> ## Content Index
> Fetch the complete content index at: https://nnutshell.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Cheaper building shifts scarcity to choosing which problem to solve
- URL: https://nnutshell.com/cheaper-building-shifts-scarcity-to-choosing-which-problem-to-solve/
- Published: 2026-10-25T23:09:54.000Z
- Updated: 2026-10-25T23:09:54.000Z
- Description: Sprinto CEO Girish Reddicar, a bootstrapped founder turned VC-backed operator, argues that as AI makes building cheap, the hard part becomes finding painful problems and earning enterprise trust.
- Author: Admin
- Tags: Girish Reddicar, Sprinto, Recruiter Box, Enterprise SaaS, Agentic AI, Venture Capital, Cybersecurity, #rec-2026-10-24, New

*Recorded 24 October 2026 · 35 min*

As building software gets cheaper, scarcity shifts to deciding what to build, argues Girish Reddicar, co-founder and CEO of Sprinto. In an interview with Founder Story host Danielle, he draws on two companies: Recruiter Box, a bootstrapped business that reached thousands of customers before he sold it, and Sprinto, a venture-backed compliance company.

The tension is that cheaper creation helps competitors as much as founders. Reddicar says customer empathy, domain expertise and enterprise rigor do not get easier, while marketing and problem selection get harder. He also argues that software should run outcomes autonomously, yet in compliance, where wrong evidence can reach an auditor, that autonomy has to clear an unusually high reliability bar.

1. **Marry the problem** — [▶ 4:26](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=266s&ref=nnutshell.com)  
Before Recruiter Box worked, Reddicar tried five or six ventures over 3 years that did not. His lesson is to commit to the problem rather than to solutions or ideas. He says building is addictive, and that having power at your fingertips can lead you to keep building things nobody wants, a trap he claims to have fallen into more than most. Recruiter Box succeeded as a simple CRM for hiring, which he calls boring and unsexy but necessary, filling a hole of that size and supporting a viable business.
2. **Scarcity shifts to selection** — [▶ 6:28](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=388s&ref=nnutshell.com)  
As building becomes cheaper, Reddicar argues, deciding what to build becomes the scarce and costly act. His test: is this a top-three problem for someone, and are they ready to pay? In B2B, he says, a buyer unwilling to spend money signals a problem that is not large or deep enough. A few hours of quick coding does not replace being close to the customer's pain and understanding it as intimately as they experience it. Cheaper building, he says, simply means more competition and more hungry people chasing the same problems.
3. **Mom Test before code** — [▶ 8:05](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=485s&ref=nnutshell.com)  
Reddicar credits the book The Mom Test with changing his approach. Its premise is that polite people will encourage an idea the way a mother would, then not pay for the product. He promised himself he would write no code for Sprinto until he had interviewed prospects. Although he could have claimed to be his own customer, having faced the problem before, he interviewed about 30-40 other startups about security and privacy due diligence. The idea itself shifted between the first conversation and the last, and only then did he judge it worth building.
4. **Expertise survives, marketing costs more** — [▶ 11:21](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=681s&ref=nnutshell.com)  
Asked whether customers can now simply build their own solutions, Reddicar concedes that a class of problems will disappear that way. Another class will remain, he argues: problems deep enough that customers do not know how to solve them, where domain expertise is still a source of advantage. The flip side of democratized creation is cost elsewhere. More products mean more noise, so standing out and knowing whether you are building the right thing are both more expensive. In his view, taking a product to market is harder than it was 5 years ago. He offers no data on where the boundary between the two classes falls.
5. **Beating the incumbent by 2-3x** — [▶ 13:15](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=795s&ref=nnutshell.com)  
Selling to small businesses involves one decision point; enterprise buying, Reddicar says, is extremely risk-averse and organized around downside protection. Security and privacy are the baseline, and a startup needs a good answer to why not an incumbent that has been there for 20 years, echoing the joke that nobody gets fired for using IBM. His rule of thumb: a product that is 20-40% better will lose, and it must be 2x-3x better to offset the risk of buying from a nobody. He also underestimated the customization and rigor required, and says committee buying lets one stakeholder's no derail a deal.
6. **Boring problems, digital trails** — [▶ 16:03](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=963s&ref=nnutshell.com)  
Reddicar is drawn to unglamorous problems that are painful and valuable. His logic: every business depends on others, that dependency leaves a digital trail, and counterparties increasingly want assurance the trail is secure and private. His example is a window-dressings and curtains company that needs compliance because it supplies Bed Bath & Beyond and handles its order and supplier information. Internally, Sprinto frames its mission as increasing the GDP of B2B commerce by building trust and removing friction. He adds that the market has repeatedly proved larger or more painful than he estimated, and that market sizing is imprecise but useful for training how you think.
7. **From promise to proof** — [▶ 20:27](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=1227s&ref=nnutshell.com)  
Reddicar says the dial moves from promise to proof as a startup progresses through pre-seed, seed, A, B and C. Early investors back the team, expecting the product, customer or market to change while the founders remain the constant. The business is mainly a vehicle for conversation. Later rounds resemble public-market analysis: growth, gross margins and NDR in a spreadsheet, with a forecast built on them. At some point, he says, fundraising becomes financial modeling rather than promise. He cautions that this is his first VC-funded startup and that he does not understand everything about how VC works.
8. **Founder as bottleneck** — [▶ 22:46](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=1366s&ref=nnutshell.com)  
Bootstrapping taught Reddicar every job in a SaaS company, but left him as the bottleneck: the company could grow only as fast as he could learn. Sprinto was designed to grow a large company fast, which meant building top-down, hiring people better than him at each stage and letting them tell him what to do. Venture capital was downstream of that choice. The first round closed before any code existed, on a napkin explanation to an investor who had watched him build and sell his earlier company, which he cites as proof that overnight successes take years.
9. **Outcomes, not tasks** — [▶ 27:32](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=1652s&ref=nnutshell.com)  
Most AI use today is babysitting, Reddicar says: assign a task, review, steer, rework. A human at the wheel remains the bottleneck. Real leverage, he argues, requires software committed to outcomes rather than task sequences, and reliable enough to run in the background. In compliance, hallucinations are unacceptable because wrong proof cannot be shown to an auditor. His analogy contrasts today's intern with an overenthusiastic domain-expert employee who charges ahead and pulls you in only when stuck. The need is acute, he says, because governance, risk and compliance workloads are exploding while team sizes stay the same.
10. **Watch what won't change** — [▶ 32:30](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&t=1950s&ref=nnutshell.com)  
Reddicar declines to forecast. The AI frontier moves too fast, he says, for guesses based on today's understanding to be reliable, though he expects life in 3-5 years to look very different. His hedge draws on a Jeff Bezos idea: pay attention to what will not change. For Reddicar, that means people still buy from people, so understanding customers, empathy and personal connections remain important even in a world of bits and screens. The host adds that opportunities may lie where few are looking, in what stays constant.

