Richard Conway is the Founder and CEO of The Optimisers and serves on the IAB Emerging Technology Council. With over 25 years of experience in digital marketing and search, Richard is one of New Zealand's leading voices on AI-driven search evolution, helping Kiwi businesses navigate the rapidly shifting landscape of discovery, visibility and agentic commerce.
Google I/O 2026 was not just another round of AI product announcements. For marketers, publishers, agencies and digital teams, it was something more important. A clear signal that Google sees the next era of search as conversational, multimodal, personalised and increasingly agentic.
In layman’s terms, people will not just search for information. They will ask, compare, plan, refine, delegate and, in some cases, transact. The internet is evolving, and Kiwi businesses would be wise to evolve with it.
For the last 25 years, search has largely been built around a simple exchange. A person types a query. Google returns links. The person clicks, compares and decides. That model is not disappearing overnight, and the data does not support the lazy “SEO is dead” narrative. But the shape of the journey is changing quickly.
At I/O, Google announced that AI Mode has surpassed one billion monthly active users, while AI Overviews now has more than 2.5 billion monthly active users. Google also says AI Mode queries have more than doubled every quarter since launch, and that overall Search queries reached an all-time high last quarter.
AI search is now becoming mainstream, and the trajectory is sharply inclining. The more useful question is what that means for Kiwi marketers.
One thing to note however, is that not everyone is happy with this transition, TechCrunch reported a 30% increase in DuckDuckGo (still a tiny market share) installs as some searchers defect from Google to the privacy focused alternative.
The big shift: from answers to actions
The most important message from Google I/O was not that AI answers are growing, (That much should be obvious to anyone). The shift is that Google is trying to move Search from an answer engine to an action engine. I predict the big evolution to agentic commerce and delegated decisions with acceleration in earnest in 2027.
AI Mode is becoming the multi-modal conversational layer. Google already has the lions share of the data and audience, the move is keeping users. The search box is being redesigned to accept text, images, files, videos and Chrome tabs. Users will be able to move from an AI Overview into a follow-up conversation in AI Mode, with context carrying through. Google is also introducing Search agents that can monitor the web in the background and send updates when something changes.
I remember when Google coined the term ‘micro-moments’. Being in the right place at the right time and directing a searcher to a solution immediately. That idea is changing; agents turn search into something that can persist over time.
A home buyer can ask an agent to monitor houses that meet a detailed set of criteria. A traveller can ask for options that match budget, dates and preferences. A business buyer can ask for suppliers that meet compliance requirements, service location and evidence of experience. A student can use AI Mode to explore, quiz and refine. A shopper can compare, check availability, monitor prices and eventually buy.
This is where the implications become much bigger than SEO, the definition of which is not yet clear. Agentic SEO, AEO or even Search Everywhere Optimisation. Exciting times!
If an AI agent is helping a customer decide, the brand must make sure that their information is readily available, formatted and trusted in the way that agents want. If the agent is comparing suppliers, your proof points need to be structured, credible and easy to retrieve. If the agent helps someone transact, your product data, booking paths, forms, feeds, reviews and availability must be clean enough for machines to use. Agentic readiness is the next critical aspect of search that New Zealand business should be all over.
Search is becoming multimodal
Google’s own AI Mode data shows that people are no longer searching traditionally, more than one in six searches in the US now use voice or images, and that image searches are growing by more than 40% month on month. It also says the average AI Mode query is three times the length of a traditional Search query. This shift in searchers behaviour is critical for marketers to understand, a short keyword search may be something like “best running shoes”.
An AI Mode search is likely to be longer, something like; “I am training for my first half marathon, I overpronate slightly, I run mostly on pavement, I want something under $250 and I care more about comfort than speed.”
That second query isn’t just longer. It contains intent, context, constraints, preferences and commercial signals. For marketers, this means the old keyword-only mindset has evolved and we must evolve with it. Keyword data is still relevant, but so is the need to understand the real questions people ask when they are exploring, deciding and doing. Think about your own search behavior, it has likely changed in a similar manner.
Google says planning queries in AI Mode have grown 80% faster than AI Mode queries overall in the past six months. Brainstorming queries have grown 30% faster than overall AI Mode queries since launch. The nuance here is that people are now using search to think, not just search.
Search is no longer one channel. It is a portfolio of discovery surface.
Traditional search is not dead
I’m sick of seeing these clickbait headlines from supposed ‘thought leaders.’ That’s why I love the Datos/SparkToro Q1 2026 State of Search report as it cuts through the hype. It shows AI adoption is accelerating, but it also shows that traditional search remains remarkably resilient. This matches our own internal data.
In the US, AI tools’ share of desktop events increased from around 0.41% in January 2025 to 0.93% by March 2026. On a quarterly basis, AI tools increased from 1.31% in Q1 2025 to 1.65% in Q1 2026. In the EU and UK, AI tools nearly doubled from 0.54% in January 2025 to 1.08% by March 2026. That is growth, but it is still small compared with traditional search.
