Overcoming AI obstacles: Learnings from AI practitioners within the Enterprise – Cyber Tech
In our steady effort to know and help our clients higher, we commonly conduct end-user surveys. Just lately, we sponsored a examine with IDC* that surveyed groups of knowledge scientists, information engineers, builders, and IT professionals engaged on AI tasks throughout enterprises worldwide. Our objective was to establish the highest challenges they face and the very best practices of extra mature AI groups. The examine revealed that enterprise AI challenges range with the maturity degree of the groups as they start to check and operationalize AI and GenAI. Nonetheless, a number of key findings stood out as persistent points throughout all maturity ranges which are actually informing our method to growing clever information infrastructure in help of AI.
One of many high findings was that 63% of respondents indicated that right-sizing storage for AI wants main enchancment or a whole overhaul, with storage bottlenecks being a persistent drawback in slowing down AI modeling. To handle this, we have now partnered carefully with NVIDIA to qualify options for mannequin coaching and have optimized ONTAP for SuperPOD qualification, which is at present in testing. Our CEO not too long ago introduced a brand new information infrastructure designed for the AI period, which is able to independently scale capability and efficiency to deal with the wants of the biggest foundational mannequin growth and scale down for inferencing. This infrastructure will run each on-premises and as software program on the planet’s largest public clouds, providing ONTAP information administration companies essential to environment friendly and accountable AI.
One other important discovering was that respondents cited information entry on account of infrastructure restrictions because the #1 explanation for AI challenge failure. To sort out this problem, we’re centered on simplifying and unifying information storage to help higher information entry. We offer a single information and management airplane throughout the NetApp information property, spanning the sting, information middle, and public clouds. This method natively helps all information codecs and environment friendly information motion, bringing information to the proper sources both on-premises or in one of many hyperscalers for every stage of the information pipeline. Moreover, we expose our capabilities to the instruments information groups use, similar to AWS SageMaker, Google Vertex, and Azure ML Studio.
The third key discovering was that solely 20% of respondents have mature, centralized insurance policies for information governance and safety for AI. That is an space the place we’re putting important emphasis going ahead. NetApp differentiates itself with policy-driven safety on the information layer, using constantly up to date AI/ML fashions to detect and reply to threats in actual time with 99%+ precision. This may be utilized to defending each information and fashions throughout the AI workflow. We’re additionally growing instruments for information scientists to get their work carried out safer and quicker, in compliance with privateness legal guidelines. These instruments will speed up and simplify information discovery and curation, present assurance of secured and compliant AI, assure accuracy and traceability, and combine with information science workflow instruments.
In conclusion, the insights from the IDC survey have considerably influenced our planning and method to growing AI-ready clever information infrastructure. By addressing storage bottlenecks, enhancing information entry, and enhancing information governance and safety, we’re higher positioned to assist our clients leverage AI and GenAI successfully. For extra particulars, you’ll be able to:
- Watch this dialog between NetApp’s Gabie Boko, CMO & Hoseb Dermanilian, International Head of AI GTM & Gross sales.
- Entry the IDC examine to self-assess your group’s AI maturity and be taught greatest practices
- View the NetApp-IDC-NVIDIA webinar to dive deeper into the survey outcomes
- Perceive our CEO’s AI imaginative and prescient from the current Perception Convention
*N=1220 – IDC Transformation Examine, Mar 2024 |