Landing AI is an Artificial Intelligence (AI) company responding to the challenges businesses face in adopting AI with the goal of realizing practical value.

As with any new tool, businesses face the challenge of answering ‘how do we’ questions. However, the technical complexity of implementing an AI project along with the anticipated high return on investment makes answering these questions both daunting and difficult. What should we tackle first? Acquiring technical knowledge to assess what is, and is not, technically feasible is difficult. This is further exasperated given the complexity of the technology and current wide-spread misunderstanding about what AI is today, relative to the perceived end-state of AI.

Where’s the data and can we use it? Acquiring large, high-quality datasets to train and operate, effectively, requires both an understanding of data sources and techniques to manipulate or condition data. It’s worth noting many companies already have troves of data in their data warehouses - being able to assess the value and use cases for the data is the mystery companies will need to solve.

What’s it going to cost? The high cost of implementing bespoke AI solutions tends to push leaders to choose projects with large returns ($1M and up). This leads to joyful business outcomes like: layers of management oversight to control costs and ensure returns, missed deadlines, paralysis-by-analysis and ultimately, failed grand-plans. Landing AI suggests businesses start small ($50 - $200k) and focus on growing their technical capabilites.

Are we going to create ethical and regulatory problems? Data privacy, compliance w/ AI law and regulations and bias in algorithms need to be assessed and managed to ensure AI soltuons are transparent, compliant and fair.

Is our culture, objectives and structure such that we can succeed with AI? Integrating AI into established business processes requires, typically, significant changes to company culture, objectives and organizational construct.

Landing AI seeks to offer an AI platform to help developers more easily build and deploy AI solutions. While the company is targeting the following industries: automotive, agriculture, electronics, food & beverage, medical devices, infrastructure, pharma, EV manufacturing, manufacturing and life sciences, the focus of this study is on their approach to bringing their AI solution methodology to the manufacturing sector.

In the manufacturing sector, the company is focused first on a product called LandingLens. Andrew Ng said, in an interview with Joe Miller of protocol.com on May 18 2021:

We ended up building LandingLens, which is a data-centric, MLOps platform for computer vision. We help companies — starting in manufacturing but we have interest in other computer vision vertical applications — be 10x more efficient and often much more successful as well in building and deploying computer vision systems. (https://www.protocol.com/enterprise/andrew-ng-ai-strategy)

In summary, Landing AI seeks to make computer vision easy for a wide range of applications across all industries, starting with manufacturing.

Landing AI was founded in 2017. The company is based in Palo Alto, California.

Andrew Ng is the sole founder of Landing AI. He serves today as the company’s CEO.

Andrew Ng was the co-founder of Coursera, former Chief Scientist of Baidu, and founding lead of Google Brain. He founded deeplearning.ai and is an Adjunct Professor at Stanford University. In 2013, he was named to the Time 100 list of the most influential persons in the world. He is regarded as a pioneer and globally recognized leader in AI. https://www.andrewng.org/

The idea for Landing AI seems to have originated with Mr. Ng’s recognition that businesses need to adapt and restructure themselves to take full advantage of AI technology. He founded Landing AI in 2017 to help companies navigate that transformation. (https://venturebeat.com/ai/google-brain-cofounder-andrew-ng-unveils-landing-ai-to-bring-intelligence-to-manufacturing/)

Now, Landing AI is uniquely positioned to lead the development of AI from a technology that benefits a few to a technology that benefits all. (https://landing.ai/)

The company’s focus is on machine vision systems. In a manufacturing environment, these systems process images to aid with part identification, sorting and quality control; robotic pick-and-place and optical character recognition for labeling and serialization.

Series A funding round organized by McRock Capital included Intel Capital, Samsung Catalyst Fund, Insight Partners and Canadian Pension Plan Investment Board.

Since, AIFund, Taiwania, WAISN and Drive Catalyst have invested in the company.

In all, $57M has been invested in Landing AI

The private company is believed to be realizing revenue from it’s first product, LandingLens. Stanley Black & Decker and Foxconn are among the company’s first customers.

(https://www.cbinsights.com/company/landingai/financials)

Landing AI is solving the problem of scaling AI.

The company’s approach is to offer an AI platform (with ready-built tools) that enables customers to develop, deploy, track, and maintain their AI solutions.

LandingLens, the company’s flagship product uses cutting-edge AI technology to “enable computer vision applications to be built and deployed 10x faster than before, thereby creating significant ROI. (https://landing.ai/careers/)

Landing AI is focused on the manufacturing sector. Their target customers are manufacturers.

From a market or industry sector perspective, Landing AI is focused on sub-industries within the broader manufacturing industry which can benefit most from AI-powered machine vision systems. The company lists these manufacturing sectors as their target industries on their website: automotive, electronics, food & beverage, medical devices, life sciences, agriculture, manufacturing, infrastructure, pharma, EV manufacturing.

