← Insights

Lifting the Coasean Ceiling: How AI Rewrites the Economics of Insurance Distribution

A top life insurance carrier set out to build a next-generation life product. Then, as I have seen so many times, senior leadership grew impatient. They wanted early returns, and when those returns were delayed, they shut the program down. It was a great product that served the market well. It is gone, not for lack of vision, but for lack of time and resources.

Executive summary

AI does not abolish the Coasean ceiling. It lifts it. That changes the math on how big a firm can grow before complexity eats its gains, and how small a transaction can be before it stops being worth doing.

Insurance has lived under a low ceiling and a high floor for a century. Advice, underwriting and servicing were too costly to deliver one-to-one at scale. So the industry served large, profitable cases and left about 99 million under-protected Americans below the floor.

AI changes both edges at once. The ceiling rises: one firm can coordinate far more distribution, data and decisions without breaking. The floor drops: a small monthly policy, sold inside a benefits portal in minutes, can finally cover its own costs. The winners will be the firms that pair that capacity with people. Labor and capital have always been complements; AI raises the return on both.

The gap is not vision. It has always been time and resources. AI makes time move faster and resources go farther.

Coase, the firm, and the two limits

Firms exist because markets are expensive to use. In 1937 Ronald Coase asked why, if markets work so well, anyone bothers to build companies. His answer: finding partners, negotiating terms and enforcing contracts all cost money. Inside a firm, a manager's direction replaces many of those deals. He won the Nobel Prize in 1991 largely for that insight.

The same logic sets a limit on size. A firm grows until organizing one more activity inside costs as much as buying it outside. Past that point, every new person, product line or region adds more coordination cost than value.

Clay Shirky named the two edges of that logic in Here Comes Everybody (2008):

  • The Coasean ceiling is the size above which an institution becomes so unwieldy that the cost of managing it stops it from working. Fred Brooks saw the software version in 1975: adding people to a late project makes it later.
  • The Coasean floor is the size below which an activity is not worth any organization's overhead. Shirky's point was that the internet dropped that floor, so millions of small acts became possible.

Every business lives between the two. Its ceiling caps how much it can coordinate. Its floor sets the smallest customer it can serve at a profit.

Breaks or lifts? The ceiling moves; the theory holds

The honest claim is that AI lifts the ceiling rather than repealing it. Coase himself said that a technology like the telephone cuts both the cost of using the market and the cost of organizing inside a firm. Which one falls faster decides whether firms grow or shrink. AI is cutting both at once, and quickly.

The ceiling is still visible today. GE split into three public companies. Comcast spun off its cable networks. In each case, coordinating very different businesses under one roof cost more than it returned. A seasoned investment banker put it to us plainly: AI does not break the ceiling, it creates a higher pressure point before a breakup is needed. We agree. The theory holds; the math changes.

What changes in the math:

  1. Internal coordination gets cheaper. Status updates, handoffs, reporting and quality checks that once needed layers of managers can run through agents and shared data. A firm can grow further before overhead overwhelms it.
  2. Market transactions get cheaper too. Finding a partner, pricing a risk and drafting terms are tasks agents can do at low cost. That pushes the floor down and lets small firms buy what used to require scale.
  3. The ratio between labor and capital shifts. Every investment mixes people and money in some proportion. AI moves where that ratio lands, company by company and industry by industry.

So AI produces two kinds of winner at once: bigger platforms that coordinate more, and smaller specialists that plug into them. The firms in the middle, too large to be nimble and too fragmented to coordinate, feel the squeeze.

Labor and capital are complements

At some point labor (L) and capital (K) are always complements. When a tool makes a person more productive, that person's hour is worth more, not less. That is why wages rose after mechanization, even as specific jobs vanished.

The fear that AI simply replaces people echoes the machine-breakers of the Industrial Revolution. They were right that their particular jobs were at risk. They were wrong about the economy: the machines made the remaining work, and much new work, more valuable.

The practical version for executives is simple. Financial capital that recognizes it can amplify human capital is better placed to grow than capital that only tries to cut it. Human and financial capital have always moved together; AI changes the ratio at which they combine, not the fact that they do.

AI raises output per unit of both inputs. Where it raises the value of a person's judgment, it raises what that person is worth.

Why insurance has lived under a low ceiling and a high floor

Insurance is a business of transaction costs. Before a policy exists, someone has to find the buyer, explain a product most people do not understand, gather medical and financial facts, judge the risk, and keep the policy in force. Each step has historically needed a person.

