What the Top 100 AI Apps Tell You About Your Own Team

The most useful AI research for your business this autumn is, oddly enough, a ranking of consumer apps.

The seventh edition of the Top 100 Gen AI Consumer Apps came out this week. ChatGPT is still miles ahead, Claude has become a clear third, there's a wave of personal agents, and a song generator sits at number seven on consumer spend.

So why should anyone running a team, a marketing function or an HR department care about any of that?

Because every one of those users has a job. The habits in that ranking walk through your front door at nine o’clock every morning.

89% of office workers who use AI for work first came across it in their personal lives. They taught themselves on their phones, on their own accounts, and carried the habit into the office.

I put AI into large organisations every week, and before we’ve run a minute of training the habits are already there. Nobody designed them and nobody signed them off, so the organisation ends up catching up with its own staff.

So I’ve read the consumer data with a manager’s head on. What does it tell you about how your people and your customers use AI, and what should you do about it on Monday?

How Big the Home Habit Has Got

About half of American adults now use AI chatbots, up from a third in 2024. Roughly one in four use them every day.

The UK is moving at the same speed. ChatGPT had 1.8 billion UK visits in the first eight months of 2025, against 368 million in the same months of 2024. Almost five times as many in a year.

Worldwide, AI now has about two billion users. That number grew 11% this year. Consumer spending on AI tripled to an estimated $40 billion, which tells you the same people are going a lot deeper.

And look at what they’re doing with it. Among people who do each of these activities, the majority now use AI for six of them: writing (61%), coding (58%), help with school and work assignments (56%), image creation (53%), presentations (52%) and building games (52%). A year earlier, only writing had crossed half.

That list is pretty much a description of an office job.

Among employed American adults, 38% already use chatbots for tasks at work. They’re also juggling tools: the average AI user now uses three general assistants, up from 2.2 a year ago.

Now look at who pays for the big three assistants.

A quarter of U.S. consumers use AI daily. Fewer than one in 20 pay for a personal plan with one of the big three assistants.

A quarter of people using AI daily, and 4.5% paying for a personal plan with ChatGPT, Gemini or Claude. That figure has doubled in a year, from 2.1%, and it’s still small. Plenty of daily users are on free plans.

Why does that matter to you? Because a free plan is a personal account, and when those people open a browser at work, their AI comes along with everything they’ve saved in it and whatever they paste in next.

The Money and the Skill Sit With a Small Group

The spending data is where this edition gets really interesting for employers.

The top 1% of consumer AI payers account for more spend than the bottom 50% put together.

The top 10% of payers account for roughly half of all consumer AI spend. The top 1% alone account for 19.5%, more than the entire bottom half (16.6%). That top 1% spends an average of $903 a month on AI, up 80% in 18 months. The median payer spends $25.

What are the big spenders buying? Mostly tools that automate work or make things, like workflow automation, design, video and agents. And 29 of the top 50 companies by consumer spend are missing from the web and mobile traffic rankings altogether. A lot of the money goes to tools that never show up on a traffic chart.

Time follows the same curve: one in ten AI users spends more than two hours a day with it, while almost three in ten spend less than 15 minutes.

So what? Your organisation almost certainly has the same shape. I see it in nearly every team we train.

Workplace data shows the same shape. Across 12,004 employees in 40 countries, 19% reported no time saved with AI at all. And 73% of the most productive AI users were managers or executives.

So the skill sits with a small group near the top of the org chart, much like consumer spend sits with the top 1%.

If you only count licences and logins, that top group hides everyone else. Your dashboard says adoption looks healthy, when on the ground a handful of people are flying and the rest have barely started.

Your Best People Are Building Their Own Kit

One chart in this edition says a lot about where work is heading.

Claude, positioned as the work assistant, has far more of its consumer payers on a $100-a-month plan.

The consumer market has started to split by what people use it for, with Claude pitched at work, Gemini at everyday consumers and ChatGPT trying to do both. 7.3% of Claude’s consumer payers are on its $100-plus Max plan, against 1.3% for Google and 1.1% for ChatGPT on their equivalent tiers.

Who are these people? Tech and finance professionals, often with an advanced degree, who treat AI as equipment. 82% of Claude users pay for AI, and they’re twice as likely as the average user to spend $100 or more a month.

These are consumer plans, on personal cards. So how many of those $100-a-month subscriptions do you think are being used for the day job?

Who pays tells the same story. Among consumers who pay for AI, 48% cover it entirely themselves. Only about one in five has an employer or school picking up the bill. When the employer does pay, people spend more: $95 a month on average, against $55 for those paying out of their own pocket.

Companies are buying, to be fair. 56% of the 70,000-plus American businesses on Ramp’s spend platform paid for an AI product in August, up from 45% a year earlier. And OpenAI launched ChatGPT Work this summer, chasing exactly this crossover.

The trouble is the habit got there first. Your most capable people have already picked their tools and built how they work around them, often on their own card. Whatever you roll out has to be at least as good, or it ends up as the thing they use when someone’s watching.

