Palantir and IBM have come up with a solution for simplifying the development of AI-based applications in the enterprise.
Leveraging AI-based tools has become an essential piece of the puzzle for many companies in their efforts toward digital transformation—yet this adoption is still often a hurdle for many IT leaders. While 43% businesses reported accelerating AI tools during COVID-19, according to Todd Moore, vice president, Open Technology, IBM, 59% of IT leaders recently admitted that the new technology felt threatening to them, as TechRepublic recently reported.
Palantir for IBM Cloud Pak for Data, powered by Red Hat OpenShift, is presenting a solution for this, by giving tools to developers to connect data and AI models, even those without high-level technical expertise. Designed around the “no-code/low-code” framework, which has become increasingly popular as a springboard to launching AI projects, the product integrates IBM Cloud Pak for Data services with Palantir Foundry, a data and analysis platform.
No-code offers tools and platforms for simplifying the software development—the kind of software that might be used on platforms like Facebook, Lyft and Google Docs, which was traditionally created with code. According to Moore, Palantir for IBM Cloud Pak for Data will help companies overcome these hurdles by making the creation and deployment of AI applications easier. Users can merge data from a variety of inputs into Palantir applications, as well as “use models developed on Cloud Pak for Data to improve predictions for fairness and accuracy, so businesses can more confidently apply AI,” he said.
“The immediate advantages of ontology-driven no-code/low-code is speed of delivery (MVPs can be developed in hours/days as opposed to months), ease of maintenance, and approachability for broader user groups,” said Nanfei Yan, product manager, Palantir.
A primary benefit is that employees with a great range of skills can all use the technology —everyone from a seasoned data journalist to an operational analyst. Users can pose “what-if” questions, creating scenarios rather than creating technical inputs/outputs for the model, Yan said.
As COVID-19 continues a major shift in business operations, with online operations becoming an even more critical piece of workflow for many companies, there is more need for “closer synergy between human-data-AI decision making than ever,” said Yan. Instead of relying on models or manual decisions, Palantir for IBM Cloud Pak for Data offers an “ontology-driven no-code/low-code AI-infused solution” that pairs “best-in-class AI recommendations with context-specific data to operational applications in record time, so the end users can make the most informed decisions,” she continued.
How it works
The product automates the collection and organization of data in the hybrid cloud. This is then captured by IBM Watson Catalog and transformed into a unified data asset to be available for a range of people who can analyze the results.
The major benefit, according to Moore, is that companies will have a “shared business ontology,” which can help integrate data across different business operations, eliminating silos, and will end the need to create new applications from the ground up. “Analysts don’t need to treat these applications as black boxes, and engineers can collaborate directly with end users on iteration,” Moore said. “On the other hand, common concerns to this approach resolve around the technical scalability and depth of configurability of the solution.”
“Efficiency has always been a driving factor in technology, and the ontology-driven no-code/low-code environment is an obvious step in the continuous journey for faster speed to delivery,” he continued.
While the solution can be used in a range of industries, one example is how Fiserv, a financial services company, uses it for securely harnessing data and developing models for intelligent decision-making. “Palantir for Cloud Pak for Data is designed to help companies like Fiserv ensure they have the business-ready data they need to scale AI with confidence,” Yan said.
Tools like this can bridge the gap between data scientists and analysts, by “identifying potential root causes for deviations, and suggesting potential cohort changes,” she said. “The end user can leverage both models and contextual data to perform scenario analysis, make their final decision, and that decision can be written back to the system to influence future campaigns.”