Small Teams, Big TAMs: The Ultra-Lean AI Startup Movement and How it Will Feast on $100Bn+ of Financial Services Middle Office Spend

The rise of ultra-lean AI startups is reshaping the startup landscape, proving that artificial intelligence can drive substantial ARR (>$10M) with minimal staffing (<10 people) and limited capital (<$10M). We refer to these companies as 10x10x10 AI startups. By embedding AI into both their core products and internal operations, these startups demonstrate that small teams can achieve outsized impact.  We explore this trend—what we term Small (Staffing) is Beautiful—later in this blog. 

One particularly compelling area where these startups are making significant strides is financial services middle-office automation.  While not the most glamorous sector, it presents a vast, high-value opportunity for automation. We estimate that financial services businesses spend at least $100 billion annually on middle-office operations—and the true figure could be far higher. This space is especially exciting to us at Core, given our fintech roots and our commitment to enabling more affordable financial services.

My first encounter with the “shadowy middle market” was in the 1990s while working at JPMorgan. The back office was responsible for processing, the front office for driving sales—but what was this enigmatic middle office? It was staffed by high-wage, “middleware-like” employees who manually validated data, processed paperwork, and managed countless routine tasks—whether filling out compliance forms or confirming information via phone, fax, or email. Despite variations across firms, these middle office workflows follow a similar, labor-intensive pattern within financial services subsegments such as P&C insurance brokering, mortgage lending, and auto leasing, to name a few.

This makes the middle office a prime target for AI-driven automation, offering greater cost efficiency, improved auditability, and enhanced visibility. For example, Pyq, a startup we’ll discuss later in this blog, automates the middle office for property and casualty insurance brokers, delivering substantial efficiency gains. Pyq’s AI-driven platform eliminates the need for brokers to manually fill out carrier portals or sift through extensive documentation, reducing quoting time by as much as 90% compared to using internal staffing or BPO solutions. ROI is typically achieved within weeks—often under four weeks and sometimes in as little as one. However, it’s important to note that achieving this level of efficiency requires deep industry-specific expertise; without it, the product risks inaccuracy, unreliability, and, ultimately, a loss of trust.

The future of financial services isn’t just about front-end AI chatbots—it’s about fundamentally rewiring the middle office. This shift will allow employees and teams to focus more on innovation and growth rather than the labor-intensive, repetitive workflows that have long burdened the industry—costing companies and their customers dearly. Given the intense competition in banking, insurance, and asset management, we expect much of the automation-driven savings to be passed on to customers in the form of lower costs and more innovative, market-responsive solutions.  When will this seismic transformation occur?  We answer that question later in this blog.

Caught in the Middle (Office): A $100+ Billion Automation Opportunity

Financial institutions invest heavily in their middle office to meet stringent compliance demands and maintain operational precision. Across banking, asset management, and insurance, middle office spending is estimated to exceed $100 billion annually:

• Banking:
  Functions such as risk management, regulatory compliance, treasury operations, and market analytics account for roughly 10–20% of operating expenses—translating to approximately $100-200 billion (sources: Fred, Core estimates) 
• Asset Management:
  Portfolio risk management, compliance oversight, and performance analytics typically consume 5–15% of operating expenses, roughly $10–$20 billion.  (Sources: McKinsey, Core Estimates)
• Insurance:
  With claims management, risk assessment, compliance, and actuarial analysis, spending again ranges between 10–15% of operating expenses—amounting to about $25–$60 billion. (Sources: US Treasury, Core Estimates)

This vast, process-heavy expenditure is ripe for disruption by a new class of ultra-lean AI startups—a trend we call “Small (Staffing) is Beautiful.” Teams of fewer than 10 people, primarily engineers, armed with sAI and modest pre-seed or seed capital can now achieve millions of dollars in customer sales and profitability, which previously required much larger organizations. In this new paradigm, profitability at the seed stage is not an exception—it’s becoming the norm.

