A few years ago, the enterprise software sector was buzzing about robotic process automation (RPA). These services were designed to automate repetitive business tasks — but more often than not fell short. However, generative AI may finally provide the tools necessary to create robust systems that actually accomplish these goals.
Seattle-based Tektonic is one startup working in this space. The company emerged from stealth Wednesday and announced a $10 million seed funding round led by Point72 Ventures and Madrona Ventures. It uses generative AI in tandem with traditional approaches to enable users to automate workflows using natural language.
For instance, automating quotes and renewals involves many different manual tasks unique to each business. Tektonic co-founder Nic Surbatanu previously held leadership roles at Tanium, UiPath and Microsoft; he said generative AI could usher in new possibilities for automation.
“[Generative AI] combined together with software … gives you the best results,” Surbatanu said. “It’s flexible, it understands user intent.”
According to Surbatanu, RPA vendors have long approached automation with a preconceived notion of what can be done based on available technology. This means most of these platforms lack the adaptability required to interpret user intent accurately.
He added that some capabilities introduced by foundation models have been particularly helpful for entity extraction — Tektonic uses both open models and foundation models for this task — but generally speaking there are still issues around usability.
Point72 partner Tripp Shriner called the startup’s combined approach “powerful” because it merges “the creativity and intelligence of humans” with machines’ ability “to scale up knowledge work.”
Madrona managing director Tim Porter suggested generative AI would help sales reps by automating processes needed before they can spend more time talking with customers.
Tektonic is currently in early development and working with design partners to test its system. In a few years, the company hopes to become a SaaS business connecting to customers’ systems via APIs; for now, businesses must install Tektonic as a container in their virtual private cloud.
Source: techcrunch
