Every day, millions of people interact with Atolls platforms, leaving behind a treasure trove of data in the form of searches, clicks, and discussions.

We have an amazing team of data engineers and analysts who are skilled at navigating this data, but there’s only so much they can do. A single deep dive into one part of the data could easily take days, if not weeks, even with the help of ‘standard’ AI tools.

In recent weeks, that changed.

With our new internal AI chat tool, connected via Model Context Protocol (MCP) directly to our data warehouse, it’s now possible to have a real-time conversation with our data. I ask a question in plain English, and seconds later, I’m looking at patterns across millions of data points. Something that once required multiple tools, queries, and manual cleaning can be done from one interface and with natural language.

This shift doesn’t just mean faster answers; it fundamentally changes the kinds of questions we can ask. We’re no longer constrained by pre-built dashboards or static reports. Instead, we can follow our curiosity wherever it leads, often uncovering opportunities we didn’t even know to look for.

To show what’s possible, here are three insights I uncovered in a single afternoon from seven months of hotukdeals data (January–July 2025).

Insight 1: LEGO’s Year-Round Demand

LEGO’s popularity on hotukdeals is undeniable. It consistently ranks among the top searched brands and products on the site — roughly on par with “iPhone” — and unlike most toys, its search volume remains remarkably consistent throughout the year. Most toys follow predictable peaks and troughs: big surges around new launches, Christmas, or school holidays, and quieter periods in between. LEGO breaks that mould. The search pattern looks more like a staple commodity; something consumers are actively seeking all year round, regardless of season or sales events. But it’s not just about search volume. The LEGO audience on hotukdeals is highly organised.
  • One user maintains a monthly thread of retailer exclusives, alerting the community when availability changes.
  • Others use historical price trackers and community verification before committing to a purchase.
  • There’s a culture of sharing niche buying strategies and insider tips.
What it means for brands: LEGO’s strength isn’t just its brand, per se; it’s the community around it. By cultivating collectors rather than just casual buyers, brands can tap into an audience that is always “in market” and actively looking for ways to engage. For other businesses, the lesson is clear: invest in building passionate, informed customer groups that keep your product top-of-mind year-round.

Insight 2: Community Behaviour Around the Nintendo Switch 2’s Launch

The Nintendo Switch 2 launched on 5 June 2025, but the data shows the community’s interest started long before that. After the official April announcement, searches saw a 500x spike before settling and then surging again at launch. Beyond search volume, the community’s conversations revealed three key behaviours:
  1. Rapid technical analysis – Within days, users were benchmarking the Switch 2 against competitors like the Steam Deck, complete with detailed performance breakdowns.
  2. Immediate compatibility workarounds – Forum members found that third-party controllers could be made compatible with simple firmware updates, sharing step-by-step instructions.
  3. Creative pricing strategies – Some users devised upgrade and resale tactics to secure games at significantly lower net prices.
What it means for brands: Tech-savvy consumers are not passive recipients of marketing messages. They are proactive, informed, and collaborative. In some ways they act like a live, public R&D team. Rather than ignoring these grassroots discoveries, brands can lean into them: formalise compatibility features, acknowledge community pricing hacks, and engage in open dialogue to shape future products.

Insight 3: Air Conditioning Buyers Know Their Specs

When hot weather in the UK arrived in June, searches for air conditioning units on hotukdeals grew by more than 4,300% compared to winter. That seasonal jump isn’t surprising, but the depth of consumer knowledge was.

Discussions went far beyond “what’s the best (or cheapest) AC unit?”:

  • One user mapped the manufacturing origins of budget portable AC units, revealing they came from a “very respectable” factory that supplies multiple high street and online brands.
  • Others offered guidance on refrigerant safety, F-Gas regulations, BTU sizing for different room sizes, and even local noise compliance.
  • Practical insights came from real-world comparisons. For example, one user explained why a quieter, smaller 5,000 BTU unit might be better than a 12,000 BTU one for certain homes.

What it means for brands: These buyers don’t want vague marketing claims; they want specs, sourcing details, and honest, experience-based advice. Meeting them with transparency and detail builds trust, loyalty, and long-term brand equity.

The Bigger Picture: Speed + Depth = Competitive Advantage

What previously took me weeks to uncover, I can now find in a single afternoon. The combination of AI and direct data access has made consumer intelligence faster, more flexible, and far more accessible across the company.

And it’s not just about efficiency, but also opportunity. We can generate insights to improve our products. We can help brands understand what consumers are looking for. And I can find stories in the data to share with the world.

We’re still early in our AI journey at Atolls and I’m excited about the prospect of what’s to come.

One more thing – journalists, get in touch. If you’re looking for unique, data-backed insights to inform your stories, look no further. Whether it’s toy trends, tech launches, or consumer electronics buying patterns, I’d be happy to explore the questions that matter most to you.

A longer version of this article was originally published on LinkedIn here.