Override your instincts and find your way with AI


Posted on: April 25, 2019 by Jonas Bull

Life in an AI lab

My role at Atos is to connect our data scientists and engineers to business processes and bottom lines using the best technology available. The 1's and 0's don't have meaning outside the operational context, and operations need to get the 1's and 0's in a way that can be used to drive improved decisions, or safety, or otherwise create value. Making those connections, and working with some of the best, most creative people from Atos and from our clients is the highlight of my day.

Data skewed your playing field

Just a few years ago, you could run a business with a spreadsheet and your instincts. Remember those days? Maybe you didn't have the most efficient methods, and you took some losses from unexpected downtime, or you had to take on some unknown risks. But those were costs of doing business, and your competitors were facing the same losses, the same sorts of downtime, and the same risks. That was then.

The point where anyone could be an ostrich (head in the sand) about the volume of data in the world is past. We live in an amazing time where we have more data points about a wider variety of things than ever existed before. And it is increasing. Paul Challoner, VP of network solutions at Ericsson, recently stated that they saw 79% growth year-over-year in mobile connections. IDC reports healthcare data will reach more than 2,000 exabytes by next year. Connected cars — not self-driving, just regular newer cars — will send 25GB per-hour per-car to the cloud. Add to that smart meters, industry 4.0, smart homes ... That's raw data.

One way or another we will have to get smarter.

The time to follow fast is past

So imagine yourself as that spreadsheet-driven business owner, or CTO, watching as every aspect of your business — every cooler, every vehicle, every hospital bed — is flooding you with data. You see your competition making announcements about strategic AI or machine learning initiatives. And you're thinking, "maybe I can wait until AI matures." But that data is still piling up. And the beds are losing money. And the trucks are breaking down. And the coolers are running too hot.

For businesses, "fast follower" may be too slow when it comes to adopting AI. Data scientists aren't commodity staff. They are hard to find, hard to vet and hard to hang on to. Advanced analytics aren't plug and play. Most importantly perhaps, understanding your own use cases and knowing how to use your data to turn questions into value, requires some knowledge and processes.

Fast following is an outdated instinct. It takes time to develop these competencies, and time to lay the other foundations.

Three conditions for AI success

Thierry Breton, Atos Chairman and CEO, gives three conditions necessary for success with AI, which I will take the liberty of reordering:

  1. We must have the computing power to train specific algorithms to execute and learn complex tasks. That means combining the locality and speed of supercomputing platforms with the flexibility and scaling of public cloud to deliver a hyper-converged solution capable of tackling extremely complex tasks.
  2. We must use our knowledge and intelligence to define the right use case. That means identifying the right problems, the right solutions, and the right way to connect the solutions into the existing infrastructure.
  3. The final critical condition is trust. Everyone wants the benefits of AI and advanced analytics — but everyone also fears what that data and those analytics can reveal. Your AI program must do more than just comply with regulations. You must ensure that the data at every stage is handled securely with integrity and propriety.

Imagine yourself as the business owner again, looking at a tsunami of data, and you know it is time to make a move. There is risk no matter what choice you make, but the one choice you know will leave you behind is doing nothing.

This is where AI labs like mine come in. We bring together cross-functional teams and business- and subject-matter experts from your organization with our data scientists, architects and business experts. And we work together in a creative environment to discover what can be done, design solutions and define a roadmap to move you forward confidently on your AI journey.

An AI lab can help you find your way. I'm excited to be a part of that.

At the Atos Technology days, we’ll offer a personalized view of how #AI can benefit your business. Learn more on the event here.

Share this blog article


About Jonas Bull

Head of Architecture, Atos North America AI Lab

Jonas Bull is a consultant, solution architect and strategic problem solver with 18 years’ experience planning and executing infrastructure for business, analytics and data science. He has a reputation for being resilient, cool under pressure, and unafraid of complications and uncertainty. He is deeply collaborative, knowledgeable of technology, and a curious observer of human nature. Jonas holds a B.S. in Anthropology from the University of Southern Mississippi.



Follow or contact Jonas