Shadow AI Is the Default Setting

Put the consumer data next to the workplace surveys and you can see what’s going on.

Home AI habits come to work with people, along with the data they paste in.

88% of employees who have company AI access also use personal AI tools for work. Two-thirds of office workers have used AI at work they believed wasn’t allowed. 88% have shared work information with a public AI tool: emails (43%), meeting notes (40%), customer data (34%), and financial information or confidential documents (31%).

Why would people take that risk? Because it works for them, and they can feel it working.

People who mix company and personal tools are 1.7 times more likely to report significant time saved than people who stick to the company kit. They’ve got three assistants at home and know which one suits which job, and at work you’ve probably given them one.

And they think they know better. 72% of office workers believe they understand how to use AI for their job better than the team that manages AI at their company. 81% think leadership plays by different AI rules from everyone else.

So what happens when you crack down? Of the people who’d used tools they thought were off-limits, 48% faced formal consequences, a warning or worse, and the behaviour carried on somewhere you can’t see it.

Ban personal AI and you’re banning the people saving the most time. Ignore it and your customer data ends up in someone’s personal chat history. What works is bringing the habit out into the open, onto tools you control, and teaching people to use them properly.

Agents Are Already in Their Inboxes

The newest category in this year’s ranking is personal agents, meaning AI that goes off and does things for you, like booking a table or chasing an invoice.

They’re spreading fast. One new agent that works inside iMessage reportedly passed 100,000 users in three weeks, while still invite-only. Meta’s Muse crossed five million downloads in under a month. And in early user groups, the most discussed task for agents was coding and technical automation.

41% of AI users have tried an agent. 24% use one regularly. More worrying, 32% have already let AI act on their behalf without a final check.

To make that work, people hand over the keys. Consumers have given agents access to their email (36%), web browser (33%), messaging apps (31%), cloud storage (29%) and calendar (27%).

Now picture someone who’s happily let an agent run their personal inbox for six months. They start a new project at work. What’s the first thing they’ll try?

Look at your own AI policy. It probably covers what people can type into a chatbot. Does it say anything about what an agent can open, send, book or delete on someone’s behalf? Your people are already well past that point at home.

If your policy doesn’t mention agents yet, it’s already a year behind your staff.

What Your Customers Are Telling You

Everything above is also true of your customers. They’re the same people, on the same apps. And what they expect is shifting in ways your marketing and service teams need to hear.

Trust matters more to them than convenience now. AI users rank accuracy (45%), trustworthiness (40%) and security and privacy (36%) ahead of ease of use, which fell from 38% to 32% in a year.

They’re also getting wary of AI-made content. 36% say they’re less likely to engage with something if they know it was AI-generated. Only 14% say it makes them more interested. Among people who don’t use AI at all, 83% would prefer to deal with a person.

At the same time, AI is already shaping what they buy. Of shoppers who use AI, half use it to compare prices, 43% to narrow a shortlist, 37% made a final purchase decision based on its recommendation and 26% let it complete a purchase.

Why Training Is the Lever

Add it all up and you’ve got self-taught habits, a few experts, a lot of people who’ve barely started, data drifting into personal accounts and agents on the way. Where do you start?

With the organisation itself. In a 20,000-person study this year, organisational factors, things like culture, manager support and how people are rewarded, accounted for 67% of the AI impact people report. Individual mindset accounts for 32%.

Right now, in most places, the organisation is getting in the way. Only 13% of AI users say they’d be rewarded for reinventing their work with AI if results slipped. Only 26% say their leadership is clearly aligned on AI.

The teams getting furthest with AI do their learning together, out loud.

Look at what the leading teams do differently. They’re about twice as likely to brainstorm AI ideas together (63% against 32%), share what they’ve learned (61% against 36%) and talk about what good AI-assisted work looks like (54% against 29%), and their managers use AI where everyone can see it.

That is what good training looks like: people learning from each other on real work, with someone senior making it feel normal.

And your people are asking for it. 75% of office workers say they’d be likely to look for a new job that offered better AI skills development. At billion-dollar companies it’s 80%. 77% say their company’s AI restrictions hold back their career.

Here’s the uncomfortable bit. Your best AI people taught themselves, on their own money. If you don’t give them somewhere to grow, they’ll take those skills to someone who will. And the dabblers and holdouts won’t catch up on their own, because nobody’s shown them what good looks like for their job.

Training is how you get the home habit onto safe tools, spread the skill well beyond the top few, and hang on to the people who already have it.

Six Moves for Monday Morning

If you lead a team, this is the useful bit.

Start with the survey, because it tells you where the other five need to land first. After that it takes someone senior to admit what’s already happening and decide to make it work for the business.

Your people have already run the pilot on their own phones and their own money, and the results are sitting in their browser history.

Do you know which AI tools your team paid for themselves this month? And what they put into them?

Sources & References

Frequently Asked Questions

Roy Murphy is co-founder of G3NR8. This post draws on the seventh edition of the Top 100 Gen AI Consumer Apps and 2025 to 2026 research on how people use AI at home and at work. Published 9 October 2026.