Participation in the movement of the business leaders in the business for about two decades. Changing VB brings people builds on the actual approach to Enterprise Ai. Learn more
While businesses face the challenges of deployment of AI agents of critical applications, a new, more pragmatic model that recalls people as a strategic protection against AI failure.
An example is BlendA platform that uses a “commenue-in-the-loop” method to make AI reliably reliably for critical mission work.
This method is a response to growing evidence of perfect autonomous agents a high-stake gamble.
The high cost of the unexpired
The problem with Ai thoughts becomes a visible risk while companies are exploring AI applications. In a recent event, AI-Powered Coded Editor Editor Cursor detected its own bot support Inventing a fake policy restricting subscriptions, spark a wave of cancellation of public customer.
Similarly, Fintech Klarna company is famous Returned course To replace customer service agents with AI after admitting the step resulting in lowest quality. In a more alarming case, New York City City Chatbot advises entrepreneurs in Join illegal worksHighlights the risks of following the danger to unexpected agents.
These incidents are symptoms of a larger gap of capability. According to a May 2025 Salesforce Research paperTop agents now only succeed in time of work alone and 35% of people’s steps with world world capabilities. ”
The model being co-worker
To bridge this gap, a new method focuses on human handling structures. “A AI agent should work in your direction and for you,” Mixus co-founder Elliot Katz speaks VentureBeat. “But without the construction of organizational management, fully autonomous agents generally generates many problems than they solve.”
This philosophy undergoes MixUs-In-The-loop’s model, embedding direct verification of the person directly to automated operations. For example, a large retailer might receive weekly reports from thousands of stores that contain critical operational data (eg, sales volumes, labor hours, productivity requests from headquarters). Man’s analyzes should spend a lot of time reviewing data and making decisions based on heuristics. With Mixus, AI AG Agutor autuated the heavy lift, analysis of complex patterns and flaging anomalies such as unusual pay requests or productivity outliers.
For high-stake decisions such as payment permissions or policy violations – the workflows that a human user as “high risk” – the agent stops before continuing. The division of labor between AI and people are mixed with the agent’s creative process.
“This method means people will only participate if their skills are actually increasing value – usually decisions with 30-95% of the usual tasks with automatic effects,” Katz said. “You get the speed of whole automation for standard operations, but manage the Kicks accurately if the context, the most important.”
In a demo shown by the Mixus Team to make an agent an intuitive process to be instructional instructions. To establish an agent who examines the truth for reporters, for example, the co-founder Shai Magzimaf has simply described the measures of the specified measures and a claim that has the consequences or consequences of reputation.
One of the main platform strengths are engagement of tools such as Google Drive, email, allowing users to participate in agents directly to participate in email approval agents
Platform participation capabilities expand further to meet specific business needs. Mixus supports the Model Context Protocol . Combined with companions for other business software such as Jira and Salesforce, it allows closing agents to open tickets to an inprador and reporting status of a slack manager.
Manage man as a strategic multiplier
AI AI space currently undergo a review of the truth while companies move from the production experimentation. The agreement with many industry leaders is that the people in the hole are a practical requirement for agents to continue reliably.
Mixus collaboration model changes the economy of scaling ai. Mixus prophesied that by 2030, the deployment of agent can grow 1000x and every man’s overseer can be 50x that AI agents can become more reliable. But the general need for human management will improve.
“Every man’s overseer manages more AI work for hours, but you still need more fully managed while AI deployment explodes your organization,” Katz said.
For business leaders, it means that human skills develop rather than disappear. Instead of being replaced by AI, experts emphasize the roles in which they have ordained AI’s and handling high-stakes decisions that are flagged for their review.
In this design, the construction of a strong humanitarian handle can be a competitive advantage, allowing firms to conduct AI more aggressive and secure with their opponents.
“Master companies This multiplication will reign in their industries, while those who are pursuing the perfect automation will fight for reliability, obedience,” Katz said.
