Data is now a currency, but one which carries extra responsibility for the holder especially where personal information is involved. This data currency will come under regulation in 2018, where failing to get a clean audit will have similar reputational and monetary consequences as failing a finance audit.

Getting data storage infrastructure and dataflow right is essential to equip organisations to take advantage of the machine learning and AI technologies that leading businesses are already deploying.

>See also: Plain sailing: a smooth journey for the CISO towards GDPR compliance

Being able to shift data and workloads between clouds to take advantage of multi-cloud and hybrid architectures as cloud usage and the data regulation landscape evolves, is going to become a key capability for IT teams. To support this, delivering native integration across public clouds and on-premises platforms has to be a priority for vendors.

Below, Alex McMullan – CTO EMEA at Pure Storage – outlines his predictions for 2018, regarding the importance of data in enterprise.

Data stewardship must become a core competence in 2018: IT and culture must change to support that

GDPR implementation is imminent in the EU. It fundamentally requires organisations to be good data stewards. That means knowing and showing where personal data is, where it is not and demonstrating fine grained data control from ingestion to deletion.

Becoming an exemplary data steward or responding effectively to security incidents, is near impossible with systems that take days to backup, index or restore data. To get to the stage where this level of control and safeguarding is possible we are going to see significant investment during 2018 in faster networks, search and indexing. Tools and platforms to improve the visibility, manageability and performance of data pipelines in general will also see substantial investment.

>See also: Accelerating automotive development with HPC simulation

That said, organisations that rely on technology alone for GDPR compliance will struggle. The regulation applies equally to everything from a 20-year-old backup tape stored by a security vendor to today’s cloud technologies, and the card references or microfiches gathering dust in a neglected storeroom. This means that technology alone is not enough. There will also be significant cultural and procedural changes needed to achieve and maintain GDPR compliance.

The right cultural approaches need to be led by senior management, and the right tools need to be implemented to support that behaviour. IT can help but it has to be part of an end to end approach, starting with the data architect and permeating the organisation from the back office through to every customer facing representative.

Leveraging AI and machine learning

The problems of storing and delivering the data needed to train AI and machine learning (ML) are largely solved by advanced technologies like Pure FlashBlade and NVIDIA DGX-1. So the challenge for CTOs in 2018 is to filter out the noise to identify what is viable now, and what can deliver cost effective business benefits.

As an IoT company Pure has already successfully incorporated ML techniques into both its software development and support ecosystems and can attest to the real value of using this approach to bring focussed business benefits to both internal development teams and customers seeing the downstream benefits.

>See also: Boardroom take note: ‘Security is a journey not a destination’

Training a data pipeline to learn, filter by exception and apply business rules to the exceptions is now within reach of any enterprise, whether that is autonomous driving, software development, medical imaging or cybersecurity.

Compelling stories such as Google DeepMind’s AlphaGo Zero learning to play Go better than the previous AlphaGo system (which had already defeated all comers) give a sense of the untapped potential of ML and Restricted Learning (RL) in this case.

For 2018, I expect to see ML and RL applied more widely across a range of industries to streamline labour-intensive tasks and increase the execution quality of those tasks in parallel. Robotics is one of the most interesting applications of AI technology, no longer just for industrial robots in manufacturing but everything from drone lifeguards on beaches to help swimmers and watch for sharks all the way to more advanced prostheses to help people with spinal injuries or who have lost limbs.

Data from Forrester suggests that 70% of enterprises expect to implement some form of AI over the next year. However, I believe that, across the full range of businesses, the benefits of ML are going to be felt more immediately. ML’s quick wins will be lower down in the business technology stack. ML based automation is already proven to save hours each day in the routine administration of IT infrastructure.

>See also: Staying connected: 5 mobile performance management tips

Instrumenting systems, via the Internet of Things, and using ML to analyse the data delivers valuable, actionable information which can be used to automatically resolve issues before they have a business impact.

We have customers who equate the automation and guidance provided by our own ML based systems with having an additional infrastructure engineer on staff 24/7. This frees up IT staff to invest time in making use of the data they are storing and securing for their organisations.