Is Your Data Ready To Handle AI?

by Ciarán Daly
Article Image

by Lindsay McEwan


LONDON - Artificial intelligence (AI) is already being applied to everyday problems across multiple sectors, with an estimated 200+ startups and SMEs currently developing AI products in the UK, and venture capital firms investing almost £1 billion in these companies last year.

The use of AI is particularly prevalent in driving the relevant,
personalised experiences today’s consumers expect. It is no longer enough to
simply insert a customer’s name into a preformatted marketing email, instead
companies need to use a breadth of information – from demographic data and
previous purchase history to personal interests and immediate context – to
understand and predict consumer needs and fulfill them, all in real time.

AI has the power to achieve this heightened level of personalisation,
and its output is used for various purposes, from informing chatbots to serving
dynamic ads across multiple touchpoints. According to Forrester research, 88% of marketers are now using AI
in some form to enhance decision making, speed up execution, and improve
productivity. But the majority of these companies are facing challenges that
prevent them from taking full advantage of the technology’s potential, such as
a lack of uniform data.


Related: Is personalization actually possible?


It all starts with data

When AI initiatives don’t live up to expectations it is often because
the technology is implemented without first establishing a strong data
foundation. No matter how advanced the technology, the results achieved through
AI are only ever as good as the data used to train or feed the machine. So when
inaccurate or inconsistent data is used the results will be of little value.
Before launching into AI initiatives, companies must take a holistic approach
to data orchestration, making sure all data is synchronised from initial
collection stage onwards to establish a robust data foundation that can be used
to feed AI tools and technologies. 

Here are three steps to getting marketing data ready for AI:

Step one: Centralise data management

Historically marketing data is collected, stored and processed in silos,
with distinct teams managing their own marketing channels and their own
information. But to create a complete 360-degree view of the consumer,
marketers need to bridge these silos, centralising data collection and
processing, and allowing all data to flow through a single hub, such as a
customer data platform (CDP). This practise creates a single source of
integrated data that delivers a complete picture of the consumer and can be
used to feed AI tools, across the entire organisation. 


Related: Implications of AI go far beyond personalization


Step two: Minimise inherent bias

If biased data is used to train or feed AI technologies,
the resulting insights will inevitably be distorted, so companies need to
ensure they are starting with clean, balanced information. Raw unstructured
data such as content taken directly from the internet is often used to train
machine learning algorithms, but this type of data will inevitably contain an
element of bias. Instead, businesses must implement advanced collection and
management technology within the data supply chain to cleanse and prepare data
sets for use in AI. This must be an ongoing process, with technology used to
identify and cleanse potential areas of bias as they emerge.                

Step three: Respect the consumer

A year on from the enforcement of the General Data Protection Regulation (GDPR), most businesses understand the importance of regulatory compliance. But those that get the best results from data-driven AI are the ones that make ethical use of consumer information a core value rather than something they have to do to avoid penalties.

Companies can respect the consumer by only collecting and processing the data they really need, by being transparent about the purpose of data processing, and by giving consumers real choice over how their data is used – as well as who it is shared with – through a clear consent mechanism. Implementing ethical data practises will help build strong, direct relationships with consumers and will provide a robust data foundation for AI initiatives.       


Related: AI bias isn't a data issue - it's a diversity issue


As the use of AI becomes routine across the marketing industry and other sectors, establishing a robust data foundation with which to feed AI tools is essential. By integrating information so it flows through a single hub, by taking steps to identify and cleanse areas of bias, and by implementing an ethical data strategy with consent and transparency at the core, companies can make the most of the AI opportunity and deliver the relevant, personalised experiences today’s consumers demand.


Lindsay McEwan is VP and Managing Director EMEA for Tealium

Practitioner Portal - for AI practitioners

Story

Hesai and Scale AI open-source LiDAR data set for autonomous car training

6/2/2020

Scale claims this is the first time such data has been released with zero restrictions

Story

IBM adds free AI training data sets to Data Asset eXchange

5/28/2020

Big Blue has something for you

Practitioner Portal

EBooks

More EBooks

Upcoming Webinars

More Webinars

Experts in AI

Partner Perspectives

content from our sponsors

Research Reports

9/30/2019
More Research Reports

Infographics

Understanding the advantages of AI chatbots over rule-based chatbots

Infographics archive

Newsletter Sign Up


Sign Up