Seagate’s Data Pulse study identified various areas within an organisation that will pose challenges in an organization’s AI journey, such as the lack of a clear strategy, lack of adequate infrastructure or budgets, or the lack of appropriate organisational talent. To maximize the potential of AI in their journey of adoption, it is important for organizations to overcome these challenges.

Most importantly, it is essential for organisations to first ensure a solid data infrastructure at the start of their journey in adopting AI.

After all, data is the lifeblood behind AI applications; rising AI adoption will lead to increase in data volumes. In fact, nearly all respondents in our survey said that further investments in IT infrastructure are much needed to address the increasing stream of data.

To help organisations catch up and address your questions around this aspect of AI, here are some insights and useful tips from our online panel of experts:

  1. For organisations adopting AI, what are their top challenges in terms of data storage infrastructure? What can organisations do to address them?

BS Teh, Seagate

According to our Data Pulse survey, respondents indicated some of their biggest roadblocks in implementing AI included poor strategy and direction, inferior IT infrastructure, budget, commitment from the leadership team, and a lack of talent.

If we zoom into data storage, the survey revealed nearly all respondents believe the growing use of AI demands greater need for data storage solutions (95%).

However, over 20% of respondents believe they are not ready to handle the increasing stream of data. 15% believe their organisation has not invested in the right data storage infrastructure to ensure they are AI-ready.

I’ve seen similar sentiments echoed among many of our customers, highlighting IT infrastructure as a key challenge. To grow their use of AI, they are looking to improve various aspects of their infrastructure surrounding data storage (which is why they come to us), as well as software applications, computing hardware and network equipment.

  1. From an infrastructure perspective, what needs to be improved to ensure effective implementation of AI, as well as the ability to scale and evolve in future?

BS Teh, Seagate

A recurring theme when it comes to any dialogue about adopting AI is the fact that data volumes are growing exponentially. According to the Seagate-sponsored IDC whitepaper Data Age 2025, the global ‘datasphere’ will grow to 163 ZB by 2025.

The rapid expansion and adoption of predictive analytics, machine learning and AI technologies will only increase global data volumes. The same report projects that data we use for AI will grow by a factor of 100(!) to 1.4 ZB. That’s a lot of data for existing IT infrastructure to manage, process and store.

For organisations to ensure they are able to leverage the wealth of data to effectively implement AI, three critical issues about the infrastructure needs to be addressed.

First, organisations require a consistent data management strategy that ensures the effective storage and organization of their data while maintaining its integrity and security.

Second, they need a robust data storage infrastructure that ensures capacity, performance, reliability, and scalability.

Third, they need a strong data security strategy as the risks and costs of breaches continue to rise.

Moving forward, we’ll continue to see an improvement in data analysis in terms of frequency, flexibility, and immediacy across industries and applications to drive strategic decisions and actions among organizations.

Much of this analysis is set to happen at the Edge, where AI, machine learning, and the increasingly urgent need for real-time responsiveness will drive the rapid expansion of Edge intelligence.

Overall, to be prepared for future evolutions of AI, organizations require their current IT infrastructure to be flexible and agile, accommodate large, varying volumes and types of data efficiently, be able to manage the full data lifecycle, and lastly, be largely autonomous.

Johnny Chou, Viscovery

To implement AI, an organisation must consider the computing power, storage capacity, and network bandwidth that its IT infrastructure can provide.

In terms of computing power, GPU resources provide the computing power for AI technologies. These are high-cost investments which organisations must consider how to use and share their resources efficiently.

As for storage architecture planning, organisations should consider the scenario of an immense increase and the migration in the volume of data.

If the organisation deploys AI in its cloud architecture, they should evaluate if the bandwidth and transmission speed of their cloud network can sufficiently meet the demand for rapid access to large amount of data.

  1. AI isn’t just for large enterprises. Do you have any advice to small-medium businesses who want to implement AI?

Akash Bhatia, Infinite Analytics

The application of AI is everywhere now. Small-medium businesses should focus on getting quality data to drive their AI – this will greatly determine whether they get an ROI from using the technology. Getting the foundation right is half the job done.

Also, do not be misled by hype and buzzwords. A lot of what is now labelled as AI has existed earlier as well – the only difference being that data has increased exponentially and  computational capacity has improved since.

After data, talent will be extremely crucial for AI implementation. Ensure that you have the right talent in place to recognize what is required in whatever AI process you intend to execute.

Johnny Chou, Viscovery

Small-medium enterprises (SMEs) can consider the basic AI framework offered by existing public cloud providers, such as AWS, GCP and Microsoft Azure. These platforms also provide the basic hardware for GPU computing.

However, SMEs must keep in mind that implementing AI is a long-term investment. As the volume of data being collected, analysed, and stored increase in tandem with time and business growth, SMEs should consider preparing their own hardware and server rooms.

  1. What potential downtime can I expect when implementing AI in my processes (or IT infrastructure)? What are some tips to minimise this and ensure smooth operations?

Akash Bhatia, Infinite Analytics

I don’t think a downtime is warranted. When Google Translate switched to AI, they did it without any significant downtime.

You could essentially just “switch over”. Ensure that you have adequate back up, before doing the switch. Ensure that the AI platform works flawlessly, before making the switch. Test, Test, Test!



Leave A Comment