Whilst businesses have been busy optimising for traditional search, voice AI has quietly evolved into something far more sophisticated – and far more powerful. This is backed up by the data. According to HubSpot's 2024 Consumer Trends Report, voice search and audio SEO opportunities are growing, with consumers increasingly relying on voice assistants for various tasks.
We've invited Jeremy Smith, founder of Neural Voice and voice AI expert, to share what he's seeing from the frontlines of this transformation. With years of experience building AI systems that handle genuine conversations rather than robotic commands, Jeremy offers a refreshingly honest perspective on where voice technology is actually headed.
From his work creating an AI Father Christmas that had meaningful conversations with hundreds of children, to building voice agents that genuinely understand context and emotion, Jeremy's insights cut through the hype to reveal what's actually working in voice AI today – and what businesses need to know to stay ahead.
In this guest piece, he explores why voice search has fundamentally changed, how AI agents are revolutionising customer interactions, and which industries are seeing the biggest wins from voice technology. More importantly, he shares the practical challenges nobody talks about and the real-world results his clients are achieving.
Whether you're curious about voice AI or wondering if it's right for your business, Jeremy's candid take will help you understand what's possible – and what's just marketing fluff.
Connect with Jeremy on LinkedIn or read on to discover why voice AI might be the competitive advantage you didn't know you needed.
Hi, I'm Jeremy, founder of Neural Voice. I've been obsessing over voice tech for the past few years, and I keep seeing the same pattern: businesses are missing what's happening right under their noses while people are already talking to their devices like they're old friends.
I wanted to share what I'm seeing from the trenches, because honestly, most of the advice out there about voice AI is either too theoretical or just plain wrong.
Source: Market.us
Remember when voice search meant shouting "PIZZA NEAR ME" at your phone like an idiot? Those days are dead.
Last December, we built an AI Father Christmas that ended up talking to hundreds of kids during the holiday season. Not scripted responses or keyword matching, genuine conversations. Kids would tell him about their pets, their school problems, what they wanted for Christmas, and he'd remember details from earlier in the conversation and reference them naturally.
One kid spent twenty minutes explaining why his little sister was annoying but also why he still wanted to get her a good present. Father Christmas didn't just listen, he asked thoughtful questions and helped the kid work through his feelings. The parents were amazed.
That's where we are now. People don't bark commands anymore, they have real conversations. And if your business content still sounds like it was written by a robot for robots, you're invisible to voice search.
Instead of optimising for "best pizza Camden delivery fast," you need to answer questions like "Where can I get pizza delivered to Camden that won't take forever?" See the difference? One sounds like a human asking a friend, the other sounds like a caveman with WiFi.
Here's where it gets interesting. Everyone keeps comparing what we're building to Alexa or Siri, but that's like comparing a smartphone to a calculator because they both have numbers on them.
Alexa is great if you want to set a timer or hear a dad joke. But try having an actual conversation with it about something complex, and you'll want to throw it out the window. It's built for quick transactions, not relationships.
Source: Neural Voice
What we're building at Neural Voice is different. Our AI agents actually remember what you said five minutes ago. They pick up on whether you sound frustrated or confused and adjust accordingly. They don't just answer questions, they ask better ones.
Source: Neural Voice
Last month, one of our agents was handling calls for a travel agency. A customer called asking about "a trip somewhere warm." Instead of just listing random beach destinations, it asked about their budget, travel dates, whether they had kids, if they'd been anywhere tropical before. By the end of the call, it had basically planned their entire honeymoon to Barbados.
Source: Neural Voice
The customer didn't even realise they were talking to AI until the follow-up email. That's when you know you've built something special.
What's made this leap possible? Honestly, it's been a convergence of several breakthroughs happening at once. Speech recognition accuracy has gone from "maybe it'll understand you if you speak like a robot" to genuinely reliable conversation-level understanding. Natural language processing can now handle context, interruptions, and even emotional undertones.
But the real game-changer has been latency improvements. Early voice AI had that awkward pause that made conversations feel stilted. Now our systems respond as fast as a human would, which makes all the difference in creating natural interactions.
The accessibility angle is huge too. Voice interfaces work for people who struggle with traditional UIs (User Interfaces), whether that's due to visual impairments, mobility issues, or just not being comfortable with technology. We've had elderly customers who never touched a computer have full conversations with our AI about their healthcare needs.
I'm going to be straight with you, most AI chatbots are terrible. We've all experienced that frustration of being trapped in some awful automated phone tree that clearly doesn't understand what you're saying.
But when voice AI actually works, it's transformative. That travel client I mentioned? They were missing 60% of after-hours calls before we implemented our system. Now they capture every single lead, qualify them properly, and their sales team wakes up to a CRM full of warm prospects instead of a pile of missed call notifications.
Source: Neural Voice
Another client in finance was spending thousands on a call centre just to book appointments. Our AI now handles that entire process, understands insurance questions, manages cancellations, and even follows up on no-shows. Their staff can actually focus on patient care instead of playing phone tag all day.
The ROI isn't theoretical, it's showing up in their bank accounts.
