How to Find AI Opportunities in Boring Businesses

TL;DR

The strongest AI opportunities often come from automating recurring workflows that cost ordinary businesses time or money. Start with a process you understand deeply, especially one dependent on phone calls, email, spreadsheets, or manual paperwork, then build a narrow prototype, show customers a measurable before-and-after improvement, and iterate quickly using their feedback.

Transcript

I find that building something meaningful often takes either real technical depth uh and or deep cusper insight and integration with customers. And yes, we can now, you know, use AI to code something in a few hours, but that's a small piece of the puzzle. >> If you're trying to make money with AI, this is probably where you're stuck. The tools are ... Read More

Key Insights

• Building with AI is no longer the primary obstacle because prototypes can now be created quickly and cheaply. The harder problem is identifying a recurring business need that customers consider valuable enough to pay to solve.
• The largest implementation gap is between experimenting with AI and embedding it into core operations. A cited Goldman Sachs survey found that 76% of small business owners used AI, while only 14% had integrated it into core operations.
• Boring business workflows are promising targets because many still depend on phone calls, email, spreadsheets, handwritten estimates, and repeated follow-ups. These manual processes create visible delays, errors, missed appointments, and lost revenue that automation can address.
• Domain knowledge is a competitive advantage because experienced workers possess tacit information that is not fully documented online. Their familiarity with recurring errors, difficult customers, unusual exceptions, and workflow bottlenecks can guide the design of useful specialized agents.
• Property management loses money when customers wait between discovering a listing, requesting details, calling, attending a viewing, and submitting a deposit. An industry-specific agent could connect these stages and reduce delays for customers who want to decide quickly.
• The product management bottleneck is deciding what should be built, not merely producing software. Andrew Ng recommends learning AI, building rapidly, speaking with customers, and repeatedly testing ideas against problems that consume measurable time or money.
• Local marketing for physical businesses is presented as an accessible starting area. Useful applications include bringing previous customers back, responding to inbound emails, and converting missed messages into appointments, rather than merely generating generic social media posts.
• Home-service companies are attractive automation candidates because activities such as preparing estimates may still be completed manually. HVAC, roofing, pool service, and repair businesses offer processes where a builder can examine paperwork and operational friction directly.

Questions & Answers

Q: How can I find a profitable AI business opportunity?

Start by identifying a recurring process that costs a business time or money, especially one still managed through phone calls, email, spreadsheets, or paperwork. Speak directly with the people performing that process, learn its exceptions, and build a narrow prototype quickly. The goal is to demonstrate a clear before-and-after improvement that a customer can connect to saved time, recovered appointments, or reduced revenue loss.

Q: Why are boring businesses promising markets for AI?

Ordinary businesses often contain essential workflows that have not been deeply automated, even when employees already use general AI tools. Dental practices can miss calls, construction managers can prepare estimates by hand, and accounting firms can chase documents through email. These problems are valuable because they occur repeatedly and affect operations, while fewer builders may be competing to create specialized solutions for them.

Q: Why does domain knowledge matter when building an AI product?

Domain knowledge reveals details that general models and public information may not contain. An experienced worker knows which customer calls repeatedly, which document commonly returns with an error, and which unusual exception slows the entire process. That tacit knowledge can shape an agent's prompts, rules, and workflow, giving the resulting product a practical advantage over a generic tool built without customer integration.

Q: What is the product management bottleneck in AI?

The product management bottleneck is the difficulty of deciding what to build after the cost of building has fallen. Rapid coding does not establish whether a product solves an important problem or whether customers will pay for it. Andrew Ng's suggested response is to learn AI, build quickly, talk with customers, develop judgment about their needs, and iterate around recurring problems that consume time or money.

Q: How is using AI different from integrating it into core operations?

Using AI can mean opening a general chatbot for occasional assistance, while operational integration means placing AI inside the workflow that produces business results. The transcript cites a Goldman Sachs survey in which 76% of small business owners said they used AI, but only 14% had built it into core operations. The opportunity lies in connecting AI to tasks such as calls, estimates, appointments, documents, and follow-ups.

Q: How could AI improve a property management customer journey?

An industry-specific system could support the customer from discovering a property through reviewing details, calling, attending a viewing, and leaving a deposit. The central problem is delay between these stages. When prospective renters want to decide quickly, waiting for replies or paperwork can cause lost revenue. Agents can create value by connecting the stages and helping the company respond and advance each customer more promptly.

Q: What AI services can local physical businesses use?

Local physical businesses can use AI to reactivate previous customers, answer inbound emails, and turn missed messages into appointments. These services are more directly tied to operating results than producing batches of generic social media ideas. They are also presented as relatively easy opportunities to test because a builder can show the business a before-and-after comparison without first acquiring years of specialized industry expertise.

Q: How can my current job lead to an AI business idea?

Look closely at the repetitive task you already perform by hand each week. Your experience gives you knowledge of its normal sequence, recurring mistakes, awkward exceptions, and financial consequences. The transcript argues that the person inside the process may be well positioned to solve it because prototypes are now inexpensive. Package one painful workflow into a focused product, test it with customers, and improve it through direct feedback.

Summary & Key Takeaways

• Building software has become cheaper and faster, so selecting a valuable problem is now the main bottleneck. Strong opportunities often exist in ordinary businesses where essential processes remain manual. Builders should investigate recurring delays, errors, missed communications, and administrative work that directly consume time or cause lost revenue.

• Industry experience can become a competitive advantage because employees understand undocumented details that general AI models may not capture. They know which documents return with errors, which customers require follow-up, and which unusual exceptions disrupt operations. That tacit knowledge can be translated into specialized prompts, agents, and workflow software.

• The recommended approach is to find one painful workflow, speak with customers, build a quick prototype, and demonstrate a clear operational improvement. Easier opportunities include local marketing and home-service administration, while more specialized problems demand deeper expertise because mistakes carry greater consequences and customer integration becomes more important.

• The opportunity is not simply giving businesses access to a chatbot. Many small businesses already use AI without integrating it into core operations. Valuable products connect multiple stages of real work, such as responding to inquiries, arranging appointments, processing paperwork, analyzing metrics, and preventing revenue from disappearing during delays.