Listen to this story
Salesforce has consistently ranked as one of the leading CRM providers globally. According to Grand View Research, the global customer relationship management market size, valued at USD 52.4 billion in 2021, is poised to grow at a CAGR of 13.3% from 2022 to 2030. To cash in on the opportunity, Salesforce is leveraging AI and ML.
Sign up for your weekly dose of what's up in emerging technology.
The AI revolution will not only transform the consumer world but also change the work world as well, said John Ball, then GM of Salesforce Einstein, while introducing Salesforce Einstein in 2016. Salesforce claimed that, with Salesforce Einstein, companies will be able to make better predictive and personalised customer experiences across sales, service, marketing, commerce etc. Einstein has applications across various domains.
- Sales Cloud: The customers can use Einstein to analyse data and predict which leads and opportunities are most likely to convert. Einstein Activity Capture syncs email and calendar to Salesforce automatically along with a prebuilt activity dashboard.
- Marketing Cloud: Einstein can help deduce customer choice better with predictive insights drawn from their marketing engagements, brand interactions, and conversations across social media. Businesses can also create personalised messages and content based on customer preferences and intent.
- Service Cloud: Salesforce’s AI-based chatbots can work on checking
claims status or modifying orders using NLP on real-time channels like chat and messaging.
In 2018, Salesforce open-sourced its machine learning tool, TransmogrifAI- the engine powering Salesforce Einstein.
Merlion is an open-source Python library for time series intelligence. It comes with a standardised and easily extensible framework for data loading, pre-processing, and benchmarking for time series forecasting and anomaly detection tasks. The models here include classic statistical methods, tree ensembles, and deep learning methods. Merlion also comes with an evaluation framework that simulates the live deployment and re-training of a model in production.
This library aims to provide engineers and researchers a one-stop solution to rapidly develop models for their specific time series needs and benchmark them across multiple time series datasets.
Image: 2109.09265.pdf (arxiv.org)
Last March, Salesforce released CodeGen, a large scale language model which turns simple English prompts into executable code. The idea is to democratise access to the world of writing software, allowing anyone to develop apps in conjunction with an “AI assistant” or “teacher” without the need to learn programming in the traditional way.
CodeGen, a 16-billion parameter auto-regressive language model, has been trained on a large corpus of natural and programming languages. It can be applied to both simple and complex problems, using natural language.
Using CodeGen, people with little or no programming knowledge can solve relatively simple coding problems. Through this, Salesforce also aims to bring disadvantaged groups into the programming world.
Eli Levine, (then) software engineering architect at Salesforce, said ML Lake helps application developers and data scientists to easily build machine learning capabilities on customer and non-customer data.
“It is a shared service that provides the right data, optimises the right access patterns, and alleviates the machine learning application developer from having to manage data pipelines, storage, security and compliance,” he added.
He said one of the main reasons to build ML Lake was to make it easier for applications to get access to the data they require while centralising the security controls needed to maintain trust. The common industry practice is to carefully maintain and curate a number of key datasets for ML or analytics use cases. The metadata in ML Lake is vital for model training and serving and compliance operations.