Build the Platform Small Businesses Can't Live Without

01 /The Concept

Small local service businesses – trash haulers, AC repair shops, pest control operators – run on hustle. Owners and employees are almost always on the move, either doing the work or meeting clients.

As a result, customer acquisition tends to happen in the background – either organically, or by handing it off to a student who knows the word "marketing," or by paying a real agency real money.

In the US, more than 1 million companies fit this profile: small local service operations without a dedicated growth engine.

Cohesive is building a platform to automate growth for exactly these businesses.

The platform is a CRM – but one that populates and updates itself through an AI engine.

The AI regularly scrapes Google Maps and other public data sources to extract information on companies and property owners in the surrounding area, then enriches records with contact details. (The website currently emphasizes commercial businesses; the project description also references residential property owners.)

Once potential customers are identified, the AI inserts them into the CRM and sends personalized outreach emails – tailored to the specific services that each company or property owner is likely to need.

If a response comes in that opens any door to continued conversation, the platform immediately flags it for the business owner and updates the CRM record accordingly.

Pricing is a flat monthly subscription. The exact number isn't published yet.

That's because Cohesive was only founded last fall and is currently in Y Combinator, still working out precisely what the product is and what it should cost.

02 /Why It Matters

Cohesive cites 1 million potential customers – small local service companies.

That may actually be an undercount. There are roughly 35 million small businesses in the US. Small businesses represent 99.9% of all registered companies in the country.

Strip out the storefronts, cafés, and gas stations that attract customers through location and signage, and what remains are businesses that must actively find new customers – essentially using the same neighborhood-scanning logic. The only variable is which signal the AI uses to filter which companies or homeowners are actually likely to buy.

Even within purely commercial targets, building that prospect list is non-trivial. LinkedIn – the dominant US business database – has records for fewer than 10% of local companies. That means any serious approach requires scraping other sources, matching across platforms, or using purpose-built tools designed to surface this data.

This is exactly the category Cohesive is entering – and it's not alone. Fellow Y Combinator graduate Resquared ([related review](/review/zarabatyvaj-na-teh-kogo-trudno-najti)) raised $5 million last summer on the same core thesis: scrape local business data and turn it into actionable outreach, all in one platform.

Openmart ([covered here](/review/jeto-uzhe-dengi-no-mozhno-zarabotat-eshhjo-bolshe)) also emerged from YC and raised $2.75 million last summer. Its angle: the AI doesn't just find local businesses – it layers in unusually granular detail. The platform can filter for venues that have jukeboxes, for instance, because its AI found that fact buried in a customer review somewhere.

03 /Opportunities

AI is coming for small business, eventually. There will be AI-native tools that become as indispensable to these companies as phones and spreadsheets are today.

In many markets, small businesses represent the overwhelming majority of companies. 35 million and 99.9% is probably the outer bound, but even capturing a sliver of that market can sustain a substantial startup.

The broad direction: build AI platforms with a real shot at becoming default tools for small business.

One obvious version is a Cohesive / Resquared / Openmart-style self-filling CRM – one that identifies prospects, handles initial outreach, and surfaces warm conversations for business owners to pick up. After a while, these businesses will find it hard to imagine how they ever found customers manually. This fits neatly into a Paul Graham maxim: imagine a picture of the future and then build what's currently missing.

Here's another angle the review surfaces. It's often said that the goal of a startup is to replace someone's job. But 82% of small businesses in the US have no employees at all. There is no person whose job it is to find new customers. Most solo operators can't afford freelancers or agencies for this, and the owner ends up doing it haphazardly in whatever spare moments exist.

They want to do it properly. They just don't have the time or the expertise. And expertise is acquirable – but only if you have the time to acquire it. Which they don't.

That points to a different framing for startup ideas: give people the results of tasks they don't have time for. You're not replacing anyone's job – because nobody's doing the job consistently. "Give people time back" is a principle on its own.

Who else is running chronically short on time for something that matters to them? Where can AI be wired in to start generating results from tasks that people currently can't get to?