Although employees spend much of their day communicating and coordinating with each other on various projects, this effort is often undermined by the availability of certain individuals. When a colleague with critical information is absent because of vacation or other reasons, the rest of the team must delay progress until that person responds, writes TechCrunch.
The website specialized in technology also mentions a specific problem for companies that turn to outsourcing, a practice also found in Romania, especially in the IT sector: what happens if you have an urgent task to solve, but the colleague you need to help you works from a country located in another time zone?
A common solution is for employees in the country where the activities are outsourced to work in the same time zone as those in the country where the employing company is located. But that comes with obvious inconveniences for employees who have to work a very different schedule than what would be normal in their country.
A Silicon Valley company has already raised millions of dollars to solve the problem
Ashutosh Garg and Varun Kacholia, the founders of Eightfold—a Silicon Valley recruiting startup based on artificial intelligence and recently valued at $2.1 billion—believe they have found a solution to this problem, as well as other situations that can arise in a company.
They believe that advances in large-scale linguistic models (LLM) and privacy technologies are the way forward. Earlier this year, they launched Viven, a startup focused on “digital twins” with a mission to give employees access to crucial information from colleagues, even when they're not available.
This week, Viven secured $35 million in seed funding from several venture capital firms and individual investors.
How the digital “twin” works
Viven develops a specialized LLM for each employee, practically creating a digital twin by accessing internal electronic documents, such as emails, Slack channels or Google Docs files. Other employees in the organization can then query a person's digital twin to get immediate answers related to joint projects and shared knowledge, just as you can ask questions to ChatGPT.
“When every person has a digital twin, you can just talk to their twin as if you were talking to that person and get the answer,” Ashutosh Garg told TechCrunch.
A major obstacle is that people cannot simply share any information with anyone who asks. Employees often handle sensitive information or have personal files that they want to keep confidential from the rest of the team.
Garg claims that Viven's technology solves this complex problem through a concept called “privacy and relational context.” It allows the startup's LLM models to determine “precisely,” according to their developers, what information can be shared and with whom within the organization.
How colleagues are discouraged from asking the chatbot prying questions
Viven says its LLM models are smart enough to recognize personal context and know what information should be kept private — such as questions about an employee's personal life. But perhaps the most important safeguard is that each person can see the history of queries to their digital twin, which acts as a deterrent against inappropriate queries.
“It's a very difficult problem, and until recently it was unsolvable,” Ashu Garg, a partner at Foundation Capital, one of the firms that invested in Viven, told TechCrunch.
On the competition, Ashutosh Garg says no other company is yet tackling the concept of “digital twins” for the enterprise environment.
TechCrunch reminds us, however, that just because there are currently no direct competitors does not mean that other companies will not develop such LLM models for companies to use in the future. Even Garg points out that Anthropic, Gemini, Microsoft Copilot, and OpenAI's enterprise search products already include a personalization component.
He hopes, however, that if the big tech companies enter this market, Viven will have a competitive advantage through its “relational context” model.
