Peace Island Park runs along a few kilometres of wooded riverbank in northeastern British Columbia, so when a dog runs off or a bear is spotted nearby, facility manager Matthew Weeks can’t just yell for everyone to keep an eye out.

Last winter, he asked software supplier Let’s Camp to make it possible for him to email all the vacationers on site en masse. A few weeks later, the new tool was ready.

Let’s Camp is “really user-friendly,” says Weeks. Customers use the small tech firm’s platform to book one of the park’s 120 tent or RV sites for a few nights. Weeks and his team use it to manage reservations and sell firewood and propane, while municipal staff use it to co-ordinate on-site events like weddings and jet boat races.

Based in Saskatoon, Let’s Camp has been in business for more than a decade, but the speedy development Weeks recently witnessed is the product of a major change two years ago. That’s when Let’s Camp was acquired by Beacon Holdings, a Toronto-headquartered firm that buys small software companies that sell to brick-and-mortar businesses, then infuses them with AI. 

Nilam Ganenthiran founded Beacon at the start of 2024, after working in consumer packaged goods and consulting, followed by a long run up the executive ranks at Instacart and a short stint as a deal maker at D1 Capital Partners, a New York hedge fund. “I felt I could be that translation layer between Silicon Valley, Wall Street and Main Street,” he says.

In just under three years since, Beacon has raised US$560 million from backers like General Catalyst, HarbourVest Partners and Intrepid Growth Partners and acquired 45 companies.

Ganenthiran believes he’s found a new model for transforming business with AI. Beacon holds onto the people and service on which the customers of its portfolio firms rely, but uses automation to upgrade their often dusty products. Rather than building the next hot AI app, it’s assembling a profitable network of software serving tiny markets like campgrounds, car movers, tire storage facilities and youth soccer leagues. 

For many businesses, technology today is a “very small line item,” says Ganenthiran. “I think it’s going to be a bigger line item.” Beacon, he hopes, will be the place customers call when they realize “there’s stuff to automate.” 

Born in Malaysia and raised in Scarborough, Ganenthiran had spent much of his career trying to sell new technologies into what he calls the “real economy.” In early 2024, he saw another such opportunity after trying one of the first AI coding agents, developed by a startup in which he was considering investing. 

Some in Silicon Valley predicted such tools would end the software business, since firms could now program their own internal tech systems. Ganenthiran saw it differently. Small, brick-and-mortar businesses had already been slow to digitize, he thought. With AI, they risked falling even further behind.

Initially, he designed Beacon as a kind of AI incubator, which would rapidly build and market software for niche client bases. Its first product was a facility-management system stuffed with new features for go-kart tracks. It failed at the starting line. “We couldn’t get a facility to sign up, even for free,” Ganenthiran recalls, because track managers already had software “that worked.” 

Main Street businesses relied on their existing technology providers, he realized, and weren’t really looking to switch, even if there were better features available. So rather than try to replace those small software firms, Beacon began acquiring them. “We’re buying relationships [and] trust, and then the tech follows,” Ganenthiran says.

Beacon typically targets companies that have annual revenues between $1 million and $25 million, and are already turning a profit. Starting in the black is key for Ganenthiran, who’s scarred by the many years it took Instacart to break even.

Rolling up smaller software businesses is hardly a new model; early on, Ganenthiran says he filled a spreadsheet with more than 180 competitors. Canada has produced Constellation Software, now a $64-billion conglomerate of over 1,100 firms, as well as Valsoft. 

Ganenthiran and his backers prefer to call Beacon a holding company, a semantic distinction that avoids the negative connotations of roll up, like private equity financial engineering and layoffs. “All of our profitability is driven through growth so far,” Ganenthiran says, adding, “we don’t take headcount out.”

Beacon tries to leave portfolio firms’ founders and existing employees in place to do sales and customer service. A shared back office handles all the accounting, payroll and legal work. 

