A panel discussion at this year's TechCrunch Disrupt conference is putting a spotlight on a shift that's already underway in many workplaces: AI agents that don't just automate a task, but act more like a team member with an ongoing role. The conversation, featuring payroll platform Gusto, venture firm Insight Partners, and admissions-focused startup Leland, frames this as a structural question for founders, not just a tooling upgrade.
The core idea being discussed is that AI agents are moving from single-use tools — the kind that write an email draft or summarize a document — toward systems that hold a defined function inside a company. Think an agent that manages parts of customer support, tracks recruiting pipelines, or reconciles expenses on a recurring basis, with some degree of independence between human check-ins. That's a meaningful shift from the chatbot-assistant model most businesses have gotten used to over the past two years.
This conversation isn't happening in a vacuum. Over the past 18 months, companies across HR tech, sales tools, and finance software have all rolled out some version of an autonomous or semi-autonomous agent — Salesforce, Microsoft, and Intuit among them. What's notable about this panel is the framing: it's less about the technology's capability and more about organizational design. How do you manage headcount, accountability, and culture when part of your workforce isn't human? That's a question enterprise software vendors have mostly avoided answering directly, because it complicates their pitch.
Gusto's inclusion is telling. As a payroll and HR platform built for small and mid-sized businesses, its presence signals that the AI-agent-as-teammate conversation is trickling down from Silicon Valley engineering teams to the kind of five-to-fifty person companies that make up most of the U.S. economy. Insight Partners, meanwhile, invests in growth-stage software companies and has a direct financial interest in which of these agent-based tools actually stick around long enough to generate returns.
The pattern with panels like this one is fairly consistent: high-level framing conversations tend to precede a wave of product announcements, not follow one. Expect the next six to twelve months to bring a round of new "agent" features bolted onto existing HR, sales, and finance software — often as a premium tier rather than a free upgrade. That's been the trajectory with most AI features rolled out since 2023: initial access is often free or bundled, then it moves behind a paywall once usage data shows it's sticky.
For small business owners, the practical question isn't whether to have an opinion on AI teammates as a concept — it's whether any specific agent tool solves a problem you're currently paying a human, or an expensive vendor, to do. The honest trade-off is that these agents tend to work well on narrow, repetitive, well-documented tasks and poorly on anything requiring judgment calls or relationship management. A scheduling or invoice-reconciliation agent is a very different bet than one handling customer complaints.
There's also a governance question that's easy to skip in the excitement: who is accountable when an AI agent makes a mistake that affects payroll, a customer, or a compliance filing? Larger companies are building internal review processes for this. Most small businesses don't have the bandwidth to build formal oversight, which means the risk of an unsupervised error sits closer to the owner's desk than they might expect.
Worth watching in the coming months: whether any of the panel's named companies — particularly Gusto, given its small-business customer base — ship a concrete agent product rather than just talking about the concept, and what it costs. Also watch whether early adopters report actual headcount changes, or whether agents end up supplementing existing staff rather than replacing tasks outright, which has been the more common outcome so far.
The bottom line: the framing of AI as a "teammate" is a marketing and organizational shift as much as a technical one, and small business owners should evaluate any specific agent tool on the narrow task it performs and its error-handling, not on the broader narrative.