Earlier this year, we were named finalists in two categories at the APAC Search Awards 2026, reflecting our expertise across both paid and AI search.

The Qwilr project pushed value-based bidding into territory that most B2B advertisers haven’t explored, and the Sharesies work tackled AI misinformation, a problem most brands don’t even know they have yet. Being recognised for that work by the APAC Search Awards is a proud moment for our team.

The first campaign tackled one of B2B marketing’s most persistent challenges: scaling qualified leads without watching cost per acquisition spiral out of control.

Qwilr, an Australian B2B SaaS company selling proposal software across five international markets, had reached a plateau with their Google Ads performance.

Lead volume looked healthy, but the proportion of qualified leads that matched their ideal customer profile had stagnated. While increasing spend led to more form submissions from prospects, improvement in actual pipeline was slow.

The problem lay in Google’s algorithm optimising for the wrong outcome. Every signup looked identical, so the budget naturally flowed towards leads that were the easiest to capture, not necessarily the ones most likely to become real customers.

Atlas Digital’s solution involved aligning Qwilr’s PPC strategy with genuine business outcomes, not just form submissions.

The Atlas team connected Qwilr’s predictive lead scoring model in HubSpot directly to Google Ads, so when a lead passed a specific qualification threshold, it was automatically imported as a high-value conversion event. By shifting campaigns from maximising conversion count to maximising conversion value, Google’s algorithm learned to prioritise the best-qualified leads.

Atlas Digital’s second shortlisted campaign addressed an emerging challenge that few brands have solved: what happens when AI platforms spread misinformation about your products?

Sharesies, a New Zealand investment platform, was seeing rapid growth in traffic from ChatGPT. But during an initial audit, the Atlas team found that ChatGPT was providing false information to users about Sharesies’ KiwiSaver funds.

For a regulated financial product, this wasn’t just a brand problem, but a compliance risk.

Commercial AI sentiment analysis tools couldn’t help, as they lacked the ability to trace misinformation back to its source. So Atlas developed an in-house approach, using AI agents to diagnose where false claims were originating and strategically inform repair.

The methodology captured ChatGPT’s source citations and traced misinformation backwards with precision. The team identified patterns across both controllable sources (help centre articles, app user guides) and external publishers. Sharesies implemented internal corrections within four weeks, and Atlas initiated publisher outreach with URL-specific correction requests.

The misinformation was corrected, so where ChatGPT previously stated incorrect fund numbers, it now provides accurate information plus an increase in engagement rates on AI platform traffic and monthly brand exposures.

The project highlighted the fact that GEO (Generative Engine Optimisation) is not just SEO with a new name. Companies can’t optimise for AI search without knowing what AI platforms are saying about them or where the information is coming from.

Sydney based AI-native startup MiALaw just raised $2m in seed funding, including private backers from Citibank Hong Kong and JP Morgan Australia. They’ve built a “legal expert system” that encapsulates knowledge provided by a human legal expert in a specific domain to infer solutions to problems. Their solution is designed to “strengthen professional reasoning, not replace it,” according to founder Laina Chan. Speaking to Startup Daily, Chan said,

“There was no AI system that embedded legal reasoning into its architecture. Law is not just retrieval. It is reasoning from facts and principles. MiAI Law is built to follow that same discipline.”

An expert system consists of a knowledge base, an inference engine and a user interface. the knowledge base stores declarative knowledge of a specific domain, which encompasses both factual and heuristic information. The inference engine holds procedural knowledge: the set of rules and the methodology for reasoning. It combines facts provided by the user with information from the knowledge base.

Inference is done using predefined rules according to the expert and with logical statement evaluations. Classes of problems that can be solved using expert systems include classification, diagnosis, monitoring and prediction.

Another example is Heidi’s latest launch Evidence, marking their evolution from a clinical documentation service to a real-time clinical reasoning partner. According to Heidi, “medical knowledge doubles every 73 days, making it impossible for clinicians to stay updated with new methods and research. While many have turned to general-purpose AI, these tools lack transparency and local clinical context.” Evidence solves this by providing a clinical-grade research tool integrated directly into the Heidi platform.

This week, Xero announced a deal with Anthropic. The multi-year deal will bring Claude into Xero and Xero financial data into Claude.ai for real-time financial intelligence for its clients. According to their responsible data use commitments, financial data shared between the platforms is used solely for the user's specific session, i.e. proprietary business data is never used to train Claude’s AI models. Their press release also noted that as part of the deal, Xero's engineering teams will also use Claude and Cowork to accelerate their own product development. Accounting firms asking themselves “what are we doing about AI” will be chomping at the bit to use this service and how it changes the small business owner/accountant relationship will be one to watch.

2026 brings another incredible lineup of companies, proving the value of the tech sector to New Zealand and the world, not just monetary value but solving problems across all of society and business.

Several of Atlas Digital current and past clients have been nominated this year. We’re proud to have been their growth partner on the journey to scale.

Lucy Turner from VXT is up for Xero’s Hi-Tech Young Achiever award and VXT is in the running for Duncan Cotterill’s Most Innovative Hi-Tech Software Solution. Starshipit has been nominated for the ASX Hi-Tech Emerging Company of the Year award, too.

Congratulations and we will see you at the Gala Dinner on the 22nd of May at Spark Arena.

For SaaS leaders who care about revenue, not vanity metrics. Ideal for CMOs, heads of growth, marketing leads and CROs.

Most tech businesses know their key homepage conversion points, but almost all fail to optimise the entire buyer journey. As markets get more crowded, systematic conversion rate optimisation (CRO) becomes a genuine growth advantage, not a nice‑to‑have.

What we’ll cover:

Join to learn why CRO is an offensive growth play for scale‑ups and enterprise SaaS companies and how much revenue is typically left on the table.