According to the US National Institute for Standards and Technology, the manufacturing industry accounts for 12% of total US GDP (2021), equal to $2.3 trillion. The industry employs 14.7 million people.

(https://www.nist.gov/el/applied-economics-office/manufacturing/total-us-manufacturing/manufacturing-economy/total-us)

According to the company’s website, Landing AI is pioneering the so-called Data-Centric AI movement in which “companies with limited data sets can realize the business value of AI and move AI projects from proof-of-concept to full-scale production.”  The company sees ‘data-centricity’ a key part of their overall value proposition.

A focus on using data along with deep learning, vs the traditional rules-based approach to determine an outcome produces systems which excel in subtle or nuanced inspection problems. Ultimately, higher accuracy is realized. As an example, false-rejections decrease, improving business performance.

In general, this is not a novel approach to adding AI to machine vision systems, so it would be unrealistic to say the company’s data-centric approach is a unique competitive advantage.

Perhaps, the most important unique competitive advantage the company has is that Dr. Andrew Ng is a prominent figure in the AI field, posessing a strong track record. Dr. Ng’s reputation and network connections should enable the company to establish strategic partnerships ahead of their competitors. Collaborations with industry leaders in manufacturing can lead to proof points that can accelerate market share gain and decrease customer-acquisition costs.

Landing AI is a realtively new private company using cutting-edge AI technology. And, the cutting-edge in the AI field is moving rapidly and in multiple directions.

Given that their flagship product, LandingLens, is an AI-powered machine vision system. As such, it’s likely the company is leveraging, at least, the following AI technologies to power their solution.

Deep Learning: Deep neural networks are commonly used for image analysis and object recognition. They can learn to detect and classify objects, patterns, and features in images.

Machine Learning Algorithms: Machine learning models, such as support vector machines (SVMs) and decision trees, can be applied to various machine vision tasks, depending on the specific problem and dataset.

Image Processing: Image processing techniques like filtering, edge detection, and thresholding are often used to enhance and preprocess images before feeding them into the AI algorithms.

Object Detection: Object detection algorithms are used to identify and locate objects within images or video streams.

Optical Character Recognition (OCR): OCR technology is used to recognize and extract text from images, enabling the conversion of printed or handwritten text into machine-readable data.

Pattern Recognition: Pattern recognition algorithms are used to identify and categorize recurring patterns or motifs in images, which is valuable in quality control, industrial inspection, and document analysis.

Landing AI is in the Artificial Intelligence Field. Within this field, they are focused on AI Machine Vision.

Machine vision systems, have been advancing rapidly in recent years.

These systems are a combination of camera technology, image processing software and artificial intelligence technologies.

Focusing on camera technology, improvements in recent years have followed the long-term trends enabling greater accuracy and more reliable inspections. Higher resolution, faster frame rates, and improved sensitivity to low light are, generally, the focus of camera manufacturers.

In the field of Artificial Intelligence, innovations in Deep Learning are enabling machine vision systemes to learn and adapt, making them more capable of handling complex tasks like object recognition, defect detection, and quality control

Image processing technologies are trending from 2D to 3D. 3D enables measurement of object dimensions and more accurate inspection of complex, three-dimensional shapes.

Other interesting advancements include:

Integration of machine vision systems with industrial robots allows for automated inspection and quality control. These systems can guide robots in picking and placing objects or perform inspections on the manufacturing line during the manufacturing process. Edge computing, where data is processed closer to the source of the data is gaining traction in machine vision. The benefit of decreasing the distance between data source and compute location reduces latency, critical for real-time applications.

According to CBInsights, a company that provides a business analytics platform and global private company database, Landing AI’s top competitors include:

Metaspectral: provides ingiths using AI visible-to-infrared imagery for identifying materials and chemical composition.

Aiir Innovations provides AI visual inspection software services.

Darwin AI is focused on Ai-powered visual quality inspection.

Landing AI was founded in 2017, just six years ago. In February of this year, the company began offering to anyone LandingLens for free. In doing so, the company claims they are democratizing the creation of AI for companies large and small.

Landing AI is a proponent and advocate of businesses using limited datasets that most companies already have. This ‘data-centric’ approach means businesses can implement AI machine vision faster and at a lower cost.

The impact of the ‘start small’ and ‘data-centric’ approach should be much faster adoption of this particular AI technology, with others to follow. Placing easy-to-get-up-and-running AI capabilities in the hands of more employees flattens the learning curve, enabling faster realization of practical value. In this way, Andrew Ng is achieving the original idea for Landing AI.

From an AI Machine Vision System perspective, accuracy is the most important metric. Other metcis include precision, recall or sensitivity, false negative rate, and specificity.

