Google Bard is Google’s answer to ChatGPT. On 6 February 2023, Sundar Pichai announced Bard, an experimental AI chatbot built on Google’s LaMDA language model. It is open only to a small group of trusted testers for now, and Google says the public will get access “in the coming weeks”.
If you run a website or manage SEO, the big question is simple. Will an AI chatbot from Google take clicks away from your pages? In this guide, I explain what Google actually announced, what we still don’t know, and what I think you should do this month.
What Google Announced on 6 February
The announcement came in a post on Google’s blog called “An important next step on our AI journey”. It covered three things, and I think it helps to keep them apart.
- Bard: a conversational AI service you can chat with, powered by LaMDA.
- AI features in Google Search: new ways to sum up complex topics inside the normal results page, “rolling out soon”.
- Tools for developers: a Generative Language API, first powered by LaMDA, that Google plans to open to developers, creators and businesses next month.
Most headlines focus on Bard. For SEO, I think the second point matters more. A separate chatbot is one thing. AI answers inside the search results you already rank in are another.
You can read Google’s announcement post yourself. It is short and worth five minutes.
What Bard Is
Google calls Bard an “experimental conversational AI service”. You type a question or a request, and it writes an answer in plain language. You can then ask follow-up questions, the same way you would with ChatGPT.
Google gave a few example uses. One was explaining new discoveries from NASA’s James Webb Space Telescope to a 9-year-old. Another was learning about top football strikers and then getting drills to build your own skills.
What Powers Bard
Bard runs on LaMDA, short for Language Model for Dialogue Applications. LaMDA is a large language model (a system trained on huge amounts of text so it can predict and write natural sentences). Google first showed it at Google I/O in 2021.
Bard launches with a lightweight version of LaMDA. Google says this smaller model needs much less computing power. That lets Google give it to more people and collect more feedback.
Google also says Bard draws on information from the web to give fresh, high-quality answers. This is a clear point of difference from ChatGPT, which, as launched, has a knowledge cutoff and does not pull in live web pages.
Who Can Use It and When
Right now, you probably can’t. Bard is open to trusted testers only. Google says it will combine their feedback with its own internal testing before a wider release “in the coming weeks”.
Google has not given an exact date, a list of countries or languages, or a price. I would not plan around any of those details until Google confirms them.
Why Google Is Moving Now
The timing is not a surprise. ChatGPT launched at the end of November 2022, and within weeks it became the tool everyone was talking about. People started asking whether a chatbot could replace a search engine for some questions.
In January, Microsoft also announced a new multi-year, multi-billion dollar investment in OpenAI, the company behind ChatGPT. Microsoft runs Bing, so the link between chatbots and search became very real, very fast.
Google has built AI into Search for years. The announcement mentions BERT and MUM, two language models that already help Google understand queries. But those work in the background. Bard is the first time Google is putting a chat-style AI product in front of everyday users.
What Google Bard Means for SEO
This is where I want to be careful. Bard is not live for the public, and the Search features have not rolled out. So anything about impact is informed opinion, not data. Here is how I am thinking about it.
Bard Itself Is a Separate Product, for Now
From what Google has said, Bard is its own experience, apart from the main search results page. That means your rankings in Google Search do not change on the day Bard opens to the public.
The risk is slower and more indirect. If people get used to asking a chatbot instead of searching, some questions may never reach the results page. In my experience, the first queries to move would be simple ones: definitions, quick explanations, short how-to answers and idea lists.
AI Summaries in Search Are the Bigger Change
Google says AI-powered features in Search will soon “distill complex information and multiple perspectives into easy-to-digest formats”. Its example was a question with no one right answer: is the piano or the guitar easier to learn, and how much practice does each need?
Google says these features will help people get the big picture first, then learn more. It also gave examples of where people might go next. These include finding other perspectives, such as blogs from people who play both instruments, and getting steps to start as a beginner.
That detail matters. Google is describing an AI summary that still points people toward the web. If it works that way, pages that offer real experience and a clear next step have a reason to be clicked.
Questions With No Single Right Answer
Today, Google handles simple facts well with featured snippets and knowledge panels. Questions that need opinions, trade-offs and context are harder. People usually open several pages to piece the answer together.
Those are exactly the questions Google says AI will help with. Some examples of this kind of query:
- “Is it better to rent or buy a car for a small business?”
- “Which is easier to learn, Python or JavaScript?”
- “Should I use WordPress or Shopify for a small online shop?”
If your content wins traffic from comparison and “which is better” queries, I’d watch these features most closely once they roll out.
The Click Question
Fewer clicks for some queries is a real possibility. We already live with zero-click searches (searches where the answer shows on the results page, so nobody clicks). AI summaries could push more searches in that direction.
But nobody knows the size of the effect yet. Google has not said how answers will credit sources, how links will look, or how often the AI features will appear. I would ignore anyone who quotes a precise traffic drop at this stage.
