Organizations routinely acknowledge crises, commission studies, launch initiatives — and change almost nothing. This pattern has a name.

Herbert Simon received the Nobel Prize in Economics in 1978 for dismantling the myth of the rational optimizer. His central contribution was the concept of bounded rationality: agents cannot know all alternatives, calculate all probabilities or perfectly rank all outcomes. This cognitive constraint necessitates satisficing—a portmanteau of “satisfy” and “suffice” describing a process where decision-makers forgo the search for the absolute best option in favor of one that meets a predetermined threshold of acceptability.

The distinction matters: optimizing requires identifying all possible alternatives, determining all consequences, and selecting the single alternative that maximizes the objective function. Satisficing sets an aspiration level, searches sequentially, selects the first alternative that meets the threshold, and stops immediately.

Every insurance executive has heard statistics about the number of insurance veterans expected to exit the industry in this decade and reports about millennials’ lack of interest in insurance careers.

Boards discuss it. Consultants study it. HR departments launch initiatives. And nothing changes.

The paradox isn’t unique to insurance. Decades of McKinsey research document that diverse teams outperform economically, yet companies in the top quartile for diversity are only 35 to 39 percent more likely to outperform on profitability—a gap that should have closed long ago if firms were actually optimizing. The CEO-ego-driven AOL-Time Warner merger destroyed $99 billion in shareholder value through cultural incompatibility that due diligence should have identified. The Boeing 737 MAX crisis cost the company over $21 billion because schedule pressure overrode engineering requirements that executives knew existed.

These are not isolated failures. They are symptoms of satisficing.

The “war for talent” framing is wrong because it externalizes the problem, treating it as scarcity, competition, market forces beyond organizational control. The actual problem is internal: human capital inefficiency caused by the same organizational dysfunctions that cause financial capital inefficiency. These are not separate problems with separate solutions. They are the same problem manifesting in different asset classes.

Human Capital Satisficing

The insurance industry’s response to the talent crisis is textbook satisficing. Rather than optimizing the environment to attract data scientists, carriers build internal “talent factories” to upskill existing underwriters—settling for “good enough” data skills from existing staff. It’s understandable; transformation is terrifying, expensive and carries career-ending risk if it fails. But it creates a vicious cycle. Underwriters become data janitors, spending 80 percent of their time cleaning batch-processed data exports rather than building predictive models.

The “hollow middle” emerges—a dangerous demographic structure where the 55+ workforce substantially exceeds the under‑30 cohort by more than two to one. As senior leaders retire in unprecedented numbers, far fewer qualified internal successors exist to fill the pipeline. This forced imbalance drives companies to promote less experienced juniors into roles requiring deep technical and client expertise, or to engage in expensive external recruitment wars that inflate wages across the industry without addressing the underlying capability and knowledge‑transfer gap.

One HR executive’s observation captures the dynamic: carriers hire high-quality talent to solve big problems, then “bog them down in bureaucracy… have them doing paperwork in the back room trying to get a computer that will work.” The talent leaves because the infrastructure cannot support their capability. Management concludes that “AI just doesn’t work for us” when the real problem is satisficing.

Technical Satisficing: The Ferrari in the Covered Wagon

If you want to understand why the insurance industry is losing the war for talent, stop looking at HR policies and start looking at the server room.

For years, carriers have fought a losing battle for AI talent, unable to match the lucrative compensation packages offered by the technology sector. To compete, they pitch a vision of “revolutionizing risk” and “building the future of underwriting”—only to hand new hires a login to a 40-year-old mainframe running batch cycles on their first day.

This is the corporate equivalent of buying a Ferrari engine and trying to bolt it into a covered wagon. You don’t end up with a faster wagon. You end up with a broken engine.

Every CIO knows the optimal solution is to rewrite core systems for the cloud to enable true real-time data ingestion. But a core rewrite is terrifying—it’s expensive, takes years and carries existential risk if it fails. Digital transformation projects carry failure rates exceeding 70 percent. So instead of optimizing, CIOs satisfice and hire middleware vendors to bridge the gap.

This is where satisficing becomes farcical. As I detailed in an earlier article, “Why Insurance Telematics Integrations Fail,” the legacy core still can’t ingest the live feed. So, the high-tech middleware vendor takes a stream of live telematics data and collapses it into the only format the carrier can handle: a flat file. They turn dynamic, real-time risk signals into a static PDF or Excel sheet, hand it to a human underwriter who must physically open the file, read it, and manually key the data into the policy admin system.

The workflow is the ultimate betrayal of talent. Carriers have the data. They have the people. But they’ve inserted a “flat file” bottleneck that strips out 90 percent of the value before it ever touches the decision engine.