Beacon’s real edge is an in-house AI toolkit that its centralized product and development staff uses to build new software for portfolio firms and their clients. “We’re in there shipping code with our companies,” Ganenthiran says. Leaders of Beacon’s portfolio firms say the extra technical support lets them rapidly launch new features and functionality. 

AI is everywhere at Beacon. The firm has built a set of agents to identify and evaluate potential acquisitions, with the system constantly adding to and reprioritizing the list. It has also automated some of the work of deal due diligence, using AI for tasks that would normally be outsourced to consultants like reviewing target companies’ contracts or assessing the quality of their earnings. Once acquired, staff at portfolio firms can ask Beacon’s centralized AI platform to analyze historical customer data, set goals and start developing new features. 

Mark Shulgan, a co-founder at Beacon backer Intrepid, admits he “walked in very skeptical of the AI story.” After all, every firm is talking up their automations right now. Beacon brought him around with product demonstrations and hard numbers. “The progress that they’re making is just incredible,” he says.

Let’s Camp started out in 2012 as an online booking form that founder Mike Yasieniuk’s software development firm built for a single campground client. The avid outdoorsman began selling the technology to other facilities, and “it eventually just started to snowball.”

In mid-2020, Yasieniuk spun Let’s Camp out as its own company, and the following year added a marketplace on which campers could find and book their stays. By late 2023, the firm needed investment to continue growing. It had also recently lost its sole salesperson and only other developer, leaving Yasieniuk as a one-person operation.

Venture capitalists expected Let’s Camp to increase its market by expanding into new areas. “When we started looking, it was hard,” he says, acknowledging that campground software is “a very small, niche market.”

Then, in the spring of 2024, Ganenthiran made a takeover offer. Selling to Beacon would give Let’s Camp the technical expertise and resources to expand, Yasieniuk decided. And the chance to “continue to run the company and be the visionary for it was appealing.” That July, Let’s Camp became one of Beacon’s first acquisitions.

Let’s Camp now has five full-time staff, and is on track to end the year with revenue up almost 300 per cent since the deal, according to Yasieniuk. That’s partly thanks to integrating other small acquisitions, but mostly from adding products and selling to new customers, like launching a boating platform for marina managers.

The firm also recently bid on a contract for Saskatchewan’s provincial park system. To win, Let’s Camp needed extra features like a waitlist tool, and a point-of-sale system for firewood, day passes and other concessions. Those upgrades had been on Yasieniuk’s longlist, but Beacon’s staff and AI tools built them quickly.  

On its own, Let’s Camp couldn’t “handle a project or a customer that large,” Yasieniuk says. “With the Beacon resources, we can scale up as needed.” His team now has over 1,000 campground clients.

Beacon has similarly helped ToolHound, based in St. Albert, Alta., to land and keep customers, says new CEO Ian Wilson. Founded by petroleum engineer Dean Perry in 1985, the firm sells software that companies in construction, mining and energy use to keep track of their tools and equipment. 

ToolHound had a loyal customer base, but had been slow to evolve because of financial and staffing constraints, according to Wilson. He added that Perry had been forced to spend recent years “putting out a lot of fires as opposed to trying to cut down new wood.” Now, the founder was ready to retire. In February, Beacon bought ToolHound and put Wilson—a member of its M&A team who’d worked on the deal—in charge.

ToolHound’s eight-person team soon began soliciting feedback and feature requests from its customer base. On one of those calls, Wilson learned that U.S. construction firm Bechtel, a major client, was preparing to move to a competitor. Bechtel handed over a list of several areas for product improvement. It had flagged those issues to ToolHound before, but “feature requests or small changes would take two to three years to show up,” Wilson says. Now, he claims, clients are “seeing prototypes show up in weeks.” 

Using Beacon’s team and technology, ToolHound is adding new AI functionality to its software so that customers can better monitor existing equipment, but also manage purchases of new stock and forecast future requirements. “The system that we’re trying to build for them is not just, ‘Where is my drill?’” says Wilson, “but, ‘How many drills do I need on Tuesday across all my job sites, and how am I going to supply those?’”

