Artificial intelligence (AI) is the talk of the town. Yet as enterprises ramp up use of AI and advanced analytics, they are running into roadblocks that diminish the competitive advantages of data-first business.
Organizations developed an appetite for data-related AI for good reason. According to research from Enterprise Strategy Group, data-first leaders are 49% more innovative, 54% more likely to exceed revenue goals, and 20 times more likely to beat competitors to market by multiple quarters.
What are the biggest challenges organizations face as they leverage AI and analytics to drive data-first business? We posed that question to the CIO Experts Network, a community of IT professionals and technology industry influencers. Here are their recommendations.
The power of insights
“It’s crucial that organizations leverage analytics and AI to drive data-first business to ensure they can gain real insights for winning corporate strategic decision making,” notes Elitsa Krumova (@Eli_Krumova), a global thought leader and tech influencer.
In specific industries like retail, those failing to embrace an imperative for AI-driven insights risk being left behind. “Retailers that harness AI and [machine learning] ML insights to understand their customers on a deep level – and on the flip side, identify who is not a legitimate customer – will create superior experiences in-store and online,” says Michael Reitblat (@ForterGlobal), CEO and co-founder at Forter. “Ultimately, this will lead to stronger customer loyalty and lifetime value, all while stopping fraud from impacting the bottom line.”
Challenges slow adoption
Despite the clear upsides, there are an array of challenges that undermine the benefits of AI and data-first transformation. Research conducted by Emerald Research Group for HPE found that while the lion’s share of companies (98%) report using AI in some capacity, including proof-of-concept and pilot programs, only a fraction (14%) have fully realized their AI strategy. What’s more, 89% confirm they are facing difficulties executing AI initiatives.
Among the more prominent challenges: aligning AI strategy with business strategy, building the right skills and talent, and creating a culture of data collaboration, according to Kieran Gilmurray (@KieranGilmurray), CEO at Digital Automation and Robotics Unlimited.
Mastering data quality is essential for ensuring the data used to develop AI-driven insights is complete, accurate, and reliable. Without the right governance, data quality challenges persist, and traditional methods may not be enough when dealing with new data types, notes Helen Yu (@YuHelenYu), founder and CEO of Tigon Advisory Corp.
Data privacy is another issue that needs to be properly addressed as part of strategic AI and data-first business strategies. Failure to formalize and initiate data privacy policies and practices can expose an organization to undue legal, financial, and reputational risk.
Biases can also hinder successful AI initiatives. Companies that purchase synthetic data from third-party vendors, which is prone to distinct biases, can inadvertently introduce biases into their AI platform’s foundation. Companies often aren’t aware of the inherent biases — sample selection bias, prejudice bias, and measurement bias, for example — that might exist in training data. “When the algorithm or machine learning model processes and reprocesses this data, these biases increase exponentially or manifest themselves in unprecedented ways,” says Dipti Parmar (@dipTparmar), chief strategist at Dipti Parmar Consulting and co-founder at 99stairs. “This is GIGO (garbage in, garbage out) at the next level.”
Many of the challenges to successful AI implementations come down to forging and facilitating a robust and holistic data strategy. For example, data liquidity — the art of moving data from points of origin to points of use — can be a major stumbling block. “Liquidity is the measure of how seamlessly data can move between devices, applications, and systems,” says Peter Nichol (@PeterBNichol), chief technology officer at OROCA Innovations. “Do this well and data will be available in real-time when decisions and insights are required. Do this poorly and departmental data silos will drive deeper tunnels throughout your organization, making accessing data nearly impossible.”
There’s no question that organizational silos have tremendous potential to hold back the promise of AI-driven business outcomes. “Many large organizations operate and make decisions in groups defined by P&Ls, business units, and functions,” adds Michael Bertha, partner at Metis Strategy. “This results in a technology estate wrought with disparate data and applications, each designed to drive outcomes that benefit a specific unit as opposed to the enterprise.”
Moving forward with data-first business
There are many ways organizations can address these hurdles and create a solid data-driven foundation. Among them:
- Lean into change management. Changing the status quo is a challenge with any major business initiative, and AI and analytics are no different. Create programs that educate staff on the importance and benefits of data-driven business while also providing training on best practices and solutions. “Without this, organizations face problems with employees lacking understanding of how to get the best from their data and draw out the right analytics to further their business success,” explains Ramprakash Ramamoorthy, director of AI research at Zoho.
- Embrace an integration strategy. Given all the doors AI and analytics can open, it’s important to consider a host of tools and strategies built around an open platform. “There is more that can be achieved with a system and strategy that embraces a variety of AI tools and integrations with core business applications compared to one that does not,” notes Sawan Joshi, director of information security at Cervest.
- Measure the impact of AI and analytics. Many companies are launching AI and analytics initiatives somewhat blind and not following through to measure their impact and overall business outcomes. Without those KPIs, projects and overall enthusiasm for the initiatives can stall. “Not knowing how to effectively measure the impact of what they are leveraging as it relates to the business, culture, or customers can be a challenge,” says Tom Allen, founder of the AI Journal. “Data has the potential to be an asset to a business to take it to the next level of success. It can also be a liability that collapses it overnight. We don’t often see data evangelists as the business global titans of today. Wouldn’t it be nice to see that shift?”
Click here to learn how HPE can help turn your data into intelligence.