Small (Staffing is Beautiful): The Ultra-Lean Startup Advantage

Artificial intelligence is reshaping the startup landscape by empowering small teams to compete with and even outperform larger, well-funded companies. Much like AWS spurred a startup revolution, AI on tap through LLMs and LLM services brokers is fueling a new wave of innovation and startup formation.  These ultra-lean teams can:

• Lower Staffing Needs:
  With a single-digit number of engineers and minimal administrative support staff, these teams build, test, and refine products at unprecedented speeds. A highly skilled team of < 10 employees can build an enterprise-ready product that once required a much larger workforce.  We estimate that one engineer can achieve the output of more than five (or more) by leveraging automated AI programming solutions–and this leverage will increase over time with improvements in AI technologies.  With fewer engineers needed to achieve the desired result and one engineer equating to five (or more), you don’t need to put out six REQs to build a new scrum to scale out your business.  One engineer can now produce like a scrum.  As a consequence, we believe seed-stage enterprise teams are now capable of delivering Series A metrics (or better).  
• Increase Agility and Speed:
  These lean teams are positioned to pivot quickly in response to market shifts,  Smaller human teams enable rapid, seamless communication that significantly shortens product development cycles and creates a higher bandwidth connection between customer feedback and product development.  
• Reduced Overhead:
  Without the burden of extensive finance or HR departments, founders can focus more intently on product iteration and founder sales rather than organizational management. It also means that investment dollars can be ever more focused on high ROI investments in product, engineering, and marketing (and with limited, if any, administrative cost leakage).  
• Access the Highest-Quality Talent:
  Fewer team members and fewer administrative layers allow for more significant cash compensation and equity allocation to top-notch engineers.  Historically, we’ve talked about top engineers being well worth their high salaries.  Levered up with AI, this is the case even more so.  Moreover, top-tier engineers thrive in environments where they are unburdened by excessive bureaucracy (who doesn’t love working in an admin-free environment!), making ultra-lean AI startups highly desirable employers for leading engineers.  Also, with the ability of small teams to deliver massive outcomes, this improved quality of work does not need to involve an equity upside tradeoff.
• Scale On-Demand:
  Say goodbye to the stair-step build-out of start-ups, where each step is punctuated by the founder diverting time and energy to pursue a significant capital raise.  Startups can now scale operations continuously and in response to market demand, funding new AI-leveraged hires on a just-in-time basis through revenue growth or smaller fundraising rounds. 

Yes!  Seed Stage and Profitable: Disruption Delivered without Disrupting Your Cap Table

A key element of this new model is achieving significant progress with modest early-stage venture investment. In the past, before ZIRP compromised startup building frugality and discipline, companies like eBay demonstrated how venture capital may not be a necessity when lean teams encounter ready markets and deliver a solution that delivers transformative value. With every corporate CEO and CFO focused on how AI can deliver cost efficiencies, the market is definitely ready for selling AI solutions that streamline the middle office.  We have seen AI middle office solutions enabling a predictable >5x ROI in within a few months.  As a result, the demand for these solutions is robust.  Much like eBay, these startups won’t need to rely on heavy venture capital—this next generation of AI startups can deliver substantial profitability and growth without substantial capital.

Notable Ultra-Lean Startup Examples

Reducto

• What They Do:
  Reducto addresses one of the toughest challenges for large language models—interpreting complex documents. By developing next-generation OCR technology, Reducto transforms unstructured data from PDFs, spreadsheets, and images into structured formats that LLMs can readily use.  
• Funding:
  Founded in 2023, Reducto secured $8.4 million in seed funding.

Pyq

• What They Do:
  Pyq automates property and casualty (P&C) insurance broker workflows. In smaller brokerages with revenues around $10M, a significant portion of the team is dedicated to these labor-intensive roles. Pyq’s automation addresses this inefficiency, creating a ripe market for its solutions.
• Funding:
  Founded in 2022, Pyq has raised an amount consistent with the companies in this blog.

Pasito

• What They Do:
  Pasito automates a range of time-consuming workflows in employee benefits. These include preparing open enrollment guides detailing the policies offered by benefits providers, enabling employees to easily compare plans (including doing so based on their personal usage history), automating new hire onboarding workflows, and personalizing communications for employees to maximize the value of benefits.  Today these tasks are managed manually by carriers, vendors, benefit brokers, and employers, and the cross-coordination is people-intensive.  Pasito offers a platform that automates and streamlines these cross-company workflows.
• Funding:
  Pasito is seed stage and has raised $5M.

Other Notable Examples (Outside Financial Services)

These companies achieving tremendous growth outside of financial services increases our confidence that we will see similar companies addressing financial services. 

