With changing consumer demands and market conditions, more and more organizations are looking for new ways to drive innovation. The key, according to Google’s thought leaders, lies in data and the rapidly evolving artificial intelligence (AI) technologies.
They revealed five mutually dependent trends that can help organizations unlock the potential of data, embrace the latest AI developments, and innovate, while enabling greater unification and flexibility across their work processes.
The trends are detailed in Google’s latest Data and AI Trends report which includes findings from both Google and market intelligence firm IDC.
Here are the five trends that amount to an interconnected data strategy, according to Google’s thought leaders:
- Unifying data to unveil insights on digital experiences and workflows
“Organizations are realizing that their siloed data storage and warehouse strategies can’t keep up with modern demands,” said Andi Gutmans, general manager and vice president of engineering at Google Cloud. “With the amount of data that devices and applications generate every day, it doesn’t surprise me. They need a better way to store, manage, analyze, and govern all this data, while also cutting down on the extra work, costs, and conflicting insights caused by silos and redundant systems.”
With a unified data cloud, users can have the right information exactly when they need it, accelerate decision making and development cycles, and improve customer experience. For instance, retailers are unifying their data to have better visibility of customer insights, which is helping them increase conversions.
2. Support an open data ecosystem
More organizations are adopting open source software and open APIs to avoid data lock-ins and silos and to protect their freedom to move data between platforms as needed to support workflows, insights, and data monetization.
Gerrit Kazmaier, vice president and general manager, data & analytics, Google Cloud explained that this could help organizations become more competitive, adding, “imagine, all your employees, customers, and partners taking part of your data ecosystem as contributors rather than bystanders.”
Organizations are also making use of available datasets such as weather, trend, and location data to extract valuable insights and develop revenue generating applications. Seventy-five per cent of organizations use location data across a wide range of business functions such as supply chain, public transportation and personalized customer experience.
3. Embrace the AI tipping point
According to the report, at least 90 per cent of new enterprise applications releases will have an AI-embedded functionality.
“Organizations are adopting AI and ML (machine learning) tools and technologies because with them, they can pull out so much more information from the data they have and solve real-world problems with scale and accuracy,” said June Yang, vice president, cloud AI and industry solutions, Google Cloud.
To overcome the AI/ML skills gap, organizations are empowering citizen data scientists to develop models using pre-trained models or low code training models. Eighty per cent of organizations say that having citizen data scientists could help them meet their AI/ML goals.
The report also highlights some AI adoption tips for organizations:
- No need to start from scratch. Instead, use templates, models, and other ready-to-use assets that allow for customization
- Model tracing is key to understanding when the model is trained, who trained it and what data it uses
- Start small. You just need a model that does a task better than what you are doing now
- Using ML to improve search click-through rates can make a massive difference
- Successful AI solutions build reliability and stability into the model at the outset.
4. Infuse insights everywhere
Rethinking business intelligence (BI) and analytics strategies can help organizations’ customer acquisition and retention and improve their decision making.
The report noted that spending on data and analytics is forecast to reach US$200 billion worldwide in 2026. Furthermore, over 70 per cent of organizations improved decision-making and the delivery of actionable insights to all users in the workflow with BI adoption.
While these applications do not necessarily need to have AI/ML, 87 per cent of organizations find it important for their analytics software to embed predictive models.
Organizations are recommended to serve up a consistent semantic layer of data for people to interact with, rather than just raw data. People need to only see the data that’s relevant to them, and they need to know it’s accurate and up to date.
5. Know your unknown data to increase security
Knowing where all your structured and unstructured data is, and classifying it, is the first step in data risk management.
“If you don’t know what data you have, you cannot secure it. You also don’t know what security risks you are incurring, or what security measures you need to take,” said Anton Chuvakin, senior staff security consultant, Google Cloud.
Chuvakin added that unstructured data, from chat applications or log files, for instance, can cause significant headaches for organizations, especially if they unexpectedly contain sensitive data like personally identifiable information (PII).
Organizations need to make their data discoverable before they classify it. Turning to ML and other automation tools can help with that process, which is very hard and sometimes impossible to achieve manually.
The report also highlighted taking a collaborative approach to securing data. By 2025, growth in data marketplaces, privacy regulations, and data sovereignty concerns will lead 60 percent of G2000 organizations to include chief data officers, along with chief information security officers and chief legal officers in their data risk management committees.
Discussion about this post