There is no doubt artificial intelligence (AI) is one of the most hyped technology trends of the past few years. From automating repetitive tasks to supporting better business decisions with machine learning, 80 percent of businesses are already using AI.
But just because something is popular, doesn’t mean it’s easy to pull off.
According to a Teradata study, one in three companies say they need to invest more in AI initiatives, just to keep pace with competitors. Meanwhile, enterprises are experiencing significant barriers to AI adoption. They cite a lack of infrastructure, talent, and funding as the top three obstacles slowing AI projects.
This pattern is a common one with any emerging technology. First comes a wave of hype. And then the monsoon of reality strikes. Businesses realize they simply aren’t ready to adopt this next-generation solution, and so the real work begins. Are you ready to adopt AI? And, if not, what must be done to get there?
3 ways to get ready for AI in the enterprise
This exact theme was featured in our latest Innovation Executive Forum digital transformation report. In it, we saw IT leaders from Ottawa to San Diego lamenting the obstacles of AI adoption.
Take for example an IEF member, who is an IT leader in the retail sector. She said her business simply wasn’t ready for AI because its data infrastructure wasn’t ready. “Our infrastructure is way too behind to even discuss AI. We are still working on the analytics piece. We are probably three years away,” she said, discussing the need to first clean up her team’s data collection abilities. Which brings us to the first thing you need to consider when getting ready for AI.
STEP 1: Ready your infrastructure
Of the respondents in the Teradata study, the most common barrier to AI adoption was “lack of IT infrastructure” (40% of respondents). Just like any major IT project, you need the building blocks in place, before you can start implementing.
When it comes to AI and machine learning initiatives, there are several key pieces of your infrastructure to prepare, according to this Whitepaper from Intel. These include:
- Networking and data center: Determine if existing environments can support the bandwidth and computing resources needed for intensive, data-driven AI projects.
- Cloud: Map out your cloud strategy, determining which workloads will be better gained from cloud. (E.g. Leveraging image recognition and natural language processing out-of-the-box from vendors such as Amazon, Google and Microsoft’s Azure.)
- Data and analytics: Determine if you have the data needed to power your AI project. If so, is it easily digestible and usable in AI applications? If not, how will you find it, process it and ultimately use it?
Laying the foundation is crucial. But there’s another key ingredient you need to move ahead with AI: Business support.
STEP 2: Building up the business case
Another IEF members touched on this point recently. He works in the marine industry and has already rolled out an AI solution to improve maintenance costs and efficiency. He said he didn’t get there by convincing technologists. He did so by speaking directly to the business leadership, in terms they understand best: dollars.
“AI can be a tough sell to non-technology people. But if you can show what you can do with it, it’s amazing, and it’s not a tough sell anymore,” he says.
IT leaders must build strategic support for AI. In fact, according to the same Teradata study, 30% of businesses said a lack of budget was slowing AI adoption. Another 19% said weak business cases were to blame. To succeed here, your business should build up business support in these ways:
- Strategic leadership: Your job is to reveal how AI is crucial to delivering a strategic advantage to your organization in the years ahead. Do this, and the budget and prioritization will follow.
- Clear business use case: Next, get focused and identify a clear, low-hanging business win. Bonus points if you can then implement a proof of concept, to show the reality behind your claims.
- Adoption and acceptance: Don’t forget your end-users. AI solutions demand the collaboration of numerous players across the organization. Make sure you are spreading awareness about the value of new AI tools. And use change management to drive adoption.
Yet, even with the right infrastructure and business support, AI initiatives risk failure if you don’t have the right people, or processes, in place to make them work.
STEP 3: The right talent and operations
The third stage of AI readiness is getting the right resources and operational procedures in place. Once again, this is a common area of frustration for would-be AI enterprises: 34% of enterprises lack access to AI talent; 23% face confusion around policies, regulations, and rights surrounding AI. IT leaders need to ensure AI deployments are operationally ready. They can do so by focusing on the following:
- Talent and resourcing: What is your resourcing strategy for AI? Decide upfront the approach that works best for your unique capacities and requirements – and get started on overcoming the IT skills gaps.
- Get agile: AI initiatives frequently rely on the continuous improvements and speed afforded by agile methodologies and DevOps practices. If you don’t have the team or practices to operate in lean, agile ways, it’s a good time to get started.
- Cybersecurity, ethics, and regulations: AI also comes with numerous, specific challenges related to data security, privacy and ethics. You should not launch an AI initiative until these crucial questions are answered.
It’s time to get ready for AI
AI has lots of promise. But to realize it, businesses must first get ready. To do that, they will need the infrastructure, the personnel, and the processes required for a smooth, successful implementation.
Want to learn more?
Dig deeper with Intel’s excellent Whitepaper, all about the three stages of AI preparedness.
Ready to join the discussion? Request to become a member.