Bard vs ChatGPT: What We Know So Far
People keep asking how the two compare. Based only on what has been announced, here is the honest picture.
| Point | Google Bard | ChatGPT |
|---|---|---|
| Company | OpenAI | |
| Language model | Lightweight version of LaMDA | A GPT-3.5 series model |
| Uses live web information | Google says it draws on web information | No, it answers from its training data |
| Who can use it | Trusted testers only | Open to the public since late November 2022 |
| Link to a search engine | Made by Google, with AI features also coming to Search | Standalone chat product |
I can’t tell you which gives better answers, because almost nobody outside Google has used Bard yet. I’ll test it properly once it opens up.
How I’m Preparing My Sites
I don’t think anyone needs to panic or rebuild their strategy this week. But I do think this is a good moment to tighten a few things that already matter. Here is what I’m doing.
- Sort your traffic by query type. In Search Console, export your top queries from the Performance report. Tag each one as a simple fact, a how-to, a comparison or opinion question, or a brand or product search.
- Find the pages most at risk. Pages that win clicks for quick definitions or short answers are the easiest for AI to sum up. Make a list of them and note how much traffic each brings in.
- Add what an AI summary can’t. On those pages, add first-hand experience, your own photos or screenshots, real test results, worked examples and clear opinions. Give people a reason to click past a summary.
- Build pages around a next step. Google’s examples point people toward blogs with real perspectives and beginner steps. Pages with a clear action, such as a guide, a tool, a template or a product, should hold up better than thin answers.
- Strengthen who is behind the content. Add or improve author bios, explain how you tested or researched things, and keep contact and about pages up to date.
- Grow channels you own. Email lists, direct visits and returning readers don’t depend on how Google shows its results. I’d put a little more effort here in 2023 either way.
- Set a baseline now. Note your current clicks, impressions and click-through rate for your main page groups. When AI features land in Search, you will want a clean “before” to compare against.
None of these steps depend on how Bard turns out. They make your site stronger whatever Google launches next.
Experience and Helpful Content Matter More Now
Google’s recent changes already point in the same direction as this announcement. In August 2022, Google launched the helpful content update, which targets content made mainly to rank rather than to help people.
Then in December 2022, Google added “Experience” to E-A-T, making it E-E-A-T. Google’s raters now also look at whether content shows first-hand experience of the topic.
Put those together with AI summaries and the picture is clear to me. Generic content that repeats what is already online is easy for an AI to replace. Content built on real use, real testing and real opinions is much harder to replace.
Google’s own guide on creating helpful, reliable, people-first content has a set of self-check questions. I run my most important pages through it at least once a year. This is a good time to do it.
Should You Use AI Tools to Write Content?
Many marketers are already testing ChatGPT for drafts, and Bard will add another option. I get asked about this a lot.
Google’s spam policies already cover automatically generated content made mainly to manipulate rankings. And the helpful content system looks at whether content is made for people. So the risk is not the tool itself. The risk is publishing lots of thin, unchecked pages because a tool made them cheap.
My rule of thumb is simple. I’m happy to use AI for outlines, idea lists or a rough first draft. But a person who knows the subject must check every fact, add real experience and stand behind the final page. If a page adds nothing a chatbot couldn’t say on its own, I don’t publish it.
What We Still Don’t Know
The announcement leaves many SEO questions open. These are the ones I’ll be watching:
- Source credit: Will Bard and the AI features in Search link to the pages they draw on, and how visible will those links be?
- Coverage: Which types of queries will show AI summaries, and in which countries and languages?
- Reporting: Will Search Console show any data about clicks or impressions from AI features?
- Accuracy: How will Google keep answers correct, especially for health, money and other sensitive topics? Google says it wants a high bar for quality, safety and groundedness, but hasn’t shared details.
- Publisher controls: Will site owners get any way to choose how their content is used in AI answers?
Until Google answers these, I’d treat any strong prediction, good or bad, with care.
Mistakes I’d Avoid Right Now
- Rushing to rewrite everything. Bard is not public and the Search features have not launched. Make steady improvements, not a panic rebuild.
- Mass-producing AI content to “keep up”. This is the fastest way to end up with exactly the kind of content Google is trying to filter out.
- Chasing “Bard SEO” tricks. There is no proven way to optimise for a product almost nobody can use yet. Be wary of anyone selling one.
- Ignoring the change. The opposite mistake is just as risky. Search is clearly moving toward more AI answers, so start the groundwork now.
Conclusion
Google Bard is Google’s first public step into chat-style AI, and it signals where Google Search is heading. Bard is still limited to testers, but Google has also promised AI summaries inside Search soon. That is the part I’d prepare for.
The best preparation is not new. Know which queries bring you traffic, make your key pages clearly more useful than a summary, and show real experience. As a next step, export your top queries from Search Console this week and tag them by type. Then fold the results into your wider plan using my SEO priorities for 2023.
