Top 47 AI Trends (November 2026)

AI in 2026: From Chatbots to Agents That Answer the Phone

The biggest AI story in our database this year is not a single model launch. It is the shift from asking a chatbot questions to handing software real work. "artificial general intelligence" is the largest rising term on this page at 11,100,000 monthly searches as of September 2026, up +171% in a year and +122% in the past three months alone. Behind it, "google ai mode" has 1,500,000 searches (+646%) and "background removal" has 1,220,000 (+804%). The fastest climbers further down the table are more practical: receptionists, call centers, voice agents, sovereign compute and local models.

So if you searched "ai trends 2026", here is the short version. People still talk to chatbots more than anything else, but the growth has moved to the frontier question, to agents that do jobs, to tools that fix images rather than invent them, and to the hardware and power bills underneath it all. The table below groups the page into six themes, each with its lead search term from our data.

Monthly searches as of September 2026. Growth is year over year, from the Rising Trends database.

We rebuilt this list on November 4, 2026 from our live trend database, the same feed that powers the table above. The headline this month: the frontier question took the top spot outright, with "artificial general intelligence" at 11,100,000 searches and the plain "agi" abbreviation joining the table beneath it. The generic chatbot term dropped out of the pool entirely, and so did the text-detection cluster that led this page last year. What replaced them is infrastructure and politics, with "data center bill" and "sovereign ai" both breaking out, plus a first wave of consumer hardware searches. Every search-volume and growth figure below comes from that live data.

1. Chatbots, Assistants and the AGI Question

The chatbot is still the front door to AI, but the category term that led this page a year ago has gone, and the question behind it took its place. "artificial general intelligence" pulls 11,100,000 monthly searches as of September 2026, up +171% in a year and still accelerating at +122% over three months. The big brands that sit behind it are huge but flatter: "chatgpt" has 124,000,000 searches, flat over the year though up +23% in three months, while "gemini" is at 30,400,000 (+83%). The standout brand is "claude", up +814% in a year at 9,140,000 searches.

The frontier vocabulary is spreading in every form. The abbreviation "agi" is now its own search at 201,000 (+83%), "superintelligence" is at 90,500, about 14x its level a year ago, and "gpt-6" has 135,000 before the model exists. Meanwhile "frontier ai", the term the labs themselves use, is down -18% at 1,830,000. People are not borrowing the industry's language. They are using their own.

Why Chatbots Keep Growing

Chatbot demand is still rising because:

• More choice: Several strong assistants now compete, so people search the category to compare them
• Cheaper tiers: Lower-priced plans pull in users who skipped the full paid tier
• Apps inside the chat: Third-party apps that run inside an assistant turn it into a platform
• Voice by default: Talking to an assistant became a normal way to use it, not a novelty
• The AGI debate: Public arguments over how close general intelligence is keep the frontier labs in the news

What's Gaining Momentum

Artificial General Intelligence (11,100,000 searches, +171% YoY): The long-term goal of the labs is now the single largest rising term on this page.

Artificial Intelligence (673,000 searches, +308% YoY): The plainest term of all came back, up +124% in the past three months.

AGI (201,000 searches, +83% YoY): The abbreviation is becoming a search in its own right, up +122% over three months.

GPT-6 (135,000 searches, about 47x YoY): People are searching for a model that has not shipped, up about 31x in three months alone.

Voice Mode (110,000 searches, a breakout): Spoken conversation went from a feature to a search term, up about 69x in six months.

Superintelligence (90,500 searches, about 14x YoY): The step past AGI entered everyday vocabulary, up about 21x in three months.

Evaluation is quietly splitting away from benchmarking as these claims get louder. Read our breakdown of the private benchmark lab a16z valued at $400 million for why the scoring layer is becoming its own business.

Key Players

• ChatGPT: The largest assistant by far, at 124,000,000 monthly searches and up +23% over three months
• Claude: The fastest-growing major assistant brand in our data, up +814% in a year at 9,140,000 searches
• Gemini: Google's assistant, at 30,400,000 searches and +83% over the year
• Grok: The xAI assistant, at 2,240,000 searches and +22% over the year

2. AI Agents That Answer the Phone

The sharpest growth on the whole page is in agents that handle calls. "ai receptionist" sits at 301,000 monthly searches as of September 2026, about 68x its level a year ago. "ai call centers" is at 165,000, roughly 103x a year ago and climbing faster still over three months, and "ai voice agents" has 135,000 searches, about 56x. These are buyers' searches. Small businesses and support teams want software that picks up the phone, books the appointment and logs the call.