Girish Reddicar's core claim is that cheaper software does not make business easier; it relocates the difficulty. When a working product can be assembled quickly, the scarce resource is knowing which problem deserves the effort. Reddicar says the burden of deciding what to build rises as the cost of building falls. He speaks from experience. Before Recruiter Box, he tried five or six ventures over 3 years that did not work, and he says he is guilty, more than most, of falling into the trap that building is addictive. Having the power to make things at your fingertips, he argues, can lead you to keep building what nobody wants.

His filter is demanding. The problem should be a top-three problem for someone, and that someone should be ready to pay; in B2B, he says, a buyer unwilling to spend money signals a problem that is not large or deep enough. Recruiter Box fit the pattern: a CRM for tracking candidates through hiring, which he calls boring and unsexy but necessary. Validation came through The Mom Test, whose premise is that polite people encourage you the way your mother would, then do not buy. Reddicar promised himself no code until he had interviewed prospects. For Sprinto he spoke with about 30-40 other startups, though he could have claimed to be his own customer. The idea shifted along the way.

Cheaper creation, in his account, also raises costs elsewhere. More people can build, so more noise competes for attention. Reddicar says taking a product to market is harder than it was 5 years ago, and knowing whether you are building the right thing is harder too. Finding a problem large enough that customers cannot solve it themselves is harder still. The host pressed the obvious objection: if customers can build their own tools, why pay a vendor? Reddicar concedes that a class of problems will vanish this way. Another class remains, he says, where customers do not know how to solve the problem and need an expert, and there domain expertise is the advantage. He offers no data on where the boundary falls, so the claim rests on his judgment.

Enterprise buyers are where trust becomes the product. Reddicar contrasts selling to a founder, a single decision point, with selling to an enterprise, which he describes as extremely risk-averse and focused on downside protection. Security and privacy features are the baseline. A startup must also answer why not an incumbent that has been there for 20 years, echoing the old joke that nobody gets fired for using IBM. His rule of thumb is that a product 20-40% better will lose to the incumbent; it must be 2x-3x better to offset the risk of buying from a nobody. He also underestimated the customization and rigor required, and notes that in committee buying one stakeholder's no can derail a deal. The 2x-3x threshold is his heuristic, not a measured figure.

