AI research assistant team plans: pricing, shared workspaces and admin controls

AI research assistant team plans differ less in how well they answer than in three things: what the team plan costs five people each month, which plan unlocks shared workspaces, and which plan adds roles and admin controls. Most team plans charge per seat, so a five person team pays for five seats every month, including the weeks when two of them barely search. As of September 2026, Notion Business with Notion Agent and Enterprise Search costs a five person team US$100 a month billed yearly, according to Notion's pricing page.

Small teams have a cheaper option. Put the shared part (the chat, research notes, links and files) in a free Barside Space, and let each person connect their own AI account, billed per use under a spending cap. This page works through the numbers for five users, shows what each kind of setup loses on a free tier, and says plainly where a paid team plan is the better choice. We make Barside, so we name its limits as clearly as its strengths.

How much does an AI research assistant team plan with shared workspaces and admin controls cost?

Pricing for an AI research assistant team plan with shared workspaces and admin controls is usually per seat, per month, with a discount for yearly billing, so for five users you multiply the seat price by five, then by twelve. The table below uses only prices we could confirm on an official page, as of September 2026, and leaves a plan's price out when we couldn't.

The last row has no fixed number because API use is billed per question, by the AI company each person connects. You can fix the ceiling, though. If each of the five sets a monthly cap of US$10 in Settings → AI, the most the team can spend through Barside is US$50 in a month, and a quiet month costs far less. The cap is your choice, so pick one that matches how much your team searches.

Per-seat plans have clear advantages: one invoice, one admin, and AI included without anyone touching an API key. Pay-per-use suits teams where research comes in bursts, such as a startup that reads hard before a launch and then builds for six weeks. If you want to add a team plan to the rest of your tools, the subscription cost calculator multiplies any plan by the number of people.

Which plans include shared workspaces, roles and admin controls?

When you compare AI research assistants on shared workspaces, admin controls and pricing, the useful question is which tier first gives you all three together. Shared workspaces let research outlive the person who ran it, and roles decide who can invite, post and delete. Admin controls cover the rest: who can join, who sees the member list, and what happens to the data.

In Barside, the owner opens Manage Space for the controls. They include Who can upload files?, Who can share resources?, Who can pin messages?, Who can post? for a read-only announcements Space, and Slow mode. When members may invite others, Invitations from members need approval is on by default. Owners and moderators also see who joined and posted in the last 7 and 30 days, and can turn off Show the member count and Show the members list in Space settings.

If your company needs identity management across hundreds of people, an enterprise plan from a dedicated vendor is the right fit. Barside's controls are built for a team of five to a few dozen people who know each other.

What you lose on each free tier

Free tiers are where team research tends to break down. A plan that works well for one person often limits sharing, history or AI as soon as a second member joins, so check the team rules before you invite anyone.

• Notion Free is generous for one person, but its pricing page says it limits blocks once two or more members share a workspace, caps uploads at 5 MB and keeps 7 days of page history. Notion AI is a trial.
• ChatGPT has a free tier, and its pricing page lists what the paid plans add for sharing and administration.
• Barside has no paid tier for Spaces, so nothing is held back for teams. What the free app leaves out is the AI itself: each person brings an AI account, and without one, AI research and AI search don't run. Chat, shared notes, files and the search bar work without any AI.

So the loss depends on the tool. On a per-seat plan, the free tier usually loses shared AI. In Barside you keep every shared feature and pay only when someone asks an AI a question.

Setting up a shared Space with each person's own AI account

This setup splits team research into two parts. The AI work happens on each person's computer, with the AI they connect, and the results go into a shared Space where the whole team can read them. Barside runs on Windows, Linux and Mac.

1. Install Barside on each computer. Everyone downloads Barside. Notes stay on each computer and need no account.
2. Connect an AI. Each person opens Settings → AI and connects Claude, ChatGPT or Gemini with an API key, or OpenRouter with a one-click sign-in.
3. Set a cap. Under AI spending limit, turn on Limit what Barside's AI spends and pick an amount for a day, a week or a month.
4. Create the Space. The team lead opens Spaces, signs in with Google or an email code, chooses Create Space, adds a name and a description, and picks a color.
5. Invite the team. Invite by email address from the Members tab, or use Share link so people choose Ask to join and wait for approval in Join requests. Invitations expire after 7 days.
6. Set the rules. In Manage Space, decide who can upload files, share resources and pin messages.
7. Research and share. Press Ctrl+Shift+A to open the AI assistant, ask your question, and let it write what it finds into a note with sources. Share that note into the Space's Notes tab, and post the key finding in Chat.

