Hops Is Live: Wrike Founder's Team AI Workspace Has No Per-Seat Fees

Most team AI tools charge by the seat and still leave the work in private chats. One person asks ChatGPT or Claude a question, gets a good answer, then copies it into a shared doc while the rest of the team repeats the same question tomorrow. Hops, a new collaborative AI workspace from Wrike founder Andrew Filev, is built on the opposite idea: the AI works inside the files the whole team already shares, and the team pays only for the AI work it uses, with no per-seat fee at all. It opened to everyone on November 10, 2026, after more than 2,500 teams used it in early access, according to the company's launch announcement. The pitch is simple and easy to test. Whether the credit meter behind it stays cheap in real use is the question this review answers, using Hops' own pricing page and launch release as the sources.

What Hops actually is

Hops is a workspace where a team and an AI agent work in the same chats, documents, slides, tables and dashboards. You ask @Hops in the thread, and it drafts from the files already in that space, writing straight into the shared document instead of handing you text to paste somewhere else. Everyone edits the same version, and every edit Hops makes shows in version history, so the team can tell what came from the AI and what came from a person.

The product comes from For Good AI, the company Filev leads after founding Wrike, which was acquired for $2.25 billion, and building Zencoder, a set of AI coding agents for developers. Filev's argument for Hops is that individual AI gains have not turned into company gains. In the launch release he says that when he asks founders and executives how much faster their company moves than a year ago, the answer is usually 10 to 20 percent, well short of double. His comparison: "The PC era didn't pay off when PCs got faster. It paid off when they were networked. I believe AI will follow the same path." That 10 to 20 percent figure comes from his own conversations, not a published study, so treat it as a founder's framing, not a measured result.

Hops also connects to more than 2,000 apps, from CRMs to cloud drives, so the agent works from the team's own data rather than from a blank chat. And when a team is missing a tool, it can describe the app it needs and Hops builds and runs it inside the workspace, using the same data and permissions.

The pricing, in plain numbers

Hops sells AI work by the credit, not by the seat. The pricing page, read on November 12, 2026, lists three plans:

The headline number is the top-up rate: $20 buys 20,000 credits, which works out to $0.001 per credit, or one tenth of a cent. The minimum top-up is $20. Credits bought this way stay valid for 12 months, cannot be refunded except where the law requires it, and sit in a shared pool the whole workspace draws from, with admins able to cap each person's use. Prices are in US dollars and exclude taxes, so UK and Canadian readers should add local tax at checkout. At November 12, 2026 market rates, $20 is about £15 or CA$29, and the $200 Team AI plan is about £151 or CA$285. Those conversions are approximations.

One detail changed between the launch coverage and the live page. Unite.AI's November 10 report says the free plan includes 5,000 credits. The pricing page now lists 20,000 free credits. This article uses the live page's figure, since it is the number a team signing up today will actually see.

There is one gap in the pricing page worth naming. Team AI costs $200 a month and mentions a shared credit pool, but the page does not say how many credits, if any, that $200 includes. Until Hops spells that out, treat Team AI as a $200 fee for team features, with credits possibly billed on top, and check before you commit.

What a credit buys

Hops publishes example credit costs for common tasks. The ranges below come from its own pricing page, which labels them illustrative, and the real cost of a task depends on the model used, how much information Hops reads, which tools it calls, how many steps it takes and how long the output runs.

Hops does not name the models it uses, saying only that it runs on top frontier models, matched to each job, so the cost of the same task can move without warning if the routing changes. The pricing page also says model availability, routing, credit packages and consumption rates may change prospectively.

The 20,000 free credits stretch further than the number suggests. At the published ranges, they cover roughly 13 to 40 deep research reports, or 66 to 200 quick answers, before a team spends anything. Hops suggests $20 per active user per month as a starting budget, which means a team where three people use the AI daily should expect about $60 a month after the free credits run out. Chat, documents and collaboration never consume credits, and when the pool runs dry Hops pauses AI work while the team keeps using everything else.

