Data management is foundational for successful AI projects; companies must focus on ensuring their data is clean, centralized, and accessible to effectively train AI models.
Unless you have been living under a rock you will have been exposed to the tremendous coverage and hype around AI and machine learning. You know when Oprah Winfrey is dedicating an hour-long show to discuss this trend that it has hit not just business and IT circles, but the mainstream media and consumer population too. But how much of this is hype versus reality?
I have spent many hours on the phone with CIOs, tech teams, partners, and more on the topic. Based on these efforts here are some observations as well as what it means to today’s ERP communities:
Most companies are still at incredibly early stages of AI experimentation and adoption: In a recent CIO study we found that 43% of executives in our study hadn’t even started any AI projects yet. Only 8% of leaders had completed more than five AI-related projects.
However, IT and business leaders are aggressively investing in learning, education, and experience: Every CIO I have spent time with over the past few months is scrambling to establish an AI knowledgebase, skillsets, and project experience. Getting any experience now is essential to gaining and keeping a competitive advantage. Our research also showed that the top challenge of implementing AI was the lack of business case and skillsets.
“We are looking to start with quick wins and leveraging partners such as Google. They have a data set of roughly one million users that are utilized to automate some basic office automation tasks such as meeting transcription. Next, we are using AI bots for customer service which is another common use case. We are also starting to experiment with algorithms to replenish goods in our stores,” comments Stijn Stabel, CTO of Carrefour Belgium.
It all starts with the data: Executive after executive that I have interviewed are spending a significant amount of time focused on their company data and making sure that it is clean, centralized, and accessible. Many companies know that their AI projects will feed off their proprietary data and need to use that data to train their models.
“We are working on our data strategy right now and examining data lake technologies. We know that our ERP group generates at least 80% of the data that is going to be important to our overall AI strategy,” comments Mark Slater, Vice President, Digital Business Solutions, Reynolds Consumer Brands.
Your Partners have invested in or are investing heavily in AI solutions: From SAP to major consultancies such as Accenture to software and hardware partners, all are aggressively investing in AI and touting their built-in AI capabilities and features. Understand how you can unlock these capabilities and any technical prerequisites that you will need to meet them.
What this means to ERP Insiders
Aggressively gain experience with AI
It’s important to gain practical experience with AI and ML so that you can build internal resources and knowledge. Partnering can help accelerate learning, but you need to preserve your internal knowledge base so that you can maintain your competitive advantage.
Know your partner’s AI strategy
Leverage solutions and technology you already have. Spend some time with your partners to explore use cases that can impact your business the most and talk to other customers who are further along the path so that you can learn from their experiences.
Evaluate your data
To unlock data’s full potential in training and development, you must be able to centralize and validate this data on an ongoing basis. Explore cloud-based data lake technologies to select the right solution and path for you.
Build an AI center of excellence
You need a centralized organization that is only focused on AI which helps set standards and governance models, explore partner solutions, and systemize learning and knowledge sharing across the company. It is important that this organization carries with it both IT and business skills and personnel to ensure that it reflects the needs of both entities.
