Artificial intelligence (AI) is no longer a technology reserved for global enterprises like Google, Amazon, Apple, or Facebook. Without extensive resources and knowledge in the area, even smaller businesses can benefit from AI-as-a-service (AIaaS). As they have done previously with software, infrastructure, and web-based platforms (SaaS, IaaS, Paas), vendors are now offering AI algorithms and models as a service to automate your business.


What is AIaaS?

AIaas is any third party that offers artificial intelligence outsourcing. Using such services will allow companies to conduct experimentation with AI without any sizeable upfront investment and less risk to existing infrastructures. Typically, experimentation permits the sampling of public cloud platforms to test machine learning algorithms.

Different provider platforms have variations of AI and machine learning to suit the needs of the organization. Before opting for a service, it is crucial to understand the features and ensure it can do what you need. For example, if the business primarily uses Google products, the Google Cloud Machine Learning platform might be better suited than Microsoft or Amazon services, fitting the existing infrastructure.


Using AI-as-a-Service

Many companies realize the benefits of AI for their business. A Havard Business Review report shows that AI will add $13 trillion to the global economy in the forthcoming decade. The opportunity is nothing short of vast, with those using AI benefitting from greater productivity, a sustainable competitive advantage, and increased revenues. 

The problem is that companies struggle to go beyond the pilot phase of AI projects due to a lack of internal skills for implementation. AI is still in its infancy, and the number of experts in the technology and science behind is relatively small. The AIaaS market is expected to see a ten-fold growth in the next five years as it helps organizations in these positions. Instead of trying to find a person with the right skills, companies need to find people who can use AIaaS tools.

As well as reducing the skill gap, AIaaS negates the need for gathering, preparing, and using the appropriate data to train their software. An AI algorithm is only as good as the data that feeds it. With AIaaS platforms, algorithms can be deployed in different environments and use cases, with the results being adapted to develop more complex models.  The broad scale of AIaaS platforms allows AI to grow at a pace.


High level vs. Low-level AI

When discussing AIaaS, it is useful to distinguish between high and low-level AI. A problem for a high-level AI model might be face recognition. An algorithm’s task is well-defined, detecting if a face is in a picture, and therefore is relatively standardized. The answer will be either “yes” or “no” meaning anyone can understand it.

A low-level algorithm could be something like logistic regression. A model such as this can be applied to all kinds of different scenarios like churn prediction or fraud detection, for example. Sometimes, they may require a lot of pre-processing, training, and optimization. These more complex solutions need expert AIaSS platforms as the skills are not readily available in-house to test different methods.


Where AIaaS is best applied

AIaaS applications work best in sectors that have a high volume of digital processes like analyzing documents, emails, or other unstructured information. Complex scenarios can be handled by ready-made machine-learning algorithms that manage vast data points in real-time.

For example, Cinelytic is offering AI tools to the film industry as a way of predicting box office success. The platform uses machine learning to analyze performance data across box office, home rental, and pirated download transactions. The information will then be cross-referenced by genre, release date, and the cast of the movie.

Cinelytic can predict the potential lifetime value of any movie, which may be highly valuable at a time when audiences are starting to dwindle.

The New York Times uses Google AI to digitize five million historical photos. The “morgue” contains pictures going all the way back to the 19th century, many of which do not exist anywhere else in the world. Google AI stores scans of the image and any notes attached to it, allowing users to categorize the semantic information they contain. AI is making the preservation of the images richer and more accessible.


Democratising AI

The potential benefits of democratizing AI should not be underestimated. AIaaS allows businesses to benefit from new technology, regardless of expertise, size, and existing infrastructures. Everyone will have the ability to progress using AI and infuse it into day-to-day processes. While AI was once a tool for global enterprises to push further ahead, that doesn’t necessarily need to be the case with the accessibility of AIaaS.

The funding for AI startups is still increasing, with a 592% increase during the last four years. This alludes to an anticipated surge in AIaaS when these companies grow and mature into different market verticals. There will be plenty of platforms for IT leaders to try and overcome technology gaps, leading to an exciting future for everyone.


Case study – UK National Health Service (NHS)

The NHS is working with the AIaaS vendor Kortical to optimize supply and demand models for every hospital in England. With blood products, shelf-life is short as platelets only last for seven days. This means that hospitals always need to stock blood of different types, antigens, and collection methods to meet patients’ needs and potentially save lives.  

The challenge for the NHS comes in ensuring that hospitals have a blood supply at all times while mitigating the overstocking of platelets, which will ultimately expire after one week. There needs to be an understanding of supply, manufacturing, distribution, stock, logistics, and economics.

The NHS Blood and Transplant unit turned to AI-as-a-service to be more efficient at predicting the supply and demand for blood. The process to help reduce Adhoc transport costs and expiries included:

  • Building advanced machine learning models that can predict which donors would attend an appointment and how much volume they will donate
  • Predicting demand for different blood types
  • Processing over 700,000 variables to make sure the right products go to the right place and minimize transport and expiry costs
  • Simulating results to optimize the strategy
  • Building an AI-driven app for the NHS team to interact with and monitor


The Kortical Cloud API turned the data into an AI-powered application, delivering a 54% reduction in expired platelets, 100% reduction in ad-hoc transports, and a full delivery rate.



AIaaS to the perfect setting for testing AI without hefty investments or new technology requirements. Businesses can quickly generate proof-of-concept ideas without having technical knowledge and sustain a competitive advantage.

Many AIaaS vendors on the market offer products like speech-to-text, chatbots, face recognition, and supply-chain prediction, amongst many other solutions. As AIaaS is a new market, there does tend to be a lack of standardization and regulation, which businesses should be aware of when selecting a vendor. Some AIaaS parties still keep their models in a “black box,” making it difficult for people to understand. If you are looking at low-level AI solutions, research vendors carefully to ensure you can interpret their approach.

Disclaimer:  The author of this text, Jean Chalopin, is a global business leader with a background encompassing banking, biotech and entertainment.  Mr. Chalopin is Chairman of Deltec International Group,

The co-author of this text, Robin Trehan, has a bachelor’s degree in economics, a master’s in international business and finance, and an MBA in electronic business.  Mr. Trehan is a Senior VP at Deltec International Group,

The views, thoughts, and opinions expressed in this text are solely the views of the authors, and do not necessarily reflect those of Deltec International Group, its subsidiaries, and/or its employees.