Financial Satisficing: The Rate-Hike Addiction

Human capital satisficing and technical satisficing connect to a third pattern: financial satisficing.

The FIO’s 2025 Annual Report documents the industry’s primary survival strategy: the P/C sector achieved record premium growth, not through operational efficiency or underwriting innovation but through rate increases. Personal auto premiums rose 12 percent; homeowners premiums rose 13 percent in 2024. Reserves ceded to offshore jurisdictions have nearly quadrupled since 2019, exceeding $450 billion.

Rate increases and reinsurance offloading are satisficing solutions: they meet profitability thresholds without requiring operational transformation. The “acceptable” outcome (hitting quarterly targets) substitutes for the optimal one (building capabilities that would make rate increases unnecessary).

Nowhere is the failure of satisficing more visible than in commercial auto. The sector has seen 55 consecutive quarters of rate increases, yet the industry combined ratio has remained above 100 for 12 of the last 13 years. (See for example, “2025 Commercial Auto Study,” Conning, Sept. 2025)

In any other industry, a management team that raised prices 55 times without fixing unit profitability would be replaced. In insurance, it is accepted. Carriers continue to pull the “rate lever” because it is a known, comfortable mechanism that fits into existing spreadsheets. It is the ultimate act of satisficing: choosing the action that is easiest to execute rather than the one required to solve the problem.

AI Satisficing: Paving the Cow Paths

The FIO report notes that 88 percent of private passenger auto insurers and 70 percent of homeowners insurers plan to use AI, based on NAIC surveys conducted in 2022 and 2023. But the primary use cases cited are “underwriting efficiency,” “faster claims processing” and “customization.”

The high-tech middleware vendor takes a stream of live telematics data and collapses it into the only format the carrier can handle: a flat file.”

Notice what’s missing: fundamental business model change. AI is being deployed to do current work faster—paving the cow paths rather than building new roads. This is AI satisficing: using transformational technology to optimize incrementally rather than restructure fundamentally.

Even when major carriers announce billion-dollar AI investments, read the fine print. Are they building autonomous underwriting engines? No. They’re building summarization tools to help humans read claim files faster. They are paving the cow path with gold.

The statistics are revealing: 70 percent of insurers explore AI, but only 22 percent have deployed to production. BCG research shows only 7 percent scale beyond pilots. The gap between technical success and operational adoption is enormous—and it traces directly to satisficing. Carriers invest in predictive models that underwriters don’t trust or can’t incorporate into existing workflows. The technology works. The adoption fails.

Why AI Satisficing Is Not Safe

Legacy carriers watching InsurTech struggle might feel vindicated. The InsurTech winter is real: global funding fell from $15.8 billion in 2021 to $4.25 billion in 2024—a 73 percent decline. Root Insurance’s stock collapsed more than 90 percent from its IPO peak to around midyear this year. Lemonade fell 75 percent. Multiple high-profile failures validated skeptics who always thought digital disruption was overhyped.

Related article: Has InsurTech Funding Found a Floor?

Here’s the problem with that analysis: Amazon’s stock collapsed 96 percent from peak to trough during the dot-com bust. Critics in 2001 declared the e-commerce experiment a failure. The schadenfreude was palpable—one Fortune 500 CFO proclaimed to me that dot-com companies “were not real companies” and the people who worked at them “were not real business people” and “real companies” would not hire them. That CFO’s company was subsequently acquired, and its brand name has since disappeared.

Stock price is a lagging indicator. Leading indicators tell a different story. Root Insurance posted its first profitable year in 2024: $31 million net income after losing $147 million the year before. Root’s gross loss ratio for Q3 2025 sits at 59 percent—better than the P/C industry average of 62 percent documented in the FIO report, and far better than many legacy personal auto carriers. The founder stayed and executed a disciplined turnaround.

Some critics note that legacy carriers are adding millions of policies while InsurTechs add thousands—confusing volume with value. That’s the same analytical error that dismissed Amazon for only selling music, movies and books while it was building Amazon Marketplace and eventually AWS.

The pattern is identical: massive stock collapse, critics declaring victory, then quiet execution on fundamentals while incumbents relaxed. Amazon achieved profitability in Q4 2001—during the crash—and used the “quiet period” to build AWS. Today its market cap exceeds $2 trillion.

Legacy carriers dismissing InsurTechs because of stock price collapses are making the same analytical error retailers made dismissing Amazon after the dot-com bust. Focusing on stock price is satisficing—it’s the easy metric to measure. Analyzing unit economics, loss ratios and operational execution is optimizing—it’s hard. And the executives who confuse the easy metric for the meaningful one will be blindsided when the survivors emerge.