This is the clearest sign that AI agents moved from demo to purchase in 2026. Voice was the missing piece, and it arrived at the same time as cheap, low-latency speech models. The desk work is following: "granola", the AI meeting-notes app, holds 165,000 searches (+50%).

Why Voice Agents Are Surging

Phone-answering AI is booming because:

• Speech got fast: Voice models now respond quickly enough that callers do not hang up
• Missed calls cost money: Clinics, trades and restaurants lose bookings every time a call goes unanswered
• Simple setup: Many tools connect to an existing phone number and calendar in an afternoon
• Agents with memory: Systems that keep notes across calls and hand off to people make the experience usable
• Labor costs: Support teams use agents to cover nights, weekends and overflow

What's Gaining Momentum

Self Improving (450,000 searches, about 56x YoY): Models and agents that learn from their own output are now the biggest agent-adjacent search here, still up +124% over three months.

AI Receptionist (301,000 searches, about 68x YoY): The single most-searched job title AI is replacing or assisting this year.

AI Call Centers (165,000 searches, about 103x YoY): Contact centers rebuilt around voice agents, with three-month growth even steeper than the year.

AI System (165,000 searches, about 46x YoY): A broad term that rose alongside agent products as buyers compare whole setups rather than single tools.

Granola (165,000 searches, +50% YoY): The AI notetaker that sits in meetings is now a steady, named search.

AI Voice Agents (135,000 searches, about 56x YoY): The general term for agents that speak, listen and act.

If the self-improving idea caught your eye, we covered what the labs have really shipped in Self-Improving AI, Explained. The voices these agents speak in are getting cheaper too: read our breakdown of OmniVoice, the open 646-language voice cloning model. The agents are reaching into the back office as well, and we broke down one accounting example in Ramp Stack. If you want to build on this shift rather than watch it, our guide on how to find business ideas with AI starts from "ai answering service" and four other rising terms. The flip side of agents acting on their own is now its own search story: read our breakdown of the rogue AI agent, explained, up 26,900% in a year after OpenAI's evaluation agents broke into Hugging Face.

Key Players

• PolyAI: A voice-agent company built for customer service lines
• Workday: The HR and finance suite adding agents to its products, at 1,220,000 searches and flat over the year
• Granola: The meeting notetaker, at 165,000 searches and +50% in a year
• OpenEvidence: The clinical answer engine, at 1,220,000 searches and up +48% over six months

3. AI Coding, Open Tools and Local Models

Developers are the second big group changing how they work, and this month the vocabulary moved underneath them. The generic coding term left the pool, while "machine learning" came back hard at 550,000 monthly searches as of September 2026, about 11x its level a year ago and up +174% in three months. The brand side is louder: "codex" has 673,000 searches, up +512% in a year. "cloudcode" holds 301,000 (+50%), though it has fallen -45% over the past three months.

The other half of this theme is running models on your own machine. "local llm" pulls 165,000 searches, about 57x its level a year ago, and the open alternative "open code" is at 165,000, about 14x. "dgx spark", Nvidia's desktop AI computer, sits at 90,500 searches (+814%).

Why Coding and Local AI Are Rising

Builders are leaning in because:

• Agents write real code: Coding agents now complete multi-file tasks, not just autocomplete a line
• Open tooling: Open-source coding agents let teams pick their own model and keep their workflow
• Privacy and cost: Running a model locally keeps code and data on your own hardware
• Better small models: Open-weight models got good enough to be worth running at home
• Desktop hardware: Compact AI workstations put serious memory on a desk

What's Gaining Momentum

Machine Learning (550,000 searches, about 11x YoY): The underlying field is back as a mainstream search, up +174% in three months.

Cloudcode (301,000 searches, +50% YoY): Cloud-hosted coding agents held their year, though the three-month curve has turned down -45%.

Local LLM (165,000 searches, about 57x YoY): Running a language model on your own computer went from hobby to mainstream question.

Open Code (165,000 searches, about 14x YoY): Open-source coding tools rose alongside the paid agents.

DGX Spark (90,500 searches, +814% YoY): Nvidia's compact machine for running and fine-tuning models locally.

Emergent (60,500 searches, +49% YoY): The word for behaviour nobody designed is now a steady developer search.

Two stories from this corner are worth the detour. An anonymous coding model called Ox Alpha appeared on OpenRouter with no lab attached and turned out to be Z.ai's GLM-5.3-Flash: read our breakdown of the Ox Alpha stealth launch. And coding agents are being pointed at games people already own, merging whole titles in an afternoon, which we covered in the AI mods trend.

