Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. Watch now.


Using artificial intelligence (AI) as part of a larger business process can often be a complicated exercise. 

That was the experience that Gero Keil and Thilo Hüllmann had as they were trying to build a tool focused on creating ease for sales professionals, by using AI to help with lead generation and validation. The process wasn’t easy, and the two developers realized there was a need for a tool that is accessible by non-technical people to build intelligent business workflows.

“We started building out some initial tooling that would make our own lives easier,” Keil told VentureBeat. “As we spoke to other companies in our network, we realized that having a platform that makes it very simple to build machine learning models and integrate them into the business workflow is something that actually a lot of companies would benefit from.”

Keil and his cofounder, Hüllmann, have developed the Levity no-code automation platform, which provides organizations with a modular approach to building business workflows. The company announced today that it has also raised $8.3 million in seed funding. A beta version of Its platform is now available.

Event

Intelligent Security Summit

Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.

Register Now

Using AI blocks to enable a modular approach 

The concept of no-code tooling is gaining widespread adoption across the IT landscape. With no code, rather than an organization needing to have developer skills to write application code, instead, there is a visual interface that enables users to build an application without coding.

There is no shortage of vendors using some form of no-code or low-code approach to enable different aspects of AI. For instance, Cogniteam, has a platform for robotics and the company Regie, uses a low-code approach to enable a sales and marketing content creation service.

Levity provides its users with AI blocks, which provide a modular set of capabilities that can be used to build out an application workflow. Keil explained that an AI block is a component that enables a user to create a deep learning model based on their own data. The models have all been pre-trained, but can be fine-tuned by users as needed.

While there is an opportunity to fine-tune the AI models, a large part of what Levity also does is educate its users so they understand what can be accomplished.

“More than 95% of our users do not have any technical background with AI,” Keil said. “We produce quite a good amount of educational pieces that you help inform people of what’s possible and what isn’t.”

How to build an AI workflow with Levity

Levity has a focus on business workflows and how to automate them.

Keil explained that a user can use Levity to build out basic and advanced workflows. 

A basic workflow can be for email sorting such that when an email arrives in Gmail the AI block will automatically read that email and its attachments, figure out what should be done with the email and then put it into the proper cloud storage folder or in the right system. Going a step further, Levity’s no-code interface provides integrations with multiple tools including Zapier and Make (formerly Integramat) for automation.

Levity can also be used to help an organization automate the process of sorting through support tickets in different systems, including Zendesk and Intercom. The AI block approach has a pre-trained model that enables users to identify urgency and sentiment to help sort and prioritize issues.

Among the more unique use cases where Levity is being used, in one case an organization is actually using its system to help automate a lab process.

“We have one customer who automates a laboratory process where they analyze dog poop samples,” Keil said. “In their case, their workflows are a bit different than most customers.”

He explained that what is similar across all use cases though is that organizations will bring their own data and decide which labels they want Levity’s system to automatically identify. Once that data is loaded, the user clicks on a button that says ‘train model.’ Then, when the user is confident that the model works as expected, it can be integrated into whatever workflow an organization needs.

Looking forward, Keil said that Levity will be expanding the AI blocks to support more capabilities and integrations.

“As of today, the platform covers text, image and document classification,” he said. “We’re aiming to introduce new types of AI blocks on our platform that can solve different kinds of tasks and expand the range of AI technologies that are simple for people to use.”

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.