Voice search has come a long way since we first asked Siri for the weather or told Alexa to play our favourite song. But if you’ve tried using voice assistants for anything more complex lately, you’ll know they often fall short. Stilted responses. Misunderstood questions. No memory. Not a hint of joining the dots between one question and another.
Yet at the same time, the rise of AI voice search — using tools like ChatGPT and emotionally aware companions like Romi — is showing us a glimpse of something far more powerful. Something more human.
In this article, I’m exploring where voice search is failing, where’s it’s quietly flourishing, and why the brands that adapt their voice search SEO strategy now will be the ones that are heard loud and clear as the future of voice search takes centre stage.
Alexa, are you OK?
Let’s be honest. Alexa is a bit rubbish.
If you’ve ever asked her a question and been hit with “Here’s something I found on the web…” followed by a completely irrelevant or outdated answer, you’re not alone.
She forgets what you said five seconds ago, can’t follow up with context, and half the time sounds like she’s reading from a broken autocue. It’s not exactly the future we were promised.
We’ve all yelled “ALEXA STOP!” like she’s an annoying housemate who just won’t take the hint.
By contrast, if you have used ChatGPT’s voice mode, you’ll have noticed a huge leap in the spoken search experience. It listens. It remembers. It responds naturally. You can have a flowing, conversational exchange that feels — dare we say — intelligent. Even empathetic.
So just why is Alexa still so flat and forgetful, while newer voice experiences feel like actual conversations?
The answer lies in what’s powering them. Traditional assistants like Alexa and Siri weren’t built on modern large language models (LLMs). Instead, they rely on pre-coded commands, keyword triggers and shallow integrations with the web. In contrast, conversational AI voice search tools are trained on massive datasets, context-driven prompts and real-time adaptation.
That means when you ask ChatGPT something, it’s not just matching your words to a database. It’s interpreting your intent. And responding accordingly.
The gap is widening. And for brands trying to stay visible in this space, that matters.
What’s actually broken with voice search?
Why does traditional voice search still feel so clunky?
The short answer is: most voice assistants weren’t built to chat. They were built to follow commands. “Set a timer.” “Play some jazz.” “What’s the weather?” Simple stuff.
But when you ask a real question — something that needs context or up-to-date knowledge — they start to crumble.
Most assistants:
- Don’t remember what you said a moment ago
- Can’t handle follow-up questions
- Pull outdated or surface-level answers from the web
- Sound robotic and lack any real understanding of tone or emotion
And the big problem? People don’t trust them. When answers are too vague, too flat, or just plain wrong, we stop using voice for anything meaningful. It becomes a novelty — not a tool.
And that’s the problem. This lack of trust doesn’t just hurt the user experience. It hurts brands, too. Because if your content is buried behind a bad interface, your voice isn’t getting heard.
But here’s the thing: when people give up on voice assistants, they’re not giving up on asking questions — they’re just turning to smarter, more intuitive tools instead. And that’s where brands need to be.
Enter the new wave of voice search
Romi, ChatGPT and emotionally aware voice AI
While voice search as we know it is struggling, a new generation of voice tools is starting to change the game.
Take Romi — a palm-sized robot developed in Japan and launched this summer with the goal of providing emotional support through natural conversation. I met him when I attended the Osaka World Expo earlier this year with Emma and our son.
Unlike old-school voice assistants, Romi can remember previous chats, respond with empathy, and even adapt to your mood. With Romi, it’s not just about answering questions any more. It’s about building a relationship.
Then there’s ChatGPT’s voice mode, which many people are discovering for the first time via the app. You can speak naturally, ask follow-up questions, even change topics mid-stream — and it keeps up. It might not be perfect, but it feels intelligent, responsive and remarkably human.
As I said, these newer tools are different from Alexa, Siri & Co because they’re powered by LLMs (large language models) — the same technology behind tools like ChatGPT and Google’s AI Overviews. They’ve been trained on vast amounts of real-world content, which means they’re capable of giving more helpful, relevant responses.
And they’re getting better fast.
Google is currently testing Project Astra — a prototype voice assistant with memory, live vision and the ability to respond in real time. And, Google Audio Overviews are being trialled as new AI voice search tools, offering spoken summaries that feel more like a podcast than a search result.
I personally see this shift as way more than just an upgrade. It’s changing the whole trajectory of the future of voice search. Moving it from tool to companion. From utility to personality. From commands to conversations.
And as this shift takes hold, it’s clear that brands need to rethink their voice search SEO approach. Because it’s no longer about being found through rigid voice commands. It’s about being part of the conversation.
Voice search in action: a real-life test
I wanted to see how today’s voice assistants really stack up, so I did a side-by-side test.
I asked three simple questions to two very different tools: Alexa (the original pioneer of the smart speaker world) and ChatGPT’s voice interface, powered by the same kind of large language models now driving AI search.
These weren’t trick questions. Just the sort of thing a business owner might ask — one practical, one topical, one factual.
