Does the Firm Become Smaller

If agents make external coordination cheaper, a firm may find it easier to work with specialized suppliers or partners. Automated discovery, evidence exchange, and routine negotiation could reduce some of the effort required to transact across company boundaries. That could make outsourcing, partnerships, or project based networks more practical for selected work.

If agents make internal coordination cheaper, a firm may be able to coordinate more activities without adding equivalent management overhead. It could expand its reach, serve more customers, or operate more complex workflows within the same organizational structure.

The same technology can therefore push in opposite directions. The result depends on which costs fall faster: the cost of coordinating internally, or the cost of coordinating through markets and partners.

Earlier analyses of this question include Malone, Yates and Benjamin’s (1987) argument that information technology could shift coordination toward markets, and Brynjolfsson, Malone, Gurbaxani and Kambil’s (1994) study of whether information technology leads to smaller firms.

A business decision cannot be based on coordination cost alone. External work may increase exposure to supplier failure, data leakage, quality variation, or loss of critical knowledge. Bringing work inside may require investment, specialized talent, and management capacity. Contracting can be difficult when outcomes are uncertain or cannot be specified in advance.

Agents may also create new costs on both sides of the boundary. Internal use requires authority, oversight, context, integration, and recovery. External use requires identity, disclosure limits, compatible semantics, commercial terms, and recourse. If these costs remain high, cheaper message exchange will not necessarily make market coordination efficient.

Bargaining power matters too. A large firm may use agents to coordinate an extended supplier network while retaining strong control over terms. Smaller partners may bear integration costs without gaining the same benefits. A more connected market does not automatically become a more balanced one.

The useful question is not “Will firms become smaller?” in the abstract. It is: for which activities does agentic coordination change the relative cost, control, and risk of internal versus external organization?

Start with a specific activity. Estimate its current internal coordination cost, external transaction cost, quality requirements, information sensitivity, and failure consequences. Then test how a bounded agent system changes each. Include the costs of specification, supervision, integration, verification, and dispute resolution. Separate labor removed from labor shifted to another party.

An activity may move across the boundary without changing the overall size of the firm. A company may contract for more specialized work while expanding its platform and control functions. Another may retain execution but reduce administrative coordination. Firm boundaries can shift by activity, capability, and risk rather than contract uniformly.

Evidence should compare real organizational choices before and after agentic coordination is introduced. Did a company bring work inside, move it outside, or change the terms of a relationship? Did the full cost fall after accounting for quality, oversight, risk, and transition? Who captured the savings? Did the company preserve the knowledge and control it needed?

These effects may take years to appear. A pilot that lowers coordination time in one workflow cannot establish that the optimal firm boundary has changed.

Agents could make firms smaller, larger, more networked, or simply different. The economic mechanism is plausible: changes in coordination costs can change organizational choices. The direction remains conditional on the relative costs of markets and internal administration, and on the institutions that govern both.