Welcome to the AI Index Report, featuring practitioner analyst and entrepreneur Toni Witt. This series is intended to deliver the timely intelligence about artificial intelligence (AI) you need to get up to speed for an upcoming client engagement or board meeting.
This episode is sponsored by the AI Summit Preconference at Community Summit, taking place October 16th in Charlotte, NC. The full-day preconference is dedicated to providing an understanding of where, why, and how AI should be applied to drive business results.
PayPal CPO John Kim mentioned that a product the company is launching within the next few months relies a lot on AI. Essentially, it’s a checkout feature that tracks the different ways users input personal information, such as their address. Then, it uses AI to predict which parts of the information to use with the right merchants.
The goal of this is to reduce fraud and minimize the errors that PayPal and merchants have in their customer databases.
Throughout the payments industry, real-time payments (RTP) have become more widespread lately. The problem with this method is that there are little to no protection measures in place. Because transactions settle instantly, they are irreversible. So, VISA built RTP Prevent, which uses AI to manage risk for RTPs.
VISA uses deep learning to provide real-time risk scores to banks, which then decide whether or not to approve RTPs.
ConverSight recently raised a $9 million Series A round of funding. The startup is leveraging generative AI to help enterprises analyze data faster and more thoroughly. This round was led by Surface Ventures. This brings ConverSight’s total funding to over $15 million.
A unique advantage of ConverSight is that it has been around already for five years, well before the emergence of newer AI technologies like ChatGPT. The company already has 200 enterprise customers that it’s serving. Additionally, the company is full-stack in the data pipelines:
- It built a self-service platform that provides enterprises with capabilities such as data sourcing, storage, transformation, analysis, and reporting.
- The platform uses generative AI to automatically generate visual reports.
- It built natural language processing (NLP) into the reporting layer to provide a conversational interface for data teams to further understand and explore data.
Conversight is building a new product called Marketspace, which is a subscription-based database for commonly used AI tools but for specific uses. For instance, if a user needs inventory optimization, they could apply filters on the platform for the specific AI tools designed to tackle that task.
Product of the Week (07:32)
Hugging Face recently announced a tool called SafeCoder, which is a coding assistant designed for enterprises. This tool differentiates itself from other similar assistants because it doesn’t expose users to a third party.
The issue with many of the coding assistants is that whatever code the assistant is given, what the model returns is exposed to a third party who operates and owns the model. This is a major concern, as it would be easy for employees to accidentally leak confidential information.
SafeCoder does not expose customers’ code bases to a third party. Hugging Face achieved this by empowering its enterprise customers to fine-tune their own models on their own code bases. It pre-trained SafeCoder with 15 billion parameters in over 80 programming languages. The deployment of SafeCoder is designed for on-premise or virtual private cloud (VPC) deployment.
Not only do enterprises avoid leaking their code base, but they also get a model that’s optimized for their hardware and fine-tuned to their specific code base and needs.