• • AI Ace:
    Developed by a group of students, AI Ace is an educational chatbot offering personalized learning experiences. It secured seed funding estimated at under $5M, outpacing traditional tutoring systems and driving rapid innovation.
  • Bolt (formerly StackBlitz): An AI-powered, browser-based development platform that enables users, including non-technical individuals, to create full-stack web and mobile applications using natural language prompts. It leverages AI to simplify the app development process, allowing users to describe the desired application and have Bolt generate functional prototypes rapidly.
  • Character AI:
    What started as a modest project evolved into a leader in conversational AI by raising seed capital under $5M, emphasizing quality and continuous iteration.
  • Cursor:
    An AI-powered platform that automates code generation, debugging, and refactoring, Cursor has achieved $100M in ARR with just 20 employees and under $5M in capital investment.
  • Midjourney:
    Known for turning textual prompts into stunning digital art, Midjourney secured early-stage funding in the low millions, enabling a small, agile team to dominate a niche market.
  • Tennr:
    Tennr’s solutions are tailored to healthcare practices aiming to optimize administrative tasks, improve operational efficiency, and enhance patient.  Tennr reads, reasons, and responds to paperwork, significantly reducing manual data entry and administrative errors. 

Challenges and Considerations

Despite the enormous potential, these startups face several challenges:

• Implementation and Integration Complexities:
  Integrating AI into existing middle office workflows requires careful planning, especially in high-compliance environments that are rife with legacy software.  So scalable startups must think through how to efficiently navigate managing consultative sales, bespoke integrations, and high-touch implementations that require organizational change.
• Data Quality and Security:
  Ensuring robust data management is crucial to provide confidence that the AI solutions match the assurance levels of current less automated middle-office operations.  That said, automation is beneficial for auditing and service consistency, so AI middle office solutions will offer a step-function improvement in reliability, consistency, and transparency vs. the old way of doing things.
• AI Programming Oversight:
  Although AI empowers lean teams, it still requires experienced human oversight for quality assurance and ongoing management of the code base.  Having a hypothetical team of AI programmers does not obviate the need for human management (so still plan to take your AI programmer out for an occasional coffee check-in:) 
• Avoiding over-reliance on consulting and services:
  While short-term profitability is possible, it necessitates well-managed consultative sales and implementation skills. The nature of middle office workflows means that one size will not completely fit all, at least not out of the box.  Some customization and hand-holding will be needed versus traditional SaaS solutions that folks may buy over the Internet with limited, if any, interaction with the associated product and engineering teams.  The AI advantages that create the opportunity for building ultra-lean startups also present risk.  
• Customer retention/stickiness: The other side of the coin to the leverage AI provides these startups is that it provides a similar low-friction entry point for competitors. To protect margins and market position, these companies will need to find ways to deeply embed their solutions into customer workflows and, over time, explore how they can become a new “system of record” layer that displaces old ERP systems and legacy software.  This new system of record will not have data embedded inside a proprietary application; rather, the AI will “play the data where it lies” and create a system of record based on a virtual database that accesses necessary information from the systems where it is currently retained.  If these middle office solution companies cannot address this issue, they run the risk of declining margins over time.  While historically, the software industry has been successful at steadily increasing prices, this may not be the case in this new AI world.  It is possible that software using AI to disrupt others may become disrupted by AI itself.  Without a clear pathway for ensuring user stickiness, it may be that impressive initial ARR traction for these companies may be a “false positive” in terms of indicating a strong product-market-fit longer-term viability.

Conclusion

The era of ultra-lean, AI-powered startups is unfolding at this moment. With financial institutions spending over $100 billion annually on middle office functions, the opportunity to automate these labor-intensive processes is enormous. Pioneering companies like Pyq and Pasito show that a small, agile team armed with advanced AI can streamline complex workflows, reduce costs, and significantly increase customer profitability—achievements that once required far larger organizations. Their success, mirrored by breakthroughs in sectors such as software development, digital art, conversational AI, and personalized education, underscores a new paradigm: lean operations–modest investments in human and financial capital– can yield monumental returns. As AI continues to evolve, it will resemble the second coming of AWS and enable a new startup renaissance that will redefine industries, including the software industry, delivering higher returns on capital and ushering in a new bull market for early-stage venture capital. 

2025 marks the Chinese Year of the Snake—symbolizing transformation and growth. We see 2025 as the year when AI-powered business transformation takes hold in the financial services middle office. Moreover, we foresee this seismic trend of middle office transformation extending beyond Financial Services across industries.  If you are building a Small (Staffing) is Beautiful AI startup that is enabling this transformation, please contact us.  We at Core would love to hear from you.