Enterprise AI: A Closer Look
As part of the 2021 Fintech and Enterprise Tech Spring batch programs, Plug and Play hosted a webinar on Enterprise AI. During the webinar, we sat down with Stefano Lindt, Senior Director of Alliances at C3.ai, one of the leading Enterprise AI companies in the world, for a virtual fireside chat to talk about opportunities and challenges around operationalizing artificial intelligence (AI) at scale. Thereafter, the event concluded with presentations from five startups that are providing tools to help deploy AI at scale.
Find out some of the main takeaways from the fireside chat in this article.
Defining Enterprise AI
In short, Enterprise AI is a category of software that leverages AI to drive digital transformation, exploit opportunities such as understanding/predicting customer needs, and solves problems such as proactive threat detection. Adopting certain tools and technologies will allow companies to develop provisions and deploy AI applications at scale.
Stefano Lindt emphasizes that organizations still face challenges to deploy ML models and operate them at scale. When a company has large amounts of ML models in production at the same time, the ability to manage the governance of those models from development to provisioning, to production becomes a really hard task to do.
Operationalizing and scaling Enterprise AI
Executives around the world recognize the strategic importance of using AI, but only 13% of companies have succeeded in moving beyond POCs to scale AI across their enterprise. Some of the challenges and barriers in scaling AI applications that Stefan has seen include:
- Identifying the right business use case and what the company is trying to solve. There may be difficulties in finding the right use cases and key low-hanging fruit use-case that can quickly demonstrate value may be overlooked.
- Finding the data. Companies want to have repeatable use cases but that requires the right data to leverage AI. Organizations are aware that they have the data, but they might not have the ability to find it, to access it, and they may not have enough history of the data or the right signals of it.
- It requires complex systems that can aggregate data and create a unified data image to build AI and ML applications and certain technologies at the company’s disposal may not be sufficient enough.
- Much of a data scientist’s time is still spent on low value tasks - data wrangling, aggregation, cleansing, and normalization - so any tools that can automate much of this away and let free up time to focus more on important tasks such as analytics has the possibility to drive better ROI.
Solutions to increase the success rates for AI and ML adoption
During the webinar, we conducted a live poll, asking our audience to identify which factors were a major roadblocks to adopting AI/ML, the shortage of talent was identified as a key barrier. This is not surprising as other studies show that there is a shortage of about 150,000 data scientists in the US alone.
The second biggest challenge was lack of collaboration across the organization. Operationalizing AI is business-driven but the business side may need data from across the organization so bringing in all the relevant stakeholders is crucial to successfully deploy these applications. Organizations also need the necessary environments where people can build and test their models, as well as having the right tools and technologies to enable them.
This is just a glimpse of the insightful discussion with Stefan Lindt. For the full recording of the fireside chat and startup pitches from Workera, dotData, LeapYear, Dataron, and Automation Hero, please click here.
Featured Startups
Workera is an Enterprise workforce skills measurement, benchmarking and upskilling solution focused on Data and AI capabilities.
dotData automates feature engineering, the most manual and time-consuming step in AI and ML projects. dotData’s proprietary AI technology automatically discovers hidden patterns behind hundreds of tables with complex relationships and billions of rows and AI features for your AI and ML algorithms.
LeapYear is the world’s first platform for differentially private reporting, analytics, and machine learning.
Datatron provides a single model governance (management) platform for all of your ML, AI, and Data Science models in production.
Automation Hero combines RPA with AI to form an end-to-end intelligent process automation platform that automates repetitive and time-consuming tasks for knowledge workers.
Enterprises are not what they used to be. At Plug and Play'sEnterprise Tech Accelerator we are in touch with corporations and startups that are shaping the future of work. Join our platform today.