Skills: teach once, the whole team uses it

Skills are the feature that separates Hops from a chatbot with a shared login. When one person works out a good way to do a recurring job, they can save the method, instructions, context and format, as a skill, and decide who else can use it. Hops then applies the same method each time anyone on the team asks. The company's own examples are sales skills: call-prep turns account context into a source-linked call brief, follow-up records decisions and next steps, deal-review flags deal risks, and account-handoff carries context and owners between teammates.

The launch release gives a concrete case. Hops' head of marketing taught the system the company's voice, which message leads, what each audience cares about, which examples to use, and now a salesperson can ask for a customer email in the team chat and get one in the company's voice, written straight into the shared doc. The head of marketing spends her time on strategy instead of line edits, according to the release. That is the company's own example, not an independent test, but it shows the mechanism clearly: the value compounds when the team's know-how is saved once and reused, instead of re-explained in every chat.

Permissions and security

Hops' permission model is the part enterprise buyers will read first. Hops will not use a file the person asking cannot open. If an answer would draw on information others in the thread cannot see, Hops holds the reply until the person who asked approves it. External sends and sensitive changes also wait for a human to approve them, and administrators can control which teams use connected tools or make those tools read-only. Any code Hops runs stays sandboxed, away from team data.

The company lists SOC 2 Type II, ISO 27001 and ISO 42001 certifications, and the launch release states that Hops never uses customer data to train models. Those certifications are company-stated; no audit report is published on the site for a reader to check. The full 2,000-plus integration library is an Enterprise feature, although the pricing page says it is included on the free Go plan for a limited time during launch, as a promotional offer that may later change with advance notice.

Hops vs the per-seat tools

The fair comparison is not Hops against a free chatbot. It is Hops against the per-seat team plans most companies already pay for. Anthropic took a similar shared-workspace step this week with Claude Sheets, Whiteboards and Meetings, and OpenAI is pushing team agents of its own, with Businesses in ChatGPT now live. Hops differs from both by charging for AI work instead of seats.

Run the math for a 10-person team and Hops' advantage depends on how unevenly the team uses AI. If all ten people use it daily, Hops' own $20-per-active-user guidance points to about $200 a month, the same as ChatGPT Business on annual billing or Notion Business. If only three people use it heavily, the bill is closer to $60, and the seven light users cost nothing, where ChatGPT Business would still charge $25 each. That is the whole bet: Hops is cheaper when usage is lopsided, and no cheaper at all when everyone is a power user.

The comparison also cuts the other way. ChatGPT Business, Notion and Slack are established products with named models, published limits and years of independent reviews. Hops is two days old as a public product, names no models, and has no independent testing yet. Its homepage carries three early-user quotes, all from founders of small companies, and its launch release carries one customer, Discern Security, whose CEO says Hops replaced Slack and Jira for his sales and marketing teams. Those are company-selected voices, not a track record.

What to watch before you switch

Four things deserve a check before a team moves its work into Hops. First, the model question: Hops names no models and warns that routing and consumption rates can change, so a task that costs $0.50 today could cost more later with no announcement beyond the pricing page. Second, the Team AI gap: the $200 monthly plan does not state what credits, if any, are included. Third, the integration library: the full 2,000-plus library is an Enterprise feature, and its inclusion on the free plan is a launch promotion that can end with notice. Fourth, the free plan's limits: the Go plan carries 5 GB of storage, 100,000 messages and 1,000 docs, with fair-use limits on collaboration features, and unused, idle workspaces may be removed after prior notice, with time to export.

Two more cautions come from the launch itself. The 2,500-team early access figure and the 10 to 20 percent company-speed claim are company-reported, with no independent count behind them. And the free-credit number already moved once, from 5,000 in launch coverage to 20,000 on the live pricing page, which is good news for sign-ups but a reminder that Hops is still tuning its offer in public.

