data teams to strike the right balance between quickly capitalising on opportunities and data resilience.
• What governance is required to ensure data is secure, current, available and of quality.