The results? Let’s just say one of them earned a gold star. The other… got told to be quiet. Again.
Question 1: “How do I get my business found online?”
✅ ChatGPT’s response:
A full, conversational roadmap. Six detailed, structured steps. Practical tips. Contextual examples. Encouragement. Jargon-free explanations. And even a polite offer to personalise the advice if we wanted.
Excerpt:
“If you want, tell me a bit about what your business does, where you are based, and who you want to attract, and I will happily suggest a tailored plan.”
It sounded like a helpful human — or at the very least, a switched-on marketing mentor.
❌ Alexa’s response:
“Search engine marketing helps your business website be found online.”
Thanks Alexa. We’ll just… Google the rest, shall we?
Question 2: “Is AI coming for my job?”
[I got our Head of Content Sarah to ask this one, as the role of the copywriter is one of those ‘billed’ as in the AI firing line.]
✅ ChatGPT’s response:
A thoughtful, emotionally intelligent answer. A breakdown of what AI can and can’t do.
Reassurance that creativity, nuance and judgement are still in demand. And even a witty metaphor about power tools and carpentry thrown in for good measure. You’ve got to love the sense of humour.
Excerpt:
“Think of AI like a power tool: if someone doesn’t know how to build a table, the drill won’t save them. But give the drill to a talented carpenter? They’ll create more, faster, better. So the future of your job? It’s evolving, sure. But for a writer who champions clarity, perfect English, and brand voice with personality — you’re not in trouble. You’re in demand.”
❌ Alexa’s response:
“I’m not quite sure how to help you with that.”
Exactly the kind of shrug-inducing reply that’s caused people to stop using voice search on smart speakers altogether! And for the record, Sarah was thrilled with the supportive answer from ChatGPT ;)
Question 3: “How tall is The Shard?”
✅ ChatGPT’s response:
“The Shard in London stands at 309.6 metres tall — making it the tallest building in the UK. If you meant different height-measurements (roof, observation deck, etc.), I can also list those.”
[Included pictures.]
❌ Alexa’s response:
“The Shard is 306 metres tall.”
It’s close — but not quite right. It certainly didn’t offer follow-up detail, alternatives, or any ability to clarify. And it definitely wouldn’t get you the point in the pub quiz.
So, what does this tell you about the future of voice search?
The difference is stark.
One assistant gave clear, useful answers in full sentences — adapting to the question and the person behind it. The other offered vague, often outdated answers with no context and no conversation.
That’s the heart of the shift in AI voice search.
The tools we used to rely on (Siri, Alexa, Google Assistant) are being left behind. Because today’s users want more than robotic replies. They want human-sounding answers, built on real knowledge they can trust.
And that means if you want your business to be part of the future of voice search, your content needs to be found by these smarter systems. Not just the old ones.
Which brings us neatly on to what that means for visibility and voice search SEO — and we’ll get into that next.
The crux of the matter: what does this all mean for brands – and why you should still care about voice search SEO
Voice search is growing up.
We’re moving from robotic, command-based assistants to intelligent tools that understand context, tone and intent. And that means people are expecting more from the answers they get.
They want clarity. Depth. Trust.
So what does that mean for your content?
It still needs to show up — in all the systems people use. That includes traditional search engines, AI tools like ChatGPT, and emerging voice-first platforms.
But now, it also needs to sound human.
Gone are the days when voice search SEO meant stuffing in long-tail keywords like “best accountant for freelancers in Kingston”. Today, it’s about answering real questions — clearly, conversationally, and consistently.
Whether someone types, taps or talks, your authority needs to be visible. If you’re not publishing clear, helpful answers in your area of expertise, you’ll simply be bypassed.
Google’s AI Overviews, Siri, Alexa, ChatGPT — they’re all scanning the web for quality, relevance and expertise. Not just content that hits the right keywords, but content that answers questions in a way that feels trustworthy.
It’s less about targeting a device, and more about shaping the knowledge graph that powers them all.
And to do that, we need to think beyond what a client’s audience types into Google. We need to consider what they say and how they say it, and how the machines are interpreting it all.
This is where voice search SEO overlaps with something much bigger: LLM optimisation. It’s about showing up not just when people type a query, but when they ask it out loud, or prompt an AI assistant for guidance.
The format might change. But the need to be found? That’s not going anywhere.
Final thoughts: speak now, or be forgotten
Voice search isn’t dead. It’s just being reborn.
We’re leaving behind clunky commands and robotic replies, and heading into a world of intelligent, memory-enabled, emotionally aware assistants.
The question isn’t “Should I care about voice?” — it’s “Will I still be visible when my audience starts talking instead of typing?”
Whether they’re using Google, ChatGPT, or whatever comes next, your expertise needs to be ready and waiting. Clear. Credible. Conversational.
At Figment, we help expertise-led businesses stay visible across all formats — from human-first SEO to LLM optimisation and content that actually earns trust in a voice-first world.
Want to stay discoverable as voice evolves?
Let’s talk about making your content future-ready.
