Meet Dr. Nicolaus Henke

Artificial Intelligence (AI) could emerge as one of the most significant technical developments of our time. AI is growing rapidly. In the Artificial Index Report 2022 from Stanford University, they report that private investments in AI in 2021 totaled around $93.5 billion—more than double the total private investments in 2020. There is also a lot of evidence on how AI becomes both affordable and higher performing.

We had the privilege to have a chat with Dr. Nicolaus Henke, who recently joined the Silo AI Board of Directors. Mr. Henke is a Senior Partner Emeritus of McKinsey & Company and the former Chair of QuantumBlack, McKinsey’s advanced analytics and AI firm, which he co-led from small beginnings to a global client service. Additionally, he helped found and scale McKinsey Analytics, where he worked with leading clients on analytics and AI work across various sectors.

 

First of all, how much value is there in Ai?

“We did some work at QuantumBlack together with the McKinsey Global Institute to assess that, and we could see 15 trillion dollars in opportunity by 2030 if we would implement everything we know and have today within AI. This is a staggering number; it amounts to about 15 percent of the world economy which runs at about 95 trillion dollars.

We looked at 400 use cases and followed where money is or can be made. We looked across 19 industries where machine or deep learning could be used, the two biggest buckets of opportunities – approximately 5 trillion each, were within Production Supply Chain management and Marketing and Sales, or in other words to better understand the customer.”

 

 

You are one of most reputable and respected experts in the sector, what attracted you to Silo AI?

“What I like most about Silo AI is that they have a proven capability to work on projects that really create an impact. And they put machine learning into both digital and physical products. It could be a truck, robotic drills, or a mobile phone.

Silo really works with very advanced customers, typically CTOs, who know a lot about the AI-space but might not have the latest capability in specific areas of machine learning, for instance computer vision. Silo, on the other hand, with their couple of hundred of PhDs and professors can bridge that gap. Very few blue-chip companies can hire these people because they want to bring their expertise to many different industries and use cases, and would get tired to “only” work on a recommender system for music sales, for example.”

 

I have heard you say that implementing AI in larger organizations can be difficult. Why?

“Most companies are incredibly distributed, there are hundreds of places which you need to impact to really capture the full effect and make AI work. To put it into action is an almost unsurmountable task.

At QuantumBllack we could create stand-alone cases, such as the fastest boat for America’s Cup, which required years of work and preparation in new ways and simulations. But take a bank for instance, if half of the work for call-centers gets automated and a lot of your decision making gets supported by the machine, almost all jobs change. Changing how a job, whether it is boat building or banking, where you have 100 of years of know-how it is done, is difficult. Changing the algorithm is sometimes easier than making all the other changes which need to happen. That is why most companies take their time to implement machine learning.”

What can a company like Silo do to address that?

“First of all, Silo works with very sophisticated clients to build the AI into the product, just like Apple builds computer vision into the Iphone so the user doesn’t need to learn about AI. This applies to devices, but also smart machines like shipping fleets for example. And Secondly, Silo is supporting companies on their Machine Learning Operations. As an example, a retailer has a number of different AI-applications who works separately from each other, but they involve the same customers and will require the same data architecture and underlying technical logic – otherwise it will become a mess.” Silo helps to architect that and build the tools required to build, test, and monitor models in operation.

 

 

Fifty years ago we thought our homes would have a robot solving our everyday needs, but that never happened. Are we bad at predicting the future?

“We might not have a robot, but we have stand-alone tools – like a dishwasher or a microwave oven, a lawn mower and a vacuum cleaner that does what science fiction thought a kind of humanoid robot would do.

The same trend we see in AI today, most companies are offered standards, like a product for call-center automation or a customer recommendation engine. Those are products that a company easily can put into their technology stack.

If you make a map of all the machine learning developed today, more than 75% of them are solving one specific problem as a vertical software.  It doesn’t need to be always the best, it could also be the one that fits the customer’s stack the best.”

Finally, what are holding companies, or what is holding AI, back?

“We are dealing with new technology and some things that we try will not work. We must focus high-impact work, and to stay focused on what works in reasonable time.

Secondly, access to technical skills. There is naturally a lack of experts in certain deep areas, where we need certain skills data engineering, data science and machine learning operations skills combined with for examples text or computer vision skills to interpret large books or input from cameras. We hope Silo can help here.

Third, translation and leadership skills in particular at the top of the house. We need people who are connecting the dots of AI to business – business translators, explaining to the business-people where opportunities might lie and helping the deep technologist to focus on the highest value problems. I have seen many very impressive boards and leadership teams for example in banking, mining, aerospace, automotive, pharma who developed a quite deep perspective on how AI can help improve or even totally change their business over time. Those leadership teams can set the right ambition and pace for appropriate investments into AI, avoiding the “hypecycle” but working to a new future, for example automating everything what happens in a mine under earth, developing the next generation of batteries, finding new medicines and testing them virtually,improving the situational awareness of shipping fleets or even automating them. Each of those examples is a vision which changes the economics of an industry, and improves growth and productivity on the way.