We're seeing massive adoption in healthcare, where our AI handles appointment scheduling, prescription refill requests, and basic symptom triage. One clinic reduced their administrative overhead by 40% just by letting our AI handle routine calls.
Travel and hospitality have been early adopters too. Hotels use voice AI for concierge services, travel agencies for initial booking consultations, and restaurants for reservations. The common thread? Industries where personal service matters but repetitive tasks eat up human time.
Source: Neural Voice
B2B sales is surprisingly hot right now. Our AI can handle initial lead qualification calls that used to require expensive sales development reps. It asks the right questions, understands complex business needs, and only passes through qualified opportunities.
Financial services are starting to experiment with voice AI for customer support and fraud alerts, though they're (understandably) more cautious about implementation.
Let's be honest, voice AI isn't all sunshine and unicorns. Privacy concerns are real, especially when you're handling sensitive customer conversations. We've built our entire platform with privacy by design but it's something every business needs to think through carefully.
There's also the uncanny valley problem. If your AI sounds too robotic, people get frustrated. If it sounds too human, they feel deceived when they find out it's AI. Finding that sweet spot takes real thought about your brand and customer expectations.
Integration complexity is another big one. Voice AI isn't just about the conversation, it needs to connect seamlessly with your CRM, scheduling systems, payment processing, and everything else. Half our implementation time is usually spent on those backend connections.
Source: Neural Voice
Source: Mckinsey
After building dozens of these systems, I've learned that the technical stuff isn't the hard part. The hard part is making it feel human without being creepy, helpful without being pushy.
Your voice AI needs to sound like that friend who actually knows what they're talking about, not like a robot reading from a script. It should understand context, if someone sounds stressed, don't chirp at them like a cheerleader. If they're clearly confused, slow down and explain things differently.
And for the love of all that's holy, integrate it properly with your actual business systems. I've seen too many companies bolt a chatbot onto their website and wonder why it doesn't help anyone. Your voice AI should be connected to your CRM, your calendar, your inventory everything. Otherwise, it's just an expensive toy.
Some of the experimental work happening right now is genuinely mind-bending. There's this thing called "vibe coding" where you describe what you want in natural language and AI generates the actual working code. It's not about being lazy, it's about focusing your time and energy on the creative aspects rather than getting stuck in technical details.
This has honestly unlocked my brain in ways I didn't expect. I can prototype new voice tools constantly now just by describing the vibe of what I want. "Make something that sounds like a wise grandfather but can handle technical support calls" and boom, I have a working prototype to test with real customers within hours instead of weeks.
We're moving toward a world where voice becomes the primary interface for everything, not just search. The businesses figuring this out now will have a massive head start over everyone else scrambling to catch up later.
I'm not going to oversell this, voice AI isn't magic, and it's not right for every business. But in the right situations, it's incredibly powerful.
We're seeing the biggest wins in inbound call handling, sales qualification, appointment booking, and first-level customer support. Basically, anywhere you have repetitive conversations that follow predictable patterns but still need a human touch.
One B2B client uses our AI to handle initial sales calls. It asks all the qualifying questions, understands complex business needs, and only transfers calls to human salespeople when there's a real opportunity. Their close rate went through the roof because their team stopped wasting time on tire-kickers.
The trajectory is pretty clear to me. We're moving toward voice becoming the primary interface for most business interactions. Not replacing humans, but handling the routine stuff so humans can focus on complex, high-value conversations.
Multimodal AI is going to be huge. Systems that can seamlessly switch between voice, text, and visual information in a single conversation. Imagine calling a business and being able to say "I'm looking at your website right now, can you help me understand this pricing table?" and the AI actually knowing what you're seeing.
Emotional intelligence in AI is advancing rapidly. Soon, voice assistants will be able to detect stress, frustration, or excitement in real-time and adjust their approach accordingly. This isn't just about better customer service, it's about creating genuinely empathetic interactions.
The ethical considerations are becoming more important, not less. As these systems get more sophisticated, businesses need clear policies about data use, conversation recording, and transparency about when customers are talking to AI versus humans.
Voice AI isn't some distant future tech, it's happening right now. The question isn't whether your business should care about it, but whether you want to be ahead of the curve or playing catch-up.
What we're building at Neural Voice is about creating genuine connections through technology, not replacing human interaction but enhancing it. When someone finishes talking to one of our AI agents, they don't think "wow, that was a good robot." They think "that was actually helpful."
If you want to see what your business would sound like with a voice AI that actually gets it, check out our Pilot Programme. Six months, fixed cost, full support, and if it doesn't work, you haven't lost anything except the regret of not trying sooner.
Anyway, that's my take. Happy to chat more if you want to dig deeper into any of this stuff.
Jeremy
As Jeremy mentions, one of the biggest hurdles with voice AI isn't the conversation itself – it's making sure everything connects properly behind the scenes. Your CRM, booking systems, payment processing, inventory management – half the implementation time is usually spent on these backend integrations.
That's where we come in. Working with Neural Voice, we specialise in ensuring your voice AI integrates seamlessly with your existing business systems, to make sure conversations translate into revenue driving action.