Demand for Kubernetes skills is growing. Whether you're a hiring manager or a job candidate, these 20 questions will help set you up for success.
Getting started with AI in 2019: Prioritize data management
If you've resolved to do more AI experiments this year, take a hard look at your data management practices first
Just before the New Year, I spoke to a CIO client and friend who was excited to talk about how to “hit the ground running” on projects that involved artificial intelligence (AI) in 2019. Like many CIOs, she’s eager to tap into the benefits and get ahead of the disruption that AI promises to bring. In our conversation, she asked me to help her understand how to prepare so her organization can “do AI right” in the coming years.
My response? Make sure your data is in order before spending one dollar on AI! If you want your AI initiatives to be successful, you must make data management a top resolution for the New Year.
That can be hard to hear for a CIO who has a team and organization chomping at the bit to “get into AI.” It can be very tempting to dive headlong into data science and AI initiatives. However, it can be difficult to make headway without first understanding the importance of data management (and other aspects of data).
Remember – AI is data. You can’t do anything with AI or machine learning without data, so you must ensure that you understand and manage the lifecycle of that data.
[ What's coming next? Read AI in 2019: 8 trends to watch. ]
Hallmarks of strong data management
Data management isn’t one of the sexiest aspects of the CIO role, but it’s vitally important to machine learning and AI. The old saying “garbage in, garbage out” fits perfectly here because if you have bad data, you’re going to get a bad model. A bad model will, in turn, tell you to do the wrong things, which can really do some damage to your organization.
That said, when your data is managed properly, AI can absolutely transform the abilities and possibilities for an organization.
To ensure your organization gets started down the right path with AI, you’ll need to take a good look at your data management practices. Among the key elements of strong data management is an understanding of:
- Where your data has come from
- Who has accessed and/or changed that data
- How your data can be used (e.g., do you have the right to use the data for other purposes?)
- When your data was collected
- What your data has been used for in the past (and how it might be used in the future.)
4 areas to examine
For the coming year, think about your goals. If AI is on that list anywhere, you need to seriously consider engaging in a few best practices that focus on data and data management. Consider them New Year’s resolutions of sorts.
First, to ensure your data isn’t garbage, you need to start with the big picture, which may sound counterintuitive. You need to build a data strategy that answers the “big” questions around your data and then think about key elements such as governance, quality, and integration. Outlined below are a few areas that I believe will help you get prepared for AI:
- Data Strategy: The “who, what, when, why, and how” for your data. Your data strategy is going to inform everything else you do. If you don’t have a data strategy, you need one.
- Data Governance: The rules and systems that are (or should be) in place to manage your organization’s data. Data governance should be driven by your data strategy. Your governance should consider (and manage) all aspects of your data, including data quality, data access, and data integration.
- Data Quality: The processes and systems that ensure your data is accurate and useful. Data quality starts the instant data is collected, and it continues throughout the data lifecycle. Data quality should be informed and driven by your data governance rules/systems.
- Data Integration: Many people lump data integration into other areas (or they just gloss over the topic), but it should be a consideration when you think about data. It will be informed and driven by data strategy and closely tie into data quality. You must spend time thinking about how data will be integrated throughout your organization and throughout the entire data lifecycle.
Data management, data strategy, and data governance might not be as sexy as talking about AI and machine learning, but you must get these data areas in order before you’ll ever be able to do AI right. When your colleagues at the next networking event are going on and on about AI, remind them how important data is. Remind them that “garbage in = garbage out” – especially when it comes to AI and machine learning.
In the coming year, I expect we’ll see more resources being shifted toward AI and AI-related initiatives than ever before. If it’s a focus area for your IT organization, set yourself up for success by getting your data in order now.
[ How can automation free up more staff time for innovation? Get the free e-book: Managing IT with Automation. ]