Two-thirds of enterprises using artificial intelligence are trying to develop a so-called AI-first culture, but only 25% of them have a broad strategy, according to an IDC survey.
The IDC survey rhymes with what we've found in our TechRepublic Premium surveys and ZDNet deep dives. The reality is AI is a boardroom buzzword and critical technology, but disconnects abound. Half of the enterprises say AI is a priority for their companies, according to IDC.
IDC found the following based on a global survey of 2,473 organizations:
- Productivity, business agility and customer satisfaction with automation were primary AI drives.
- Lack of personnel, costs, and bias were holding back AI implementations. Data management was also an issue.
- IT operations are the top business function deploying AI followed by customer service and fraud/risk management.
- 60% of organizations reported changes in their business model due to AI adoption.
- A quarter of enterprises survey reported up to a 50% failure rate on projects.
- Half of organizations have a formalized framework for the ethical use of AI, bias, and trust.
When you look at that IDC survey it's clear that the technology can be viewed with a glass half full or half empty lens. It's not surprising that 56% of respondents in a TechRepublic Premium survey said that AI projects will be more difficult than other projects.
Meanwhile, most organizations don't have a clear understanding of how AI or machine learning will help their businesses. Many companies are using pilots to test technologies.
If you connect the dots, AI projects are likely to see more disconnects going forward. Here's what I'd argue are the key issues:
The board/CEO thinks AI sounds pretty cool and is asking questions. Hint: You need some AI project pronto. Yes, AI needs to have C-level sponsorship, but a lot of this support has developed because AI provides a good narrative. I wouldn't be surprised at all if more AI projects start with a CEO asking her management team "what's our AI story?" The state of AI in 2019: Breakthroughs in machine learning, natural language processing, games, and knowledge graphs
Vendors are more than happy to sell you an AI magic bullet/black box/happy solution to fix all of your issues. It is amazing how many tech vendors have become AI juggernauts in the last year. Nearly every vendor has an AI type thing with a neat name (for an add-on cost of course) that'll allegedly run your business. These AI boxes may or may not work depending on your data preparation and management skills. Luckily, there's AI for that too.
A lot of cooks in the AI kitchen. AI and machine learning are technical and that reality means there is a lack of skilled data science talent. That issue collides with the fact that AI sounds great and everyone (including managers that have no clue what they're doing) want to be in on projects. There are a lot of folks co-managing AI projects, according to TR Premium. Good luck with that approach, but at least the consultants will be happy.
These aforementioned issues are common threads with historical technology project failures. Buckle up kids, AI is harder than it sounds.
- Free PDF download: Managing AI and ML in the enterprise
- Enterprise AI and machine learning: Comparing the companies and applications
- Free PDF download: How to win with prescriptive analytics
- A guide for prescriptive analytics: The art and science of choosing and applying the right techniques
- The ethical challenges of AI: A leader's guide (free PDF)
- The data scientist job interview: Questions to expect and questions to ask (free PDF)
- TensorFlow: A guide for IT pros (free PDF)
- Telemedicine, AI, and deep learning are revolutionizing healthcare (free PDF)
- Artificial intelligence: A business leader's guide (free PDF)