Will microphones become the new keyboards?

While generative AI is undeniably disrupting how humans work, what they value and how they interact, it doesn’t act in isolation. Instead, it’s embedded into and merges with existing technologies.

 

In some cases, gen AI simply adds a natural language interface, creating more effortless and smoother user experiences. In other cases, however, the impact of this additional layer is nothing short of revolutionary.

 

For example, an interface that can translate between SQL – a query language for interacting with databases – and natural languages radically democratises access to data.

 

Gen AI also breathes new life into technologies that had already plateaued before the pre-ChatGPT era.

 

Examples abound. Tools for handling unstructured data – emails, transcripts or social media content – did exist prior BCGPT (Before ChatGPT), but relied on manual processes and rigid rule-based algorithms.

 

Similarly, fraud detection systems that had been used in banking for decades produced false positives at rates of around 50 per cent, requiring human validation. Hyper-personalisation in marketing, targeting individuals, or “a segment of one”, was always an ambition, but tools fell short of achieving it.

 

How gen AI has transformed voice assistants

 

Before the gen AI upgrades, trust in voice assistants was declining.  According to the 2024 PYMNTS Intelligence Report, GenAI and Voice Assistants: Adoption and Trust Across Generations, between February 2023 and May 2024 the number of people believing that voice assistants could ever become as smart and reliable as humans fell from 73 to 60 per cent. Diminishing consumer enthusiasm about the technology was further highlighted by more than a quarter of respondents being explicitly sceptical about the technology’s future potential.

 

But what participants overlooked was precisely how GPT models’ language processing and contextual understanding could become game changers – through voice commands. Having realised the true potential that gen AI presents, Amazon has since upgraded its voice assistant Alexa multiple times to improve its conversational ability and context awareness and to even generate original content.

 

Meanwhile, Google, the other dominant player in the market, plans to replace its voice assistant with Gemini for Home from October 2025, designed to make household management – including shopping – more intuitive.

 

Voice for shopping

 

Voice commands already play a role across much of the shopping journey. They are most effective in restocking, where the product to be ordered is already specified. For product searches, voice assistants, being optimised for conciseness and simplicity, typically return one definitive answer at a time – Amazon’s Choice or Google’s top match.

 

Voice assistants are also well suited to retrieving order-tracking information, which Google integrates through Gmail and Amazon by connecting it to user accounts.

 

The strongest online use case, however, lies in shopping lists. Here, voice commands shine by enabling spontaneous, multitasking-friendly interactions. Realising mid-cooking that the olive oil is nearly gone and shouting “add olive oil to the shopping list” is far more intuitive than having to wash your hands, find your phone then dig around trying to update your order.

 

Thanks to heightened security and privacy risks, payment is the weakest link. Currently, only Alexa offers a completely screen-free checkout, secured by a preset voice code and verbal confirmation.

 

A physical store killer use-case

 

Anecdotally, we spend 6.8 months of our lives searching for misplaced items, which is surely only compounded by the time we waste wandering aisles hunting down shopping list entries.

 

But while a trolley that guides customers directly to products may sound futuristic, the building blocks already exist.

 

Waitrose is trialling an AI smart trolley system powered by ShopE computer vision, while Morrisons offers an in-app product finder. With a digital device clipped to it, any shopping trolley can be upgraded to guide customers on their shopping trip, as well as to capture the trail they leave behind digitally.

 

Although its current verbal directions may sound cumbersome – “with your back to the checkouts, near the end of aisle 14, on the left-hand side” – combined with indoor location technology leveraging Bluetooth, their accuracy could reach within three metres.

 

But even delivering a simple message at the right place and time amounts to a technological feat. Behind the scenes, Morrissons has migrated comprehensive product data, including availability and location, to Google Cloud’s Big Query, which also enables SKU and store layout updates in real time.

 

The app relies on Google Gemini for reasoning and contextual understanding, and there are long-term plans to add voice-enabled, hands-free navigation to the product finder.

 

Can voice replace screens?

 

While voice-enabled shopping’s main selling point is convenience, it enhances inclusivity too. For people with physical impairments that limit screen use, voice can be a gateway to the digital world.

 

Google Wallet already offers accessibility features such as TalkBack, a screen reader and Voice Access, to help visually or manually impaired users make payments.

 

And as smart speakers lead to better user experiences and earlier frustrations fade, the number of people preferring voice as their communication channel of choice is also expected to grow – particularly among younger generations.

 

To tap into the pool of consumers who prefer speaking to typing, online retailers must increasingly integrate voice commerce into their omnichannel strategy, optimising their websites for voice search, embedding voice chat and enabling voice payments.

 

But will microphones replace keyboards entirely, making voice the primary way humans interact with their devices? That remains a question only voice-controlled digital crystal balls could answer.