Compliance suited him because he is drawn to unglamorous problems that are painful and valuable. His reasoning is that every business depends on other businesses, that dependency leaves a digital trail, and that counterparties increasingly want assurance the trail is secure and private. His example is a window-dressings and curtains company that needed compliance because it supplies Bed Bath & Beyond and handles order and supplier information. Sprinto's internal framing, he says, is to increase the GDP of B2B commerce by building trust and removing friction. That is aspirational language, and the evidence offered is anecdotal. More concrete is his remark that, more than once, the problem has proved larger, deeper or more painful than he estimated. He is also frank that market sizing is imprecise but useful for training how you think about a problem.

The same shift from belief to evidence governs fundraising. Reddicar says that as rounds progress from pre-seed through seed, A, B and C, the dial moves from promise to proof. Early investors back the team, which they expect to be the constant while the product, customer or market changes; the business is a vehicle for conversation. Later rounds resemble public-market analysis: growth, gross margins and NDR in a spreadsheet, and a forecast built on them. His own first Sprinto round closed before any code existed, on a napkin explanation to an investor who had watched him build and sell his earlier company. Overnight successes, he notes, are years in the making. He also cautions that this is his first VC-funded company and that he does not claim to understand everything about how VC works.

Why raise at all after a bootstrapped exit? Reddicar says he did not see it as a sacrifice. Bootstrapping taught him every job in a SaaS company, from programming to sales, but left him as the bottleneck: the company grew only as fast as he could learn. Sprinto's goal was to build a large company fast, which meant building top-down, hiring people better than him at each stage and having them tell him what to do. Venture capital was downstream of that choice. The same detachment shaped the sale of Recruiter Box. He rejects the idea of a company as a baby, calling himself a guardian, and says selling was a logical decision because the company had reached the ceiling of what he and his co-founder could do with it.

The bottleneck idea returns in his view of AI. Most AI use today, he says, is babysitting: give a task, review it, steer, rework. A human is still at the wheel, which means the human is still the bottleneck. Sprinto's aim is software that works while you sleep, which requires two things. It must be committed to outcomes rather than sequences of tasks, and it must be extremely reliable. In compliance, hallucinations are unacceptable because the wrong proof cannot be shown to an auditor. His analogy contrasts today's intern, who needs the next task, with an overenthusiastic domain-expert employee who charges ahead and pulls you in when stuck. The demand is real, he argues, because the governance, risk and compliance workload is exploding while the number of people available stays the same.

The weak point is the arrangement's dependence on the software knowing when it is stuck. Reddicar sets a high reliability bar and describes the desired behavior, but the conversation does not say how Sprinto meets that bar or how its software decides to summon a human. That matters because the employee analogy works only if the employee recognizes the limits of its own competence, and hallucination is a failure of exactly that recognition. This is an editorial inference, not his argument. It also leaves a practical question: in a domain where errors reach auditors, who verifies the work done while the humans sleep? He is candid about other limits. Asked what he would do differently starting over, he says he has not had time for hindsight, and that lessons take time to distill.

Reddicar declines to forecast. The frontier is moving so fast, he says, that guesses built on today's understanding of the technology are unreliable, though he expects life in 3-5 years to look very different. His hedge is a Jeff Bezos idea: pay attention to what will not change. For him, that means people still buy from people, so understanding them, empathy and personal connections stay important in a world of bits and screens. The host adds that opportunity may sit where few are looking, in what stays constant. The argument fits his whole account: cheap building, noisy markets and autonomous software all leave a human decision-maker at the end of the chain who must be convinced. The unsettled question is which of those human checkpoints, whether customer, buying committee or auditor, survive automation, and who gets paid to stand at them.

**The open question.** Reddicar wants software that works unattended and pulls a human in only when it is stuck, while also demanding that it never hallucinate in audit-critical work. The conversation does not explain how such a system recognizes the limits of its own competence, or who verifies its output while the humans sleep. As an editorial inference, the pull-me-in model depends on exactly the self-awareness that hallucination undermines. Whether reliable autonomy is achievable in compliance, and what evidence would show it, remains untested here.

Watch the full video at [https://www.youtube.com/watch?v=Pt2-Lm4wlpI](https://www.youtube.com/watch?v=Pt2-Lm4wlpI&ref=nnutshell.com).