At Sentient.io, we are keenly aware that one of the major challenges companies face when adopting AI is that it is very difficult to find the relevant AI skillsets. In order to meet this challenge, Sentient has developed an AI Platform that enables software developers who have less AI skillsets to consume our ready-to-use AI models as APIs in order to accelerate AI adoption. 

Challenges to the wide adoption of AI - According to a Gartner report, Survey Analysis: AI and ML Development Strategies, Motivators, and Adoption Challenges, enterprises faces the following challenges when adopting artificial intelligence:

• Lack of deep technical knowledge - More than 56% of enterprises surveyed in the referenced Gartner Report cite this as the top challenge in adopting AI. This aligns with the fact that acquiring such specialised skill sets takes a long time.

• Lack of industry relevant algorithms - Every industry is different and the problems faced by enterprises within their domain are also different. Although AI has advanced deeply in many disciplines such as computer vision, there is still a lack of industry relevant algorithms to solve domain specific problems. 34% of the enterprises surveyed in the referenced report cite the lack of high quality data in their domain as a key barrier to developing AI algorithms that are suited to their needs.

• Significant investments needed to keep up with the AI Evolution - AI is evolving at an exponential rate. The OpenAI NLP model called GPT-3 was trained on more than 499 Billion tokens (words), and cost $4.6m dollars to train. Significant investments are needed to collect data and train large scale models. Albeit GPT-3 is a general NLP model, smaller domain specific models need to be maintained and updated regularly as well. The complexity of creating and training AI models together with developing an AI production pipeline will overwhelm enterprises especially when engaging in such activities is not their main business activity.