The same report shows Google still held 94.3% of US desktop search share in March 2026 and 95.5% in the EU and UK in early 2026. Google engagement also remained strong, with US searches per Google desktop searcher rising from 95.3 in December 2025 to 102.5 in March 2026.
Even zero-click data is more nuanced than the popular narrative. In the US, Datos reports that zero-click searches declined from 24.5% in December 2025 to 22.4% in March 2026, while organic click-through increased from 42.0% to 44.9%.
So the answer is not “Google is finished”.
The answer is more interesting.
Traditional search is still huge. AI search is growing quickly. And the way people express demand is changing.
Both things are true.
AI Mode is small in behaviour data, but large in strategic importance
One of the most interesting tensions is between Google’s user numbers and third-party clickstream data.
Google says AI Mode has more than one billion monthly active users. Datos shows Google AI Mode adoption on desktop grew quickly in Q1 2026, but from a very low base. In the US, AI Mode increased from 0.06 per cent of desktop search activity in December 2025 to 0.16 per cent in March 2026. In the EU and UK, it increased from 0.06 per cent to 0.21 per cent over the same period.
That is exactly why marketers need to avoid two mistakes.
The first mistake is dismissing AI search because current measured share is small.
The second mistake is pretending it has already replaced everything else.
The smart position is somewhere in the middle. AI Mode is not yet the dominant source of measurable traffic, especially when we are looking at desktop-only data. But it is growing quickly and, more importantly, it is changing the interface through which people ask questions.
The trajectory matters.
This is bigger than Google
Another trap is to treat AI search as a Google-only issue.
The Datos report shows ChatGPT remained the leading AI tool in Q1 2026, ending March at 34.80% share in the US and 44.82% in the EU and UK. Gemini is growing strongly, reaching 16.06% in the US and 18.88% in the EU and UK by March. Claude also accelerated sharply, reaching 8.54% in the US and 9.61% in the EU and UK.
That matters because visibility is fragmenting.
A customer may start with Google. They may ask ChatGPT. They may compare on Reddit. They may watch a YouTube review. They may check TikTok for social proof. They may use Gemini because it is embedded in a Google product. They may rely on Claude for deeper research.
At The Optimisers, we call this Search Everywhere Optimisation.
It is not about abandoning SEO. It is about expanding the definition of search.
Your website still matters. Technical SEO still matters. Content still matters. Links still matter. But so do YouTube, Reddit, reviews, product feeds, entity signals, third-party mentions, structured data, comparison content, creator content and how consistently your brand is represented across the open web.
Search is no longer one channel. It is a portfolio of discovery surfaces, we describe this as Search Everywhere Optimisation (SEO!)
Agentic commerce changes the game for retailers
For commerce brands, one of the most significant I/O announcements was Universal Cart. Google describes Universal Cart as an intelligent shopping hub that works across merchants and Google services. Users will be able to add products while browsing Search, chatting with Gemini, watching YouTube or reading Gmail. Once an item is in the cart, Google says it can monitor deals, price drops, price history, stock availability and compatibility issues.
Google also continues to push Universal Commerce Protocol, or UCP, as a standard for agentic commerce. UCP is designed to enable purchases across AI Mode in Google Search and Gemini, starting with direct buying. Google says merchants remain the merchant of record, keeping control of customer data and relationships.
For New Zealand and Australian marketers, we get to wait and see, many features roll out gradually in larger markets such as the US first. I enjoy playing with Google betas, using a VPN, and I would advise interested marketers to do the same, however we do have the benefit of waiting until there is a definitive roll out in this part of the world.
The more practical lesson is this: commerce infrastructure is becoming marketing infrastructure and agentic readiness is going to be pivotal in taking advantage of this agentic world. Product feeds, stock data, price data, returns information, payment options, merchant reputation, fulfilment speed and review quality are no longer just operational details. They are signals that AI systems and agents can use when helping a customer decide. Making sure your data is formatted correctly, cleaned up and accessible to agents is the moat for the future.
Ads are moving into the conversation
Google’s ad announcements point in the same direction. At Google Marketing Live, Google said it is testing new ad formats built with Gemini, including Conversational Discovery ads and Highlighted Answers in AI Mode. It also announced AI-powered Shopping ads and a Business Agent for Leads, where a user can chat with a brand agent inside the ad experience. Google says these formats will remain clearly labelled as Sponsored and aligns with what we are seeing with ChatGPT ads (which you can now access in New Zealand). The transition from keyword led ads to this new format is especially interesting and relevant for marketers.
If someone is having a detailed planning conversation with AI Mode, the advertising opportunity is not just the original query. It is the context that emerges through the conversation; the user’s preferences, constraints, objections and next action become part of the commercial signal. This is both an opportunity and a threat for marketers; an opportunity for those that see it and act on it, and a threat to those that put their head in the sand and ignore it!
Weak landing pages, thin product data, generic creative and poor measurement will be will have a big (negative) commercial impact because AI-assisted advertising will reward brands that can give the system better inputs.
