This article was first published in eWeek.
In 2016, we saw of artificial intelligence go mainstream and the amount of data captured by many different industries grow exponentially. Financial services, health care, telecommunications, public utilities, education, automotive and other verticals are making major moves to use predictive analytics and AI to better serve customers and unlock greater return on investments.
In 2017, we are likely to see data analytics products and data scientists speed their shift toward deep-domain expertise. Companies – especially those in regulated industries – will realize that all the data in the world is ultimately useless without the ability to address the highly specific industry requirements, challenges and opportunities in analyzing that data. Data science no longer will be (and in fact, never has been) “one size fits all.”
Here are eight reasons why:
1. Lack of In-House Data Experts an Issue
Current employees who do have deep industry knowledge often don’t have the analytical skills required to turn data into actionable insights. Last year, an MIT Sloan Management Review showed that 40 percent of companies report the lack of analytical skills as a critical challenge, but only one in five have done anything about it.
2. Enterprises That Hesitate to Use AI Risk Being Bypassed
With AI being more widely used in the consumer and enterprise worlds, companies that do nothing will risk falling behind irretrievably. And, any outside analytical support will have to reshape its general data models to specific industry needs. Accenture’s partnership with Amazon Web Services is one recent example of an attempt to help clients marry industry expertise with robust data capabilities.
3. Vertical-Specific Tools Can Provide Customized Applications Quickly
An AI system that swims in a data pool containing only credit-card transactions not only will become expert at detecting fraud, but also will be able to provide proactive suggestions. If your metadata shows you’re a frequent traveler, your bank not only will know not to deny you a coffee after your flight to Hong Kong, it also might prompt you to switch to a credit card that offers more frequent flier rewards points.
4. Companies Are Getting Smarter About Tech Investments
Technology vendors no longer will be able to make general appeals to the enterprise or try to woo consumers with aesthetics and style. The enterprise applications that succeed in 2017 must be able to map to specific customer and business paths, which varies widely by industry.
5. AI Portion of IT Market is Growing Fast
Data researcher IDG projects worldwide revenues for IT products and services will grow to $2.7 trillion in 2020, and a large proportion of that momentum will come from third-party platforms that aid companies in verticals such as financial services and manufacturing.
6. Complex Sales Cycles Mean Engineers Need to Be More Than Engineers
Selling a highly-integrated software application as a service means sales cycles are longer and product engineers must be involved right away, engaging directly with customer and prospects. Anyone whose sole job is to liaise between the two groups should be concerned about job security, even those with great “people skills.”
7. Big Companies Are Shedding Bloatware
Big companies will eschew bloatware for applications that provide essential industry-centric applications and nothing else. Bloatware is defined as software whose usefulness is reduced because of the excessive disk space and memory it requires. IDC predicts worldwide spending on cloud applications will increase to more than $141 billion in 2019 from $70 billion in 2015, with the vast majority of growth in industry-specific applications.
8. Companies That Use the Power of Vertical Expertise Will Emerge
Slack, the team collaboration software company, aggressively incorporated third-party apps to help its users build their channels into something much more than email. As Aaref Hilaly of Sequoia Partners said, services such as Slack that focus on integration and automation with existing systems will be used more often. Moreover, those services that people use daily can use AI to capture data automatically – something that big, isolated systems of record have yet to achieve.