I can not say.

I can not say.

According to the World Robotics report, there was an all-time high of 517,385 new industrial robots installed in factories around the world, an increase of 31 % compared with 2020. This brings the stock of operational robots around the globe to a new record of about 3.5 million units. International Federation of Robotics, World Robotics Report, Dec 19, 2022.

Meanwhile, lack of skilled labor is a strong driver for automation across many industries, worldwide.

Combining these realities leads to the conclusion that Landing AI is in the leading position to leverage it’s capabilities to further advance their mission of making it easy for ROBOT companies large and small to adopt AI.

The next logical extension from AI machine vision in the manufacturing industry is AI machine vision for robots used to automate factories. Therefore, Landing AI should offer, next, products and services for robotics, starting with robots that enable manufacturing automation.

Landing AI would benefit from exploring the intersection of robotics, manufacturing automation and AI because doing so leverages their presumed success in the manufacturing sector with their data-centric approach to AI and their flagship product, LandingLens.

Landing AI has established a data-centric, AI-platform approach that make it possible to implement AI without a dedicated Machine Learning team.

To at least some degree, the manufacturing industry has responded positively to the company’s offer, with more than 1,000 individuals signing up to use LandingLens before the company offered it for free to everyone in February 2023. It stands to reason, the adoption rate of LandingLens with a $0 price should increase, further advancing Landing AI’s presence (and value) in the manufacturing industry.

The intersection of Process Automation, Manufacturing and Robotics is a synergistic one. Advancements in one area often spur advancements in the others.

According to the International Federation of Robotics' 2021 World Robot Report, industrial robotic installations have increased by 10 percent, and service robots' use has risen by 12 percent, leading to a nearly doubled global robotic density. Advanced vision systems are a key driver behind this growth, enabling highly automated and networked industrial processes.

AI machine vision technology may allow robots to operate in real-time, eliminating the need for a physical definition of the product they interact with and enabling them to adapt to different tasks. Vision-enhanced robots offer greater versatility and improved interoperability in factory settings, aligning with the concept of Industry 4.0-connected factories.

According to the International Federation of Robotics, the following technologies require advancement to realize an increase in autonomy.

Scene Understanding Product: 3D scene understanding for robots, enabling perciption of, and naviation through, complex environments. Service: Integration and consulting services for robotics companies to implement 3D scene understanding in their robots.

Robotic Navigation and Control Product: Navigation and control software for robots that make use of AI machine vision to avoid obstacles, follow predefined paths, and make good real-time decisions. Service: Services to optimize navigation algorithms to enhance the efficiency and safety of robot movement.

Human-Robot Collaboration Product: Machine vision systems that enable robots to work alongside humans safely. Service: Establishment of ethics and safety guidelines for human-robot applications.

In order for a technology to be appropriate for a solution, the technology needs to serve the needs of the application and it’s long-term evolutionary vector.

In this case, these technologies are fit for purpose and scalable, as they are aligned with the specific needs and requirements of advanced robotic manufacturing automation. And, the technologies are cost effective - certainly when combining the the decline in sensor and AI technology cost over time with the cost savings associated with automation. In addition, the core technologies in AI, machine vision and robotics are now are stable, secure, reliable and compatible with existing systems.

(http://www.landing.ai) (https://www.andrewng.org/https://postindustria.com/why-everyone-is-seeking-ai-engineers-an-emerging-trend-that-is-here-to-stay/) (https://venturebeat.com/ai/google-brain-cofounder-andrew-ng-unveils-landing-ai-to-bring-intelligence-to-manufacturing/) (https://datacentricai.org/) (https://www.sciencedirect.com/science/article/abs/pii/S0957417422024757) (https://techcrunch.com/2021/11/08/landing-ai-machine-learning-operations-tools/) (https://ifr.org/ifr-press-releases/news/wr-report-all-time-high-with-half-a-million-robots-installed) (https://www.automate.org/new-product-news) (https://www.crunchbase.com/organization/landing-ai) (https://www.cbinsights.com/company/landingai/alternatives-competitors) (https://www.protocol.com/enterprise/andrew-ng-ai-strategy) (https://www.cbinsights.com/company/landingai/financials) (https://www.nist.gov/el/applied-economics-office/manufacturing/total-us-manufacturing/manufacturing-economy/total-us) (http://www.google.com) (https://postindustria.com/why-everyone-is-seeking-ai-engineers-an-emerging-trend-that-is-here-to-stay/)

Grade 95 / 100. Grader: Maitree Maniar, Nov 22, 2023 at 2:40pm.

My Response, here, to Maitree: That's quite gracicous, thank you, but I know it's scored with respect to limited available time for the challenge. And thank you for the insightful comments. A fine outline for a better business study, it has become!