AI Companions Promise Efficiency, but Team Trust Is the Hidden Price

Town and Tolan: Two Startups Reshaping Workplace AI

On October 19, 2026, two AI assistant companies announced major funding milestones on the same day, offering a window into where venture capital is concentrating its bets.

Town, a San Francisco-based startup, reached unicorn status just four months after its public launch. The company raised $108 million in a Series B round co-led by Forerunner Ventures and Menlo Ventures, bringing its total funding to at least $160 million. Town’s core product gives each employee at a client company a personal AI assistant, internally called a “Townie,” connected to email, calendars and core business tools. These agents do more than answer queries: they share information with one another across a team and can interact directly with external customers on behalf of staff. A new shared workflow feature automatically updates every Townie with fresh company information, reducing the need for manual configuration by individual users.

On the same day, TechCrunch reported that Tolan, an AI companion designed to help users manage relationships, emotions, daily schedules and personal goals, closed an additional $30 million in funding, doubling its total capital raised. PitchBook now estimates Tolan’s valuation at more than $268 million, though the company has not publicly confirmed the figure. Where Town is squarely an enterprise tool, Tolan occupies a different register: a digital companion that blends task management with emotional support, positioning itself at the intersection of wellness and productivity.

Together, these two rounds point in the same direction. AI products are no longer selling themselves as occasional utilities. They are selling presence, a persistent layer embedded in the daily rhythm of professional and personal life.

When AI Agents Begin Coordinating Work

Town’s design makes this ambition explicit. A Townie doesn’t simply retrieve information on command; it participates in organizational workflows, communicates with other AI agents and, in some configurations, engages directly with external customers on behalf of human employees. The trajectory is clear: toward AI systems that don’t just accelerate individual tasks but begin to absorb coordination roles previously handled by human interaction.

The World Economic Forum, in an October 2026 analysis, described this kind of setup as “hybrid workforce orchestration,” identifying it as one of the core management challenges of what it calls the intelligence era. The WEF’s argument is that AI has so far mainly affected individual tasks, but is now beginning to reorganize how work is structured, distributed and governed across teams. Town’s networked Townies are a concrete example of exactly that reorganization, already in motion inside client organizations.

What Management Research Is Warning

While the investment community celebrates the valuations, a quieter conversation is underway in management research. Harvard Business Review published two pieces in October 2026 examining how AI tools affect the social fabric of teams.

The first, published on October 15, documents what researchers call “relationslipping”: the gradual erosion of informal trust and coordination that occurs when employees habitually consult AI tools for advice, feedback or problem-solving instead of asking a colleague. The logic is not complicated. Many of the small interactions that build team cohesion, a spontaneous question to a peer, a shared moment of working through a problem, occur precisely because no more convenient option exists. When an AI assistant is faster, always available and socially frictionless, those interactions start to disappear.

A second HBR article, published on October 20, focuses on how AI affects leadership communication. It finds that AI-mediated workflows can flatten the range of signals that leaders send and receive, reducing the space for the empathy and candor that psychological safety research consistently links to high-performing teams. The concern is not that AI assistants are inherently damaging. It is that their adoption, without deliberate organizational attention, can silently alter the quality of human interaction in ways that are difficult to measure until the damage has accumulated.

Productivity Gains Without Bottom-Line Returns

There is a concrete data point that connects these research findings to business outcomes. A BearingPoint survey of 1,050 C-suite executives and senior leaders across 13 countries, cited by the World Economic Forum in October 2026, found that 27 percent of organizations that have implemented AI report productivity gains with no measurable impact on their bottom line. Tasks are being completed faster. Revenues and margins are not improving in kind.

The WEF attributes part of this gap to organizational design failures: companies often automate individual tasks without rethinking how those time savings translate into business value downstream. MIT Technology Review, in an October 2026 analysis of enterprise AI economics, adds a complementary explanation. It argues that organizations continue to measure AI expenditure in terms of input costs, subscription fees and infrastructure spend, rather than what it calls “effective cost per useful outcome”: the total cost required to complete work the business can actually use, such as a resolved customer request or a closed sales cycle.

Together, these perspectives suggest that the productivity trap may have a relational dimension. If faster individual task completion comes at the cost of less peer collaboration and weaker interpersonal trust, then some of the value AI assistants appear to create at the individual level may be quietly offset by reduced collective performance at the team level.

How Organizations Can Preserve Cohesion While Deploying AI

None of this amounts to a case for slowing AI adoption. The economic logic behind products like Town is strong, and the operational demand for tools that reduce coordination overhead is real.

Fast Company’s October 2026 survey of senior leaders noted a clear shift in how effective management is now being defined: less top-down coordination, more proximity to the actual work and to the people doing it. That proximity is harder to maintain if the default first interaction for most employees is an AI agent rather than a human colleague.

A few questions are worth asking inside any organization currently piloting AI assistants. Are managers creating deliberate spaces, team rituals, shared problem-solving sessions, where human interaction remains the natural default rather than an extra effort? Are they tracking not only task completion speed but also team cohesion and the quality of collaborative output over time?

The organizations that will benefit most from AI companions and networked agents are likely not those that deploy the most sophisticated tools, but those that treat the deployment itself as a design problem. Efficiency and trust are not inherently in conflict, but they will not coexist without deliberate attention. The startups raising hundreds of millions of dollars on the promise of smarter workplaces are building real technology. What happens to the human layer underneath it remains largely up to the people in charge.