The result is a floor that sits high. When acquiring and underwriting a customer costs hundreds of dollars, a small-premium policy does not pay. So the industry built around agents and brokers serving larger cases, and carriers built around paper and lab-based underwriting.

The cost of that floor shows up in the protection gap. LIMRA's 2025 Insurance Barometer puts individual life insurance ownership at 51% of U.S. adults, down from 63% in 2011. About 74 million Americans say they need coverage and have none, and another 25 million say they need more. The gap is widest in households earning under $50,000. Many adults, younger ones especially, overestimate what a policy costs.

The ceiling sits low too. Carriers, MGAs, brokers, benefits platforms and employers each hold a piece of the customer and the data. Coordinating across them has been so costly that each tends to stay in its lane. That fragmentation is a Coasean ceiling in action: the industry cannot organize itself around the customer because the coordination costs are too high.

What a lifted ceiling looks like in insurtech

AI lowers the cost of every step between a person and a policy. That drops the floor so small policies become profitable, and lifts the ceiling so one platform can coordinate many partners.

StepWhat made it costlyWhat AI changes
Finding the buyerCold outreach, low response, high acquisition costEmbedding in channels people already trust: benefits portals, HR platforms, financial apps
AdviceA licensed person explaining needs one-on-oneAI benefit advisors that size need from data the person has already shared, with a human on call
UnderwritingParamedical exams, labs, weeks of reviewInstant, data-driven decisions; field underwriting at the point of sale
ChoiceToo many products and riders; the paradox of choice stalls decisionsA personal recommendation of one or two right-sized options
ServicingManual changes, call centers, lapsesAgents that handle routine service and flag life events
Partner coordinationBespoke integrations and contracts per carrier, broker and platformStandard APIs and agents that make each new partner cheap to add

Swipe the table sideways to see every column.

Three consequences follow.

  1. Embedded distribution wins. When the cost of reaching a person drops to near zero inside a channel they trust, the channel becomes the store. The employer benefits platform is the clearest example.
  2. Instant underwriting becomes table stakes. If a decision takes weeks, the floor stays high no matter how cheap acquisition gets.
  3. The orchestrator becomes the scarce role. Someone has to coordinate carriers, data, channels and service. A lifted ceiling means one focused firm can play that role across many partners without collapsing under the complexity.

The opportunity is large. As an illustration, closing even a tenth of that 99-million-person gap would add roughly 10 million newly protected people.

Where the human still wins

Lowering transaction costs does not remove the value of a human relationship. In some moments it makes that relationship more valuable.

Our banker colleague offered a useful example. In Europe he ate at a restaurant where robots took orders and delivered food. It was a fun experience, but his sense was that many diners come once for the novelty and do not return. Many prefer the one-on-one service of a person. Industries will take time to learn where automation adds value and where people still expect a human.

Insurance has the same split. A 32-year-old adding term life during open enrollment wants speed and a clear answer. A family after a diagnosis, a death or a business sale wants judgment, empathy and accountability. AI should do the first job end to end and prepare the human for the second.

The design principle: automate the transaction, keep the trust. Use AI to remove friction, size the need and handle paperwork. Put licensed people where the stakes, emotions or complexity are high. That is the labor-capital complement in practice.

What it means for each player

PlayerWhat the lifted ceiling allowsThe risk of standing still
CarriersProfitably write small policies through many embedded channels at onceDistribution partners pick carriers that can decide instantly
Brokers and consultantsServe more employers with the same team; bring voluntary benefits to every client, not just the largestSmaller clients stay underserved, and the broker brings less to each renewal
Employers and HR platformsOffer personal, right-sized protection to every employee without adding HR workEmployees stay under-protected; benefits spend underperforms
InsurtechsAct as the orchestrator connecting carriers, data, channels and serviceTools that stand alone struggle to reach customers at scale

Swipe the table sideways to see every column.

The common thread: the cost of coordinating has fallen. Firms that use that to connect more partners and serve smaller customers will grow. Firms that use AI only to cut headcount will save money and miss the market.

Conclusion

Coase still explains why firms exist and where they stop. AI has not repealed that logic; it has moved the lines. The ceiling is higher, the floor is lower, and the space between them is where the next generation of insurance companies will be built.

For most of the industry's history, the barrier to innovation was not vision. It was time and resources. AI makes time move faster and resources go farther. For an industry that leaves 99 million Americans under-protected, that is not an efficiency story. It is a chance to reach the people the old economics could never afford to serve.

Sources