The web itself is becoming more machine-readable
One of the less consumer-facing but very important I/O announcements was WebMCP.
Google describes WebMCP as a proposed open web standard that allows websites to expose structured tools, such as JavaScript functions and HTML forms, to browser-based AI agents. The goal is to let agents complete tasks with more reliability and precision. For most marketers, this sounds technical. But the implication is simple.
Websites are no longer just pages for people to read. They are becoming systems that agents can interact with.
That means the future of web performance is not just page speed, UX and conversion rate, although all of those still matter. It is also whether machines can understand what your site offers, what actions are possible, what information is reliable and what pathways are safe to complete.
The best websites will be built for humans and agents.
Publishers still have a role, but loyalty matters more
There is a legitimate concern from publishers and content creators that AI answers may reduce traffic to original sources. Google is clearly aware of that tension.
After I/O, Google announced that Preferred Sources are coming to AI Overviews and AI Mode, allowing users to more easily spot links from sources they have selected. Google also says it is adding more prominent article carousels and expanding “Highly Cited” labels to help users find original and influential reporting.
This does not solve every publisher concern. But it does point to a useful direction.
In an AI-mediated search environment, brand preference becomes a visibility asset.
It is not enough to be indexed. It is not enough to publish. Media brands, creators and businesses need audiences that actively recognise, choose and trust them.
That is a very old marketing principle coming back in a very modern format.
Measurement is the weak link
Here is where we audience should lean in.
The search and advertising interface is changing faster than most measurement systems can cope with.
The IAB State of Data 2026 report says up to 75% of the buy-side believes leading advanced measurement approaches underperform on rigor, timeliness, trust and efficiency. It also notes that only 39% use incrementality testing, attribution and MMM together, despite each answering a different measurement question.
That is a problem in a world where the customer journey is becoming more fragmented, more conversational and less click-linear.
If AI Mode shapes consideration but the user later converts via brand search, how do we value the AI interaction?
If Reddit, YouTube and creator content influence the assistant’s recommendation, where does that show up in a dashboard?
If an AI agent shortlists three providers before the user ever reaches a website, what does “first click” even mean?
This is why the measurement conversation needs to mature quickly.
The IAB report also says marketers expect AI to improve advanced measurement by automating more strategic work, increasing measurement frequency by two to three times, and enabling more sophisticated approaches. But it also warns that legal, security, accuracy and data-quality concerns could inhibit progress.
That is exactly the balance the industry needs.
AI can help with measurement. But only if we have clean data, clear governance and human accountability.
What marketers should do now
Marketers must be selective about what they implement, there are so many new things to look at, I always say the best way is to prioritise by impact. Build the foundations that make your brand easier to find, trust, compare, cite and act on. We call this agentic readiness, in reality its also good business practice.
Answer Ready Content - Content that answers real questions in plain language, with clear structure, evidence, authorship and freshness is what will win out. Get rid or evolve thin category copy and generic blog posts where there is little to no real value or originality.
Build Action-Ready Infrastructure - Product feeds, location data, booking paths, lead forms, stock information, pricing, policies and service details need to be accurate, accessible and consistent. If agents are going to act, they need dependable inputs.
Build Trust and Authority - AI systems and humans both look for corroboration. Reviews, credible mentions, YouTube content, Reddit discussions, comparison pages, local citations, expert commentary and trusted third-party references all matter. AI agents care what others think of your brand, not the hype that you espouse on your own platforms!
Measure AI visibility separately from traditional SEO - Track whether you appear in AI answers, which sources are cited, how your competitors are framed, whether the answer is accurate, and whether your brand sentiment is positive, neutral or negative. Then triangulate that with organic demand, branded search, referral traffic, direct traffic and conversion quality. Remember however that AI tracking is more general than SEO tracking. AI is probabilistic (the answer changes) and people write longer prompts in a multitude of different ways.
Remember the Basics - Google is not going away anytime soon. Traditional search is not dead, just evolving. Technical SEO, page experience, useful content, internal linking, schema, site architecture and conversion optimisation are as relevant as ever.
The brands that win will not be the ones that replace everything with AI. They will be the ones that integrate AI into a strong digital foundation.
The Future
Google I/O 2026 showed us a future I’ve been talking about for a while no, one where search is less about typing keywords into a box and more about asking an intelligent system to help complete a task. That task might be learning, comparing, planning, buying, booking, researching, creating or deciding. For me the future is exciting and full of opportunity. because better intent creates an environment where we can know our customers better; a richly described AI Mode query tells us far more about the customer than a two-word keyword ever did.
One of the difficulties we now face is attribution and imperfect prompt tracking. I don’t see a short-term resolution to that. Strategy, data, content, commerce and measurement are all changing before our eyes and search is evolving to become the front end of intent and action.
The businesses that adapt early will not simply “rank” better. They will be easier to understand, easier to trust and easier for both humans and machines to choose. The next 12 months is going to be fun.