Following its acquisition by Beacon, ToolHound has managed to keep Bechtel and all of its other clients, he says. It also retained long-standing employees, whose expertise and relationships remain valuable, he adds. “That domain knowledge is fundamental.”

Software businesses have had a rollercoaster year, with investors trading down the stock prices of publicly-listed firms and marking down the valuations of private companies. The dips have been driven by fears that clients will use ever-improving AI coding agents to build their technology in-house, instead of buying it. 

Beacon, though, reckons it can avoid the so-called SaaS-pocalypse. “I do not subscribe to a world where people sit in a campground and say, ‘I’m going to vibe code everything,’” Ganenthiran says.

Main Street businesses will keep buying technology from vendors, he predicts. And he claims Beacon’s portfolio firms will be the ones to sell those tools because they have existing customer relationships and offer better service than Silicon Valley giants. Weeks, for one, sends suggestions straight to Yasieniuk. Let’s Camp’s founder is “immensely professional” and “very quick to respond,” Weeks says.

Beacon could also face a challenge from the AI model makers. Anthropic and OpenAI have developed tools tailored to sectors like finance, law and life sciences, and tied up with private equity investors to start services arms that can work with businesses to deploy their technology. Ganenthiran isn’t worried. The customer bases of Beacon’s portfolio firm are too small, and their needs too different to be served by any single technology platform, he says. “It would be a disaster.”

Plus, Beacon’s portfolio firms aren’t just looking to hold on to the likes of Peace Island Park and Bechtel. They’re also trying to sell them more software, and add more clients like them.  Ganenthiran says he sees a lot of room to grow, because of “the sheer volume of real economy that’s out there.” 

ToolHound is betting large customers will be prepared to increase their spend several-fold to use its planned AI features. Let’s Camp already has about a sixth of the North American campground market, but many potential clients are still using spreadsheets or pen and paper. In other markets, Beacon plans to pair software with offline services. For example, it owns firms that sell tech tools for homeowners associations in the U.S., as well as companies that send scrutineers to certify condo board elections to meet local regulations.

As Yasieniuk discovered a few years ago, the kind of niche software businesses that Beacon buys rarely meet venture capitalists’ investment criteria. Bundled together as part of Beacon, though, they’ve attracted huge sums. The firm announced a US$250-million Series B round last November, followed swiftly by a US$225-million Series C financing this June. Unlike its portfolio firms, Beacon is venture-backable because it can scale its business model significantly, Ganenthiran says. 

Intrepid—which closed its first, US$525-million fund in September—only backs companies that have AI at the core of their business, Shulgan says. And it only invests at the growth stage, once firms have figured out how to make real money. The investment firm has backed buzzy names like Blue J in tax law, StackAdapt in adtech and PhysicsX in industrial AI.

While Beacon targets tech-resistant markets and is less than three years old, it fits Intrepid’s criteria. The company is buying software businesses that are growing, profitable and defensible—a “super attractive” combination, Shulgan says. He also cited the rest of Beacon’s team, which includes co-founder Divyahans Gupta, a former Sequoia Capital partner; CTO Goutham Buchi, a former top engineering executive at AngelList and Coinbase; and COO Mark Schaaf, previously Instacart’s CTO.

Beacon’s CEO is key to the whole operation, according to Shulgan. He cites Ganenthiran’s long-term outlook, and his ability to convince selling founders that he will take good care of their firms. And Shulgan invokes a legendary figure known for buying high-quality businesses and holding them forever. “Sometimes when I think about Beacon, I think, ‘If Warren Buffett had AI, what would Berkshire Hathaway look like?’”

The Oracle of Omaha already has a Canadian successor. And anyway, Ganenthiran is focused on growing Beacon—acquiring more small software firms, adding AI and improving the tools available to more Main Street businesses. The company’s deal-making pace reached one a week this summer, he says. “We’ve just gotten more volume than I ever thought I’d see.”