Three forces are now converging to create existential risk for organizations that continue satisficing.

• First, the demographic cliff. Hundreds of thousands of professionals exiting by 2030 are taking institutional knowledge with them.
• Second, AI knowledge extraction. Competitors are using AI to capture and scale expertise. Lemonade’s AI processes 55 percent of claims autonomously—codifying the judgment of hundreds of adjusters into a scalable model.
• Third, patient capital. Surviving InsurTechs have capital from reinsurers and corporate VCs who learned from early failures. They’re building during the quiet period while incumbents celebrate apparent victory.

Related article: Lemonade Embraced AI in Claims From Inception, And Is Still Eying The Next Tech

A Lesson From Trimble

When I worked on telematics data services at Trimble, a lawyer in one of our German offices was assigned to work on one of my agreements. I remember thinking, “What does a lawyer in Germany know about American law?” He was exceptional—so good I requested him for every future deal. He would work late at night German time to be on calls with partners to discuss agreements, navigating the nine-hour time difference between Germany and California.

That experience taught me two things. First, exceptional talent exists everywhere if you’re willing to look beyond conventional pools. Second, and more importantly, if you schedule all your meetings for American convenience, you’re telling your global team they’re second-class citizens. Customer calls happen on customer time. Internal calls should respect everyone’s time.

Most companies say they value global teams but schedule everything for U.S. time zones and wonder why international talent leaves or disengages. The operational details reveal whether you actually respect distributed talent or just view it as cost arbitrage.

From Satisficing to Optimizing

The organizations that win will achieve both human capital efficiency and financial capital efficiency. Both inefficiencies are quantifiable. Both have the same root causes. Both can be solved—but only by organizations willing to optimize rather than satisfice. Here’s what optimization actually looks like.

Stop restricting your talent pool. Consider a thought experiment: What if the University of Alabama decided to become a national football power but would only recruit high school players from within the state of Alabama or only recruit the sons of the head coach’s friends? No one would frame that as a moral issue. It’s pure competitive math. You cannot be elite while voluntarily restricting your talent pool to a fraction of available players. Nick Saban built a dynasty by recruiting nationally, not because it was virtuous, but because it was the only way to win.

Companies complaining about a “war for talent” while only hiring locally, only considering traditional insurance backgrounds, and only recruiting people who will commute to work in an office are Alabama voluntarily restricting itself to in-state recruits—then complaining about the talent war against Georgia and Ohio State. It’s a self-inflicted competitive disadvantage disguised as a talent shortage.

Recognize homophily for what it is. Research confirms that hiring managers unconsciously rate candidates higher when they share demographic characteristics, educational backgrounds or leisure interests. When an interviewer encounters someone who “reminds them of themselves,” communication feels more fluid and trust establishes rapidly. This fluency is frequently misattributed to the candidate’s competence rather than their similarity to the evaluator. There’s a joke in Silicon Valley that it’s a “mirrortocracy,” not a meritocracy.

The research is damning: when organizations explicitly emphasize meritocracy as a core value, managers actually show greaterbias. The mechanism is “moral credentialing”—believing the system is fair lowers vigilance against unconscious bias. The belief in meritocracy becomes a license for restricting the available talent pool. Companies hiring people they’re comfortable with—based on school, age, race, gender, family background, regional accent or origin—and calling it meritocracy are satisficing on a feeling of comfort rather than optimizing for capability.

Measure output, not presence. “Butts in seats” productivity monitoring is satisficing because attendance is easy to measure. Developing real output metrics is optimizing because it’s hard. One produces attendance data. The other produces performance data. The carriers that mandate return-to-office without clear productivity frameworks aren’t optimizing for performance—they’re satisficing for the appearance of control.

Deploy knowledge extraction now. The window for capturing institutional knowledge is 2025 to 2027. An AI model gets smarter with every interaction. A company starting knowledge capture now will have a model with years of experience by 2028. A competitor starting in 2028 will have a rookie model. That gap may never close.

Kill the zombie projects. Stop funding “good enough” initiatives and ruthlessly prioritize the few optimal moves that actually change the business model. Don’t hire data scientists until you have a cloud-native environment where they can actually work. Don’t put a Ferrari engine in a wagon.

If you think you’re in a war for talent, you’ve already lost. The real battle is for human capital efficiency, and how you fare depends entirely on internal factors your organization controls. The choice is optimization or obsolescence.

At what point does satisficing—the choice to accept adequate rather than optimal performance—become a fiduciary question?

AI Disclosure: Research for this article utilized AI tools to discover and verify publicly available data sources and citations. All analysis, interpretation, conclusions and writing are original work by the author based on decades of operational experience.