Sergey Nivens - Fotolia
CIO must focus on easing data access to help data scientists build models more quickly
Data scientists struggle to get the right data in the right format to build artificial intelligence models, so to benefit from AI, CIOs will need to simplify data access
More than half of global data and analytics decision-makers say they have implemented, are in the process of implementing, or are expanding or upgrading their implementation of artificial intelligence (AI), according to Forrester.
The analyst firm predicted that intelligent process automation (IPA), which combines robotic process automation’s (RPA’s) task automation capabilities with pragmatic AI building blocks, such as text analytics and machine learning, will be deployed by a quarter of the Fortune 5000 companies.
However, in its Predictions 2020: Artificial intelligence report, Forrester said data scientists spend over 70% of their time prepping data before they can even begin to build machine learning models or gain any AI value.
The authors of the report wrote: “Data scientists often struggle to acquire, transform and prepare the data they need to start a machine learning project. Data lakes, data engineers and data prep tools have helped, but the real problem is sourcing data from a complex portfolio of applications and convincing various data gatekeepers to say yes.”
Forrester said that in 2020, senior executives such as chief data and analytics officers and CIOs who are serious about AI, will come to the rescue with a top-down mandate to get round the data problem. Its research has found that organisations with chief data officers (CDOs) are about 1.5 times more likely to use AI, machine learning and/or deep learning for their insights initiatives than those without CDOs.
According to Forrester, in 2020, senior IT executives will ensure their data science teams have the data they need in order to spend the majority of their work developing machine learning and AI models, rather than accessing and formatting data.
While many companies are experimenting with conversational AI, Forrester predicted that success will be limited. The report said: “Despite the maturation of the toolsets, including the expansion of pre-built and vertical-specific intent libraries and higher-power natural language understanding (NLU) engines, by the end of 2020, conversational AI will still power fewer than one in five successful customer service interactions.”
Read more about data access for AI
- There may be value in including artificial intelligence in backup software. Find out which vendors are already using AI to improve backup and data protection capabilities.
- The relationship between data scientists and companies using AI is evolving rapidly, shifting from a focus on trained professionals to experienced employees with automated tools.
Although non-tech firms will continue to deploy such AI technologies, Forrester expects a shift in emphasis among technology leaders towards linking AI with design. “Today, companies like Adobe and Google pair human-centred design and AI development capabilities,” the report’s authors said. “Next year, these tech elites will ramp up their efforts to find people with knowledge in both fields.”
As an example, Forrester described how Google’s Gmail product teams have focused their skills on techniques to respond to user feedback more quickly and incorporate user data into usability testing, with approaches such as “Wizard of Oz” prototyping. According to Forrester, these techniques allow them to learn quickly and to surface issues earlier.
It predicted that companies that are not in the tech sector will focus on the “look and feel” of new products rather than basing their product design around AI. It expected that in 2020, just 5% of design or AI job postings will mention the connection between the two fields and many AI-led design efforts will struggle to gain user adoption.