Who should try it

Try Hops first if your team already lives in shared docs and repeats the same jobs: sales follow-ups, meeting notes, pipeline reviews, campaign drafts. The free tier works for a proper trial: 20,000 credits, unlimited teammates, no card needed to see whether the skills idea sticks. Set a per-user cap on day one, since the pool is shared and one heavy user can drain it.

Wait if you need named models, a fixed per-seat budget your finance team can approve in advance, or a tool your IT team has already vetted. Hops' credit meter makes the monthly bill variable by design, and the $200 Team AI plan still leaves the credit question open. Teams that need audit logs, SCIM and a DPA will be talking to Enterprise sales, where the price is custom and the pitch is a sales conversation, not a price list.

Frequently asked questions

Is Hops really free?

Yes, for collaboration and a first batch of AI work. The Pay as you go plan costs $0 per user per month and includes 20,000 free credits, unlimited teammates, chat, docs, slides and tables. You pay only when you buy more AI credits, starting at $20 per 20,000 credits.

How much will my team actually spend on Hops credits?

It depends on who uses the AI. Hops suggests $20 per active user per month as a starting budget. A deep research report costs roughly $0.50 to $1.50 at the published credit ranges, while a quick answer costs about $0.10 to $0.30.

What AI models does Hops use?

Hops does not say. Its pricing page states only that it runs on top frontier models, matched to each job. The company also warns that model availability and routing can change, so the model behind a task may change without notice.

Does Hops train on my team's data?

No, according to the company. The launch release states that Hops never uses customer data to train models, and the site lists SOC 2 Type II, ISO 27001 and ISO 42001 certifications. No audit report is published on the site to check independently.

Can Hops replace Slack or Notion?

It can replace some of what they do, not all of it. Hops covers chat, docs, slides, tables and task tracking in one workspace, and one launch customer says it replaced Slack and Jira for his teams. Teams that rely on Slack's app ecosystem or Notion's databases should test before switching.

What happens when the credit pool runs out?

Hops pauses AI work until someone adds credits. The team can keep using chat, docs and everything else as usual. Credits bought as top-ups stay valid for 12 months, and admins can cap how many credits each person uses.

Bottom line

Hops is a real product with a clear idea: AI should work where the team works, and teams should pay for AI work, not for seats. The pricing is transparent about the rates and honest about the variables, and the free tier is generous enough for a proper trial. The open questions are the ones that matter for a buying decision: which models run underneath, what Team AI's $200 actually includes, and whether the skills idea survives contact with a real team. Those answers will come from use, not from a launch release. Sign up free, teach it one recurring job, and watch the credit meter for a month before you decide.

Sources

• Hops, official homepage — https://www.hops.ai/
• Hops, official pricing page (read November 12, 2026) — https://www.hops.ai/pricing
• Hops, About page — https://www.hops.ai/about
• Hops, "AI promised 10x. Most companies got 1.2x. Wrike founder Andrew Filev launches Hops to close the gap," PR Newswire, November 10, 2026 — https://www.prnewswire.com/news-releases/ai-promised-10x-most-companies-got-1-2x-wrike-founder-andrew-filev-launches-hops-to-close-the-gap-302934936.html
• Unite.AI, "Hops Opens Collaborative AI Workspace for Shared Teamwork," November 10, 2026 — https://www.unite.ai/hops-opens-collaborative-ai-workspace-for-shared-teamwork/
• Zemith, "ChatGPT Team vs Claude Team: Which Plan Is Worth It? (2026)" — ChatGPT Business $25/$20 per user per month, verified November 1, 2026 — https://www.zemith.com/en/contents/chatgpt-team-vs-claude-team
• Siit, "Notion vs Slack (2026)" — Notion Business $20 per member per month, Slack Business+ $15 per user per month, verified August 2026 from the vendors' pricing pages — https://www.siit.io/tools/comparison/notion-vs-slack
• Exchange rates: November 12, 2026 market rates, verified at publication (1 USD ≈ 0.76 GBP, ≈ 1.43 CAD); GBP and CAD figures in this article are approximations.