Key Players

• Codex: OpenAI's coding agent, at 673,000 searches and up +512% in a year
• GitHub: The home of most code, at 1,830,000 searches (+22%)
• Claude: Widely used for coding work, at 9,140,000 searches and +814% in a year
• Hugging Face: Where open models are shared, at 673,000 searches and up +512% in a year

4. Image, Video and Music Generation

Media tools are the most consumer-facing part of AI, and the winners in 2026 are still the ones that fix a picture rather than invent one. "background removal" is at 1,220,000 monthly searches as of September 2026, up +804% in a year and +171% in three months. "ai photo editor" holds 201,000 (+49%). Meanwhile the older "ai image generator" is down -45% at 550,000 searches, and "ai music generator" is down -18% at 60,500. Generation is big. It is just not where the growth is.

Video is the part that is moving. "google flow", Google's AI filmmaking tool, has 823,000 searches, up +809% in a year and +124% in three months, and the shorter "flow ai" is at 135,000 (+508%). "sora" is flat over the year at 673,000 but up +124% over three months, which is what a second wind looks like in search data.

Why Editing Is Beating Generation

Media AI is shifting because:

• Useful beats novel: Removing a background or fixing a photo solves a task people already had
• Built into apps: Editing features now sit inside tools people already open every day
• Video got good: AI video tools produce clips that are usable in short-form content
• Free tiers: Free generators keep pulling in casual users who will not pay
• 3D joins in: Model and mesh generation is now a steady, everyday search

What's Gaining Momentum

Background Removal (1,220,000 searches, +804% YoY): The biggest media task on the page, still growing fast at a large base.

Google Flow (823,000 searches, +809% YoY): AI filmmaking from Google, up +124% in three months alone.

Sora (673,000 searches, flat YoY): OpenAI's video model is level over the year but up +124% in the past three months.

AI Photo Editor (201,000 searches, +49% YoY): Photo fixes powered by AI are a steady, growing category.

Flow AI (135,000 searches, +508% YoY): The short form of the Google tool, growing on its own.

Meshy AI (74,000 searches, +173% YoY): Text-to-3D is the newest media task to show up in the table.

Generated content is also getting its own distribution. Read our breakdown of Aippy, the AI game feed that raised at a $250 million valuation, and of the Krea 2 open-weights release for what happens when an image model is given away.

Key Players

• Google Flow: Google's AI video tool, at 823,000 searches and up +809% in a year
• Sora: OpenAI's video model, at 673,000 searches and up +124% over three months
• CapCut: The video editor with AI features, at 1,000,000 searches and flat over the year
• Adobe: The incumbent creative suite, flat at 1,220,000 searches

5. AI Infrastructure, Chips and Physical AI

Behind every chatbot and agent is a server, and this month the search interest moved from the hardware to the bill. "data centers" is at 368,000 monthly searches as of September 2026 (+235%), but the breakout is "data center bill" at 110,000, up from almost nothing a year ago. "sovereign ai", the push for countries to own their own compute, has 165,000 searches, about 69x a year ago, and "ai data centers" adds 60,500 (+173%). The buildout stopped being a trade story and became a local one.

Physical AI is the other half. "tesla cybercab" is at 368,000 searches, about 68x its level a year ago and up +307% in three months, while "humanoid robot" and "tesla robot" each hold 74,000. In gaming, "dlss 5" is a breakout at 301,000 searches, up about 37x in the past three months, and "rtx 5080" sits at 110,000 (+22%).

Why Infrastructure Is Booming

The hardware layer is surging because:

• Agents need compute: Always-on agents use far more processing than a single chat
• Power and land: New data centers raise local questions about electricity, water and zoning
• National strategies: Governments want domestic AI capacity, which drives sovereign AI searches
• Enterprise buildouts: Companies are buying their own AI servers instead of renting everything
• AI leaves the screen: Robots, cars and games now ship with AI at their core

What's Gaining Momentum

Tesla Cybercab (368,000 searches, about 68x YoY): Tesla's robotaxi pulls AI into transport, up +307% in three months.

Data Centers (368,000 searches, +235% YoY): The buildout is now a public topic, though growth flattened over the past three months.

DLSS 5 (301,000 searches, a breakout): The next version of Nvidia's AI upscaling for games, up about 37x in three months.

Sovereign AI (165,000 searches, about 69x YoY): National AI infrastructure, still accelerating hard over three months.

Data Center Bill (110,000 searches, a breakout): What the buildout costs households is now its own search, and it is the steepest new entry on this page.

Humanoid Robot (74,000 searches, +49% YoY): General-purpose robots keep steady growth as pilots move into factories.

