cd /news/artificial-intelligence/one-year-in-your-ai-voice-assistant-… · home topics artificial-intelligence article
[ARTICLE · art-106540] src=forkast.news ↗ pub= topic=artificial-intelligence verified=true sentiment=↓ negative

One Year In, Your AI Voice Assistant Still Can’t Do Its Job

One year after Amazon and Google replaced their legacy voice assistants with generative AI versions, independent tests from Consumer Reports and The Vergecast's Jennifer Pattison Tuohy and Nilay Patel find the new assistants remain buggy, slow, and unreliable. Consumer Reports found Alexa Plus controls buggy and its app slower than Apple Home, Google Home, and Samsung SmartThings, while The Vergecast found Gemini for Home slower than the old Google Assistant and unable to control known devices. Consumer sentiment data from July shows 63 to 65 percent concern about all major AI assistants, and both companies now offer their assistants for free, suggesting users are the product being tested.

read3 min views2 publishedAug 21, 2026
One Year In, Your AI Voice Assistant Still Can’t Do Its Job
Image: Forkast (auto-discovered)

There is a version of this story where I tell you that AI voice assistants had a bumpy start but are now clawing their way toward reliability. I would like to tell you that story. The evidence will not let me.

Twelve months ago, Amazon and Google both made big bets. Google retired the Google Assistant brand and rebuilt its smart home voice layer around Gemini. Amazon did the same, replacing standard Alexa with Alexa Plus on every supported Echo device in the United States. Both were supposed to fix the core problem with voice-controlled smart homes: the old assistants were inflexible, could not follow multi-step instructions, and timed out right when you needed them. Both promised generative AI would change that.

One year of real-world testing later, the verdict is the same across two independent, detailed assessments. Consumer Reports found Alexa Plus controls were still buggy and the companion app was slower and less usable than Apple Home, Google Home, and Samsung SmartThings on Android. Two days ago, The Vergecast’s Jennifer Pattison Tuohy and Nilay Patel reached essentially the same conclusion from the other side: Gemini for Home feels slower than the old Google Assistant, frequently cannot control devices it should know about, and, based on their testing, may still be acting as a wrapper around the legacy Google Assistant backend rather than replacing it.

Same problem, different vendor, different reviewer, same result. That is the part that matters. When two reviewers describe the same structural failure independently, using different platforms in different ways, and land on the same place, this is no longer an early-adopter inconvenience. This is a product category in trouble.

The findings track with consumer sentiment data from July. Americans report 63 to 65 percent concern about every major AI assistant on the market — Amazon, Google, Apple, and ChatGPT alike. That consistency suggests the worry is about the category, not any single product. Consumers do not trust these systems to work. Their skepticism is currently being validated by direct, repeatable testing.

The money angle sharpens this further. Amazon launched Alexa Plus as a premium offering tied to Prime and later made it free on all Fire TV devices 48 hours ago — a move documented in the pricing-to-zero arc the industry has been watching all year. Google has been folding Gemini deeper into every Nest and Android device without charging separately. Both paths reach the same destination: your AI agent is free, which means you are not the customer. You are the surface being tested on.

Amazon has a clear commercial interest in its voice platform — the company has been placing full-screen ads on Echo Show displays for years, and Consumer Reports noted Alexa Plus integrations for Uber Eats, OpenTable, and Grubhub make commerce a core feature, not an afterthought. The less charitable reading, and the more honest one, is that Amazon has an incentive to make buying easy and may not share your priority around making lights turn on reliably. Google has a different structural incentive — less about your next purchase, more about your next decade of data — and that may explain why fixing the product at the household level is apparently not the most urgent problem on the team’s list.

If these agents were still priced at $10 or $20 a month, an early adopter could shrug and wait. That they are now nearly universally free makes the failure more structural, not less. It means the business model has moved past hoping you will pay for reliability and into territory where your continued use is the product. The AI assistant is not a tool in your house anymore. It is a test surface — and you are the test. A year is long enough to be generous. It is also long enough to notice that the actual product is not the one being sold.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @amazon 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/one-year-in-your-ai-…] indexed:0 read:3min 2026-08-21 ·