Clickocalypse, AI Shopping, SMB Uncertainty, Coders the New Milkmen

Google AI Mode: 'Goog Enough'?

We don't yet know how the majority of consumers will respond to Google's new AI Mode. We do know that, when present, Google's AI Overviews depress click-through rates – sometimes quite dramatically. The purpose of AI Mode, in essence, is to give users enough of the "ChatGPT experience" to make going to ChatGPT unnecessary, and thereby retain users and ad revenue. Google's entire product strategy these days is basically a reaction to competitors and competition, something it would aggressively deny.

The company inserted forums and other "first person" content into search results more aggressively in 2023 after it determined (younger) users were defecting to TikTok and Instagram. And after ChatGPT's November 2022 launch it went "code red" and began to rearchitect the entire search experience because it perceived an existential threat. AI Mode is a useful complement to search and provides a better UI/UX than traditional search in most respects; we can debate the quality of results. But AI Mode will quickly evolve and change so generalizations about results quality are somewhat pointless right now. It's good enough ("Goog enough") to probably maintain the inertia of Google usage for many people. The US Justice Department was concerned that Google would use its existing massive distribution (search, mail/Workspace, Chrome, Android) to squelch nascent competition in AI. And that's exactly what it's trying to do.

AI Overviews and AI Mode, assuming favorable user response to the latter, will combine to drive far fewer clicks to publisher sites for upper funnel and potentially mid-funnel search queries. This may not matter as much for local businesses (who mostly fulfill offline) but it could matter to brands and publishers of all stripes. Shall we call it "The Clickocalypse"? Eventually it could also matter to Google if paid-CTRs decline.

Shopping: AI Use Cases Expand

When we released the results of our initial consumer survey in November last year, there was perhaps one other published survey that investigated AI as a search alternative. Now there are multiple surveys asking some version of these three questions: 1) are you using AI?, 2) what are you using it for? and 3) how does it compare to Google? Our data showed that people were using AI for a wide range of functions. Asked which tool they preferred, Google or ChatGPT, across various research, planning and shopping scenarios, Google "won" (see below). But there was a substantial minority (29% - 43%) who chose ChatGPT. Most "AI vs. Search" surveys are finding the same thing directionally. There are skeptics who think the survey data is essentially worthless and Google's YoY search volume growth negates these survey findings. But there are too many consistent results, telling a version of the same story.

The latest of these is from Adobe (n=5K US adults), which offers insight into how consumers have embraced AI for product research. Adobe found that 39% of US adults had used AI for online shopping, with more expecting to do so this year. According to the survey, people are using AI for 1) product research, 2) product recommendations, 3) deal-finding, 4) ideas for gifts, 5) to look for unique products, 6) and to create shopping lists. This doesn't mean they're not still using Google. But it may mean their usage frequency is less or they're migrating certain types of queries to ChatGPT. There's another school that argues for "a bigger pie," meaning ChatGPT/AI search is simply additive to Google.

I don't think either the zero-sum or bigger pie arguments capture what's actually going on. I think consumer behavior is evolving rapidly and so are the products. Currently AI offers features that Google doesn't have – but will with AI Mode. People say they like AI's speed/efficiency, voice mode, perceived personalization and the conversational UX. Our survey supports this. People like the "direct answers over website links, conversational interaction, more comprehensive information, and the ability to ask follow-up questions."

Google has long battled Amazon in product search, with the latter being a more common "starting place" than Google. This is a kind of predictive model: people use both Google and Amazon for product search, as they probably will with ChatGPT and Google. But top and mid-funnel search queries may erode away from Google. Indeed, the product research and comparison experience on AI is better than on Google. If better local store and inventory data come to AI this very important category – product search – may become even more competitive.