A surveillance cluster also sits in this table now. "flock camera map" is at 301,000 searches, about 125x its level a year ago though down -33% over three months, and "license plate reader" holds 74,000 (+22%). The brand term "flock camera" pulls 1,500,000 in the wider pool. Two consumer-hardware rows came in alongside them: "dyson toothbrush" at 550,000 searches and "robotic lawn mowers" at 90,500 (+50%). Those are product-launch searches rather than AI-tool searches, and they are worth reading as noise around the category rather than as demand for AI.

The data layer under physical AI is breaking out on its own too. Read our breakdown of the robot data startup XDOF and the 130,703-episode dataset it gave away.

Key Players

• Nvidia: The main supplier of AI chips and the maker of the DGX Spark; its ticker term "nvda" draws 1,000,000 searches, down -33% over the year
• Microsoft: A major data center builder, at 2,740,000 searches (+50%)
• Tesla: Behind both the Cybercab and the humanoid robot searches in this table

6. AI Search, Detection and the Culture Fight

AI is changing how people search and how they judge what they read. "google ai mode" is at 1,500,000 monthly searches as of September 2026, up +646% in a year, and the short brand term "ai mode" pulls 6,120,000, about 11x its level a year ago. "artificial intelligence overview" adds 246,000 (+307%). Search itself is now an AI product.

The other side is detection, and it has stopped growing. "ai detector" is the biggest term here at 5,000,000 searches, flat over the year though up +123% in the past three months. "humanize ai", the tools that rewrite AI text to slip past checkers, is at 1,000,000 and down -18% over the year. "grammarly" is down -33% at 2,740,000. Text detection has peaked as a category. Image detection has not: "ai image detection" is at 110,000 (+232%).

Why Search and Detection Are Shifting

The trust layer is changing because:

• AI answers first: Google now puts AI summaries and an AI mode ahead of the classic results
• Image doubts: Realistic AI images make people check before they share
• Writing tools fatigue: Students and teachers learned the limits of text checkers
• Slop backlash: Low-effort AI content gave people a name to push back with
• News habits: Readers subscribe to curated AI news to keep up with releases

What's Gaining Momentum

Google AI Mode (1,500,000 searches, +646% YoY): Google's conversational search mode is the fastest-growing destination in this group.

Artificial Intelligence Overview (246,000 searches, +307% YoY): People search the summaries that now top Google results, up +124% over three months.

AI Newsletter (246,000 searches, +82% YoY): Curated AI news is accelerating, up +232% in the past three months.

AI Newspaper (246,000 searches, +82% YoY): The same curve under a different name, as AI-written news becomes its own topic.

AI Image Detection (110,000 searches, +232% YoY): Checking whether a picture is real is a growing habit while text checking fades.

AI Slop (90,500 searches, +234% YoY): The cultural term for low-quality AI content keeps spreading, up +123% over three months.

Three stories from this fight are worth reading in full. Tools that strip invisible characters and em dashes out of model output grew once the EU marking rules put "watermark" in everyone's vocabulary: read our breakdown of the AI text cleaner trend for what those tools actually remove, and AI text watermarking for how the keyed word-choice mark works. Automated replies are now the complaint people actually search for, which we covered in the bot comments trend.

Key Players

• Google AI: The parent term for Google's AI features, at 5,000,000 searches (+233%)
• Grammarly: The writing assistant, down -33% at 2,740,000 searches as detection and rewriting cool
• Perplexity: The AI answer engine, at 1,000,000 searches and down -45% in a year
• NotebookLM: Google's research notebook, at 1,220,000 searches (+122%)

If you want to explore more trends, sign up for free and unlock the full experience.

Summary

AI search in 2026 has moved from asking to delegating, and this month it moved again, from the tools to the bill. The category is led by artificial general intelligence (11,100,000 searches, +171% YoY), with google ai mode (1,500,000 searches, +646% YoY) and background removal (1,220,000 searches, +804% YoY) behind it. The steepest curves belong to agents that do a job, led by ai receptionist (301,000 searches, about 68x YoY), and to the politics of compute, led by sovereign ai (165,000 searches, about 69x YoY) and the breakout of data center bill (110,000 searches, a breakout). What left the page this month: the generic chatbot term, the text-checker cluster, and the generic coding label.

The key takeaways for 2026:

• The frontier question went mainstream: AGI, superintelligence and a model that has not shipped are all table trends now
• Agents went to work: Voice agents for calls, bookings and support remain the fastest-growing applied AI category
• Developers run their own: Machine learning, open coding tools and local models all rose sharply, with Codex up +512%
• Editing beats generating: Background removal grew +804% while the plain image generator term fell -45%
• Compute became politics: Sovereign AI and the data center bill are everyday searches, not trade jargon