AI humble beginning was first introduce in 1956 at a workshop in darthmoute college, by 1969 McCarthy and Hayes start to make a clear distinction and make it into a distinction. At first they are trying to solve “what kind of facts about the world are available to an observer with given opportunities to observe, how the facts can be represented in the memory of a computer, and what rules permit legitimate conclusions to be a drawn form these facts”. And now as the year 2021 AI has managed to solve not only what kind of facts about the world that represent the memory of computer, they can even launch predictive knowledge beyond human speed and intelligent that even able to make it’s own decision to optimise for a goal.
Despite its humble beginning AI has evolve, making a subset as what we know as Machine Learning that aims to give computer the ability to learn over and over again, to improve predictive models, and find insight without being explicitly programmed that way.
With that in mind, one do wonder how’s the AI and ML industry evolving from 1969 to the current 2021 post covid era?
Globally AI & ML verticals was forecasted to grow at 22.3% CAGR from 67.1 billion USD in 2020 to 131.5 billion USD by 2023. Including horizontal platform of AI enabled PaaS, that layers and empower technical end user to make use of specific products. Vertical platform of solution that address specific problems within the industries with varying level of AI & ML integration.
The Adoption trend has always be that some company integrate the AI in mostly pilot stages, but only 11% gain significant financial benefits (MIT Sloan Management Review and BCG October 2020). While other don’t have direct ROI metrics.Despite that COVID19 has show acceleration of AI adoption especially in supply chain and customer facing automation, AI and customer service. Some other interesting verticals also include sales and marketing automation, and HR automation all thanks the lockdowns in multiple country all over the world.
South East Asia
For South East Asia Kearney (Racing towards the future : AI in SEA) predicted that if applied and executed well, AI could add $1 trillion to SEA GDP by 2030. With Vietnam leading the way of investment in region.
New Emerging Winner Globally :
GPT-3 or more known as Generative Pre-trained Transformer 3 (GPT–3) is an autoregressive language model that uses deep learning to produce human-like text (Wikipedia) is the new emerging winner for NLP based startup.
By itself it have open a gateway to massive new wave of startup that leverage the OpenAI massive neural network via the distribution partner Microsoft. Wit most opportunities emerging form the collaboration of sales & marketing data integration, SME’s fraud prevention, predictive maintenance, and many more.
With Plug and Play, Entrepeneur First, and SGInnovate lead the investment in SEA AI & ML platform with 66 of the deal are made in 2020 alone a drop from 86 deal in 2019. Some emerging new company include Aampe, Brda, Origin Health, Surge Analytics, and Nextbillion AI in 2020 are mostly based in Singapore
New AI Public Market Performance Winner :
Recent IPO’s and AI integrated unicorn has show market high growth with more rational revenue multiples that their previous AI counterpart (Lemonade, etc), those company are :
- C3.ai is a leading enterprise AI software provider for accelerating digital transformationprovides comprehensive software solutions and services for a myriad of large companies, including 3M, Royal Dutch Shell, Raytheon, Baker Hughes and conEdison.
- Palantir (AIaaS) as the clear winner form data analytics, as already infamous model that support AI model management and compression, both leading challenges of MLOps and falls into data integration platform.
- Accolade (Personal Health) that give employee benefit recommendation and the data sources used for machine learning including insurer , digital health app, and provide.
- Sumo logic (AIOps) for AIOps Platform that predicts IT anomalies and the data sources for its prediction.
AI companies also bear more costs than conventional SaaS startups in product development:
- Data labeling: Andreessen Horowitz has estimated this cost to reach up to 15% of revenue in cases where customer data must be prepared for modeling.
- Cloud computing costs: Andreessen Horowitz has estimated this cost to reach up to 25% of revenue for AI & ML startups.
- Customer model integration costs : Professional services costs required to test and deploy models.
- Separate teams of data scientists and AI developers in addition to application developers, DevOps engineers, and IT operations teams: AI developers are scarce and costly and do not displace more conventional software engineers.
These costs can escalate unpredictably based on the success of model training and complexity of model deployment. The development of a single model can amount to over $1 million, making them multiple times more expensive than SaaS apps and increasing the capital requirements to seed an AI & ML startup