Small Biz Optimism Down

Small business owners are optimists. That shows up in practically every SMB survey you'll see, including ours. For example, an early 2025 survey of 500 small businesses (<500 headcount) from Clutch finds that "Over 4 in 5 SMB decision-makers express optimism about [ ] growth ... in the coming year." That's saying about 85% expect to grow. Significantly, however, that same survey identifies the top business challenge of 2025 as "economic uncertainty." The other top challenges are: 2) artificial intelligence technology adoption, 3) talent acquisition and retention, 4) increased competition and 5) cybersecurity threats. Indeed, the persistence of inflation, the growing threat of recession and the issue of tariffs is quickly dampening business optimism. Declining consumer/customer spending was not mentioned in the Clutch survey (perhaps SMBs aren't yet seeing it). But it has fallen sharply, as has confidence in the first two months of 2025. Dialog's 2024 SMB surveys found that 75% of SMBs expected strong or measured growth, in Q1, and 59% said the same thing in Q3. The major difference between the two Dialog surveys is that the Q3 survey focused on very small businesses and the Q1 survey included larger SMBs. But the decline from Q1 to Q3 may also reflect increased economic uncertainty, which has only worsened since the election. The NFIB Small Business Optimism Index (January data) also showed a decline in SMB confidence and optimism. The Optimism Index declined while the parallel Uncertainty Index grew to its "second highest recorded reading."

What are we to make of this? There are two potential interpretations, assuming uncertainty doesn't improve. The first is upbeat for SaaS and technology vendors: SMBs will look increasingly to technology, including AI tools, for cost savings, increased productivity and greater efficiency. The less sanguine view is that there will be a pullback and belt tightening. In this view, SMBs will be more reluctant to take risks, experiment with new technologies and expand or even sustain current marketing spending levels. This might then require more effort, persuasion and demonstrated value from vendors and agencies.

News & Noteworthy

• ChatGPT can now become the default assistant on Android.
• Claude can now search the web.
• Anthropic CEO wants to give AI the equivalent of worker's rights.
• Instagram encouraging people to write comments using AI.
• AI helping new startups grow faster, generate revenue faster.
• AI will be at the center of future franchise/fast-food restaurants.
• Nvidia and GM partner for "next-generation" vehicles.
• The low-cost AI positioning alternative that may replace GPS.
• Art created by AI, without human input, cannot be copyrighted.
• The uses and limitations of synthetic AI training data.
• Google Workspace is adding a boatload of new AI tools.
• Driverless Zoox challenges Waymo in San Francisco (WSJ).

Coders Are the New Milkmen

Only about four years ago everyone was talking about how all kids should to learn to code. You remember that, right? Then, last year, Nvidia CEO Jensen Huang said programming was no longer a skill people should learn because AI was taking over. Anthropic CEO Dario Amodei recently echoed that position, claiming that within a year (maybe sooner) 90% of code will be written by AI. Will coding/programming go the way of the switchboard operator or milkman? The World Economic Forum's (WEF's) Future of Jobs Report 2025, identified software and app developers as the fourth fastest growing job category of the future. Yet a Washington Post analysis found that more than 27% "of all computer programming jobs have vanished in the past two years, the worst downturn that industry has ever seen." This is not exclusively a result of AI but AI is a key factor. The way to reconcile these findings may be semi-semantic. The Post makes a distinction "programmers" and "software developers." The former are primarily focused on coding, while the latter have a broader skill set – though Dario Amodei independently said AI would negatively impact "software development" jobs. Regardless, numerous "white collar" and technical jobs that survived multiple decades of outsourcing are now increasingly vulnerable to AI. Coding is just one of them.

Funny | Disturbing | Sad

• OpenAI and Google argue for the dilution of copyright for AI training.
• Google naming of Gemini features (e.g., Canvas) duplicates ChatGPT.
• Netflix's "Electric State" uses AI – in a movie that criticizes technology.
• Columbia undergrad using AI to help cheat in virtual interviews.
• Everything you say to Alexa will now go to Amazon.
• Super-intelligent AI is the new nukes, coming sooner than we think.

If you're not a member and would like to explore Dialog membership, contact us. In 2025, we have expanded benefits but will limit new members.

If you're not getting this in your in-box, you know what to do...