dataanalytics > BTEX 2024: How to Build an AI-Ready Organization
Article
6 min

BTEX 2024: How to Build an AI-Ready Organization

Organizations are looking to take advantage of advancements with generative AI, but how can you identify gaps in your preparation? Join CDW Canada as we cover the hierarchy of needs to deploy AI for individuals, teams and across the organization.

CDW Expert CDW Expert
Speakers at BTEX speaking about how organisations can adopt AI
/

When we look at infrastructure or storing data, we’re already really good at that as IT people, The platform is where things start to go off the rails a little bit.

–   K.J. Burke, Field CTO, Hybrid Infrastructure at CDW Canada, presenting at BTEX 2024.

“When you start to look at the workloads, at containers and Kubernetes, there’s a change fundamentally to how we prepare. So all of the things we would need to prepare to run Kubernetes workloads, we would also have to prepare as we look at AI.

“As we continue to invest, we’re building up intellectual property; building up things that have value. That value is going to be indispensable for the company, so we have to have ways to measure it. We want to invest and create new ones, so we have to have a mechanism to scorecard where it’s at, where the new version is, and is the new version better than the old version?

“These are all things we have to do a little bit different than we’ve done in the past.”

Breaking down the AI space by users

/

“We’ve been doing machine learning for decades,” Burke said. “AI is not a new thing. But at this point we’re talking about language models, foundation models, generative AI. And as we talk about generative AI, it’s pretty amorphous. How do we define it in a better way?

“Part of the conversation I’m having with customers is ‘Who are we building the AI for? Who’s the AI going to benefit?’ We’re talking about different data sets. So the way I like to approach it is:

AI for the individual: How are we going to make our jobs easier? How are we individually going to be more effective? The risk profile is that individual having better access to their data.

AI for teams: Taking a function and making it more productive. Finding a way to help teams be more effective. This could be contract analysis, automatic RFP creation. The risk profile is the data you’re using to create the tool, which makes it a little bit easier.

Organizational AI: This is where things can get really challenging. When we talk about building tools for the organization, we’re talking about lots of data, larger data sets, more risk as most of that data is unknown. We haven’t just partitioned a bit of data and built a tool; it’s actually all of the data we have. So there’s a lot of challenges with what is that data, how do we identify the risky data and remove it, how do we make that data more effective?

Customer-facing AI: One thing to keep in mind is the reputational risk of exposing certain data to a customer.

Speakers at BTEX speaking about how organisations can adopt AI

If AI is the engine for innovation, data is the fuel

“There are a couple of things I call foundations,” said Leo Batista, Practice Leader, Data Fabrics, Analytics and BI Solutions at CDW Canada. “Let’s make sure our data is secure, our data is protected, our data is managed and our data is providing the right information to the AI engines to help us make better decisions.”

Batista used a car analogy to get his point across. “There are things in between the trunk and the engine. Data platforms are the pipes, fuel injectors, spark plugs, carburetors working in lockstep with that AI engine.

“Data is literally everywhere. And data can come from IoT sensors, marketing, social networks, ERP systems, et cetera. The types of data we have include structured data, traditional relational databases and unstructured data, where they're not exactly connected or interconnected. So how do we extract value from that data?

“We go through a process of storing, ingesting, transforming, loading, modeling, analyzing and publishing. And then data is ready.”

Core things to think about when planning an AI organization

“We need to think about data strategy,” Batista said. “Who is going to own what type of data? Who is going to own the customer records? Who is going to validate that data? Also data governance. Who approves the data?

“We need to map for core data platforms and modern data platforms. We need to have an architecture design for data visualization and reporting, as well as data science. We want to be using modern infrastructure with governance in place.

“Capacity, portability, governance, quality, resilience, these are the aspects that we need to think about when we think about data. And always think about people, processes and technology. What are the processes, the people and the technology taking care of the data flowing through AI and making decisions on behalf of my company?”

Burke added, “As you put the time into data governance and you start to curate your data, your data actually becomes more valuable. All of that work that you put in to get good data, remove bad data and identify risky data, you're not going to want to do again, right?”

Barriers to success with AI

“AI is evolving very quickly,” Burke said. “There'll be new announcements tomorrow or the day after. So there’s this fear, uncertainty and doubt, and we have to work through that.

“One of the really interesting things about AI is that AI projects are being driven from the top down. Executives want to do something with AI, but they’re not sure what.

“So investing in tools, investing in a methodology, investing in a pipeline and a framework to improve the quality of that data. That's time well spent.

“As you do this, just realize the value of your data, the importance to the organization will continue to increase because the data has become more valuable per gigabyte.”

How CDW can help with your AI strategy

“What we're working with internally around generative AI is really focusing on workshops, assessments, and boot camps,” Burke said. “Over the next couple of months, we're going to be working with customers to do workshops, identify use cases, create a proof of productivity, and then help our customers work through that architecture.

“We're going be able to bring a dozen customers into the office. Let's sit down, let's actually build something together. Let's go through the process. So start asking your account managers about that.

/

Start with something small. Find a subset of data that you can use, and then just know that you've got a partner that can help.