Expanding the AI Value Chain Inspires Breakthroughs

“I am aible” is the new “Just do it.”

Nike’s famous tagline (Ad Age ranked it the No. 2 ad slogan of the 20th century) has energized a nation of athletes to lace up their shoes and believe in themselves. Now, Intel Corporation is doing the same, one tech partner at a time. Aible, a California-based generative artificial intelligence (GenAI) company is living proof.

Aible CEO Arjit Sengupta and I sat down on my show CXO Spice. “I am aible” is a rallying cry for his six-year-old startup. While many dream about enterprise-wide GenAI, Sengupta and his team are making it happen with more than 25 use cases in industries like healthcare, education, retail, legal and more – each deploying in less than 30 days.

Yet, Sengupta is not innovating in a silo. Intel Labs and Aible have worked side-by-side. “When you make these things that simple and that fast,” he said, “the art of the possible changes. What makes technology sticky is when anyone can do it.” In short, Sengupta is equipping people with the confidence to say: “I am aible.” He even likens GenAI as a superpower and the people using it as heroes.

Sengupta is a lifelong student of AI: he studied AI at Stanford University; founded an AI start-up called BeyondCore (later swooped up by Salesforce); wrote the saucily-titled book “AI is a Waste of Money: Why Most AI Projects Fail and the Secrets of Succeeding in AI;” and, if all that wasn’t enough, he co-created and taught the MBA course entitled “AI in Market-facing Functions” at the Harvard Business School.

Still, his company has about 50 employees – small compared to partners typically chosen by the Fortune 100. This, I believe, is a game changer at scale. Intel is expanding the value chain for AI by collaborating with companies and their customers for true breakthrough innovation.

Aible, Intel and Shared Stewardship

While Aible has large enterprise customers like Verizon, Intel and Aible’s shared stewardship really transforms the future of small to midsize companies. AI requires computing megapower. It’s a dangerous notion to think that AI is reserved for the largest companies with the largest budgets.

According to the Small Business Association, there are 33.2 million small businesses (defined as companies with less than 500 employees) in the United States – which make up 99.9% of U.S. firms. These more agile small and medium-size businesses are innovating at lightspeed.

To bring AI to more people and make a bigger impact requires making it usable for many. AI deployment and adoption can be costly. Intel worked with Aible all the way through the journey: development, product testing, optimization, support. I find Intel’s commitment from start to finish exciting, delivering two significant results: 55x total cost of ownership (TCO) reduction and deployment of GenAI within 30 days.

There’s a new world for innovation, one where silos are being replaced by side-by-side collaboration with partners and customers.

Cost Reduction in Total Cost of Ownership: 55x

When Sengupta first told me about how Aible, with the support of Intel Labs, had achieved a 55x cost reduction in TCO, I thought I had misheard him. But there it was, another proof point on how shared stewardship has helped create scalable AI solutions, a true breakthrough in the race to bring AI beyond a prototype.

Aible and Intel’s serverless approach leads to more cost-effective production in the cloud. Sengupta revealed that, after talking to many organizations, the overwhelming consensus was they “don’t want to move their data” and no matter what cloud or on-prem server they’re on, there’s an Intel processor.

A serverless approach means you pay only while your workload is running on an Intel processor, a major driver in achieving the 55x cost reduction in TCO. The server, he says, is then used by somebody else after your workload is done. This is Aible’s first serverless approach powered by Intel processors and AVX 512 technology. It’s secure because it’s coming through a virtual private cloud (VPC). This, again, expands the value chain by sharing resources across many organizations. AI is power hungry. Optimizing AI workflow from data ingest to deployment reduces cost. Collaboration with Intel enhances Aible’s ability to offer scalable AI solutions.

Not to be ignored are the many organizations with older iterations of servers and processors. But, can older processors support GenAI workloads? Sengupta says to look at the time between user request and response. He said, “Even using second generation processors, the response was less than a minute.”

As Aible’s team and Intel worked through the use cases, they also began to see a pattern of customer problems – from governance to accessing the right data for GenAI to user experience. They saw that deploying GenAI felt complicated for businesses. Aible decided to uncomplicate GenAI so that any business user could leverage their unique data.

GenAI in Less Than 30 Days? Yes.

I implemented AI for a global financial institution. It took 9 months just to gain internal alignment within the division piloting with us and even more time to actually train the model and test. The pages on my wall calendar flipped by as we discussed who would be involved in the initiative, what the test data set would look like and how big the data sample should be. We looked at the transcript of the call center, turning that into AI modeling where the bot picks up the phone and faces potential questions. So when I heard Aible was deploying in 30 days – and in some cases as little as five – I was shocked.

Computer power is much more advanced now, and Intel has played a significant role in its evolution, directly impacting data center bandwidth. At Computex in Taiwan, Intel CEO Pat Gelsinger unveiled the Intel® Xeon® 6 processor. With up to 80 cores and 160 threads, its exceptional efficiency and processing power uses less energy consumption for data centers and provides a much more powerful server to train data for AI.

Still, a 30-day or less deployment window is a departure from typical AI implementation where disparate technology, prototyping and testing of different settings and processors can take months – with no promise of success on an enterprise-wide level. Aible addresses all of that, Sengupta told me, before day one of deployment.

“When you have built your technology for that speed of implementation and speed of iteration, the results are outstanding,” he said. His team’s success rate proves this to be true.

A Learning Curve with Meaning

Intel and Aible broadened the value chain to include customers, tech innovators and the people actually using GenAI solutions for greater business impact. All this leads me back to the spirit behind Moore’s law: twice the performance at half the price. The shared stewardship of Intel and Aible proves that innovating without making efficiency unaffordable is possible. That’s not debilitating, but refreshing. That’s not reactive, but intentional. Collaborating all the way through the value chain ushers in a new era to democratize AI. When the idea of “I am aible” is embraced by more people – from high school and college students to people working in small businesses to those in large enterprise – technology becomes useful in everyday work and life.

My core belief is that growth resides at the cross section of technology and humanity. The vision: leverage technology to do better and enrich the world by making technology solutions affordable and accessible to more people. This is how Intel and Aible are creating breakthrough innovation.