BrandPost: Initial Results of the Intel and Aible Benchmark and Case Studies Report Released
Earlier this year, Aible, the only enterprise artificial intelligence (AI) solution that guarantees impact in one month, announced the initial results of the Intel and Aible Benchmark Study, a collaboration that is helping enterprises fast-track benefits from advanced analytics and AI, while also evaluating server vs. serverless architecture.
According to MIT-BCG, only “a mere 10% of organizations achieve significant financial benefits with AI.”
The Gartner report, A CTO’s Guide to Top Artificial Intelligence Engineering Practices, published 29 October 2021 states, “AI projects are characterized by high failure rates and take a long time to move from pilot to production. Slightly more than 50% make it from pilot to production, and those take an average of nine months.”
In this same market, Aible has delivered significant results for every customer in this benchmark study in 30 days or less.
Success stories included:
- A Fortune 500 technology company used Aible to identify actionable insights for sales opportunities in a matter of 29 days.
- Another Fortune 500 Healthcare Provider found new insights in Social Determinants of Health (SDoH) data with a 20X improvement in speed to insight in 15 days.
- Nova Southeastern University discovered paths to potentially improve student retention by 17% by utilizing AI from Aible in just 15 days.
- A multinational CPG company used Aible to identify ways to drive $10M in additional sales in just 17 days.
- A global food company identified ways to reduce food wastage by more than 10% in 27 days using Aible.
- A global manufacturer identified ways to reduce the impact of late shipments by more than $4M annually in just 17 days using Aible.
- A Leading Food & Beverage services company used Aible to identify actionable patterns and ways to improve sales efficiency by 5% in just 13 days.
- Another leading University enlisted AI from Aible to mitigate student attrition by 12% after only 30 days of analysis.
The detailed report with case studies can be downloaded here.
“Intel is helping us change the art of the possible in AI. Through the ‘Impact from AI in 30 days’ program that we rolled out with the Intel Disruptor Innovation Initiative, we were able to prove the capability of our Intel-optimized technology and have already collected success stories from a dozen customers, including two of the Fortune 500. Plus, more success stories are coming in every week,” said Arijit Sengupta, Founder and CEO of Aible.
“When we first engaged with Aible in the Intel Disruptor Innovation Initiative, we challenged them to prove that they could indeed generate business impact from AI in 30 days,” said Arijit Bandyopadhyay, CTO of Enterprise Analytics & AI, and head of strategy of Enterprise & Cloud, DCAI Group at Intel Corporation. “They have certainly lived up to that challenge, often generating value from actionable insights in less than 15 days. In fact, much of the time in these projects was spent on waiting for data or for customer feedback. The actual analytical work was completed in a few days.”
The Intel and Aible collaboration enables enterprises to leverage the combined power of Aible’s AI automation and Intel’s superior architecture. Aible automates and eliminates all the data science and dev ops complexities, so end users quickly see massive performance gains securely in their own cloud accounts. The built-in AI acceleration of Intel Xeon Scalable processors delivers peak performance and efficiencies across the entire machine learning lifecycle.
Serverless is superior
A key technological strength of Aible is its serverless-first approach, which enables it to train machine learning models far faster than other solutions in the market today (minutes instead of days) and at a significantly lower cost. The Benchmark Study has validated the serverless approach across these areas:
- In job cost comparisons, serverless computing was found to be 2-3 times more cost-effective than server architectures for comparable dataset sizes and workloads.
- When it comes to total cost of ownership (TCO), serverless is 3-4 times more cost-effective than servers.
- In end-to-end elapsed time for training models, serverless was found to be 2-3 times faster than servers.
The Intel and Aible performance benchmark and case studies report is available here.
Lessons from the Intel and Aible Benchmark and Case Studies Report
Some thematic elements have already become clear that explain why the Aible “30 days to AI value” approach is effective:
No one has perfectly clean data: In almost every case, the original dataset was not perfect and thus had to be updated several times before the project could be completed. In minutes, Aible Sense automatically evaluates whether the data has signal, and this validation made it easier for the project teams to “fail fast” and iterate until they got to the right data. Aible Sense also automatically adjusts for many common data quality problems, plus automatically recommends derived variables (also called “features”) to improve the impact of predictive models.
Business user involvement is key: In several cases, business users suggested changes to the dataset, requested more focused analysis (country-by-country, for example), or even changed the use case (from demand forecasting to overstock prevention in one case). Aible Explore enabled the project teams to engage with business stakeholders much earlier in the project by collaborating on an open-world exploration of the data. Because teams didn’t have to wait for the predictive modeling to be completed before accessing useful insights and could get business stakeholder feedback earlier in the project, they avoided having to change the project after investing months of effort.
AI must be firmly grounded in business realities and objectives: In every project, understanding the business objectives more precisely was key to delivering value. In one case, the customer’s key focus was helping their salespeople make their first sale as quickly as possible. They had significant constraints on how much effort they could spend on training their salespeople and thus determined that the optimal action was to identify the best candidates for the costly coaching investments. Aible Optimize automatically ensures that the predictive models are optimized to deliver positive business impact, while considering business objectives and resource constraints.
Click here to see additional Intel and Aible Benchmark customer case studies as they become available.
- What is Quarto? RStudio rolls out next-generation R Markdown
- RStudio changes name to Posit, expands focus to include Python and VS Code
- Fedora ditches CC0 'No Rights Reserved' software over patent concerns
- CrowdStrike enhances container visibility and threat hunting capabilities
- How observability tools help with legacy software
- Why do businesses suck at using data?