Web research needs an AI company that offers search. Claude uses Anthropic's web search, ChatGPT uses an OpenAI search model, and OpenRouter uses its web plugin. Gemini has no web search in Barside. Searches are billed on each person's own account.

Take Maya, who runs a five person startup preparing a pricing change. Her product lead asks the assistant to research how three competitors bill teams, gets a note with sources, and shares it into the team's Space. Her engineer saves a copy with Save to my workspace, adds his own findings, and later asks his notes a question with AI search that cites its sources. Nobody pays for a seat, and the research week costs only what the five of them asked.

What are the alternatives to Perplexity Enterprise Pro for shared research threads?

Teams look for alternatives to Perplexity Enterprise Pro and to Perplexity's shared threads for team research for two common reasons: the per-seat cost for a small team, and wanting research to sit next to the notes and decisions it feeds. If your team relies on Perplexity's answers and is happy with the price, staying is reasonable. Check its current plans and sharing features on Perplexity's own site, since we don't quote them here.

The main alternatives for team research with shared threads fall into three groups:

• An all-in-one workspace with AI included, such as Notion Business, where research, docs and AI live in one shared online workspace and everyone pays per member.
• A source-focused tool such as Google NotebookLM, built around adding sources and asking questions about them. It suits a team questioning one stack of papers or reports.
• A shared Space with bring-your-own AI, such as Barside, where each person researches with the AI they choose and the team keeps the results as shared notes, links and files.

Barside doesn't copy the thread model. Each Space has one chat, and a reply quotes the earlier message, so you make a Space per team or topic. Research lands in shared notes, which keep their formatting, images and sources. Shared notes are copies: there's no live co-editing, so one person owns each note and others save their own versions.

Which AI research tools suit startup teams in 2026?

The best AI research tools for startup teams doing collaborative search in 2026 depend on how your team works more than on any ranking. Some likely matches:

• A team that already writes everything in one online workspace will get the most from that workspace's AI tier, such as Notion Business, and should budget per member.
• A team that reads a fixed set of sources closely, such as a founder preparing due diligence, may prefer a source-focused tool like NotebookLM.
• A small team that researches in bursts and works at desktop computers can run the Barside setup above, keep spend tied to use, and add team chat and paid communities in Spaces for the conversation around the research.
• People who live in Claude Desktop or ChatGPT can keep doing so: AI connections let those apps search and read the Barside workspaces you choose, read-only by default.

Mixing is normal. One person may keep a paid chat subscription for daily use while the team's shared record lives in a free Space.

Limits to know before you choose

• No phone app. Barside is a desktop app for Windows, Linux and Mac, so it suits teams who research at a computer.
• No threads. One chat per Space, with quoted replies.
• No live co-editing. Shared notes are copies; each person saves their own with Save to my workspace.
• File limits. 20 MB a file (100 MB for a video) and 250 MB of uploads per account; see file sharing in Spaces.
• Each person manages a key. Bring-your-own AI means five accounts and five bills. People who already pay for a Claude or ChatGPT plan can use it instead, through My plan on their own computer.
• Spending limits count per computer. Each computer counts its own AI use, so set a limit on each AI account too.
• AI search reads six passages per question, at most two from any one note, so broad questions like "summarize everything we found this quarter" work poorly.

Where your research is stored and who can see it

In per-seat cloud tools, your workspace lives on that company's servers. In Barside, each person's notes stay in Barside's data folder on their own computer. When someone asks the assistant or AI search a question, the request and matching excerpts go to the AI company they connected, whose data policy applies, and Barside doesn't train any model on your notes. Read the AI you connect and your data.

Spaces work differently because they're shared online. Messages, shared notes and files travel over HTTPS and are stored on Barside's Spaces server, where membership checks decide who can read them. Spaces aren't end-to-end encrypted, and hosting administrators can access the data, so keep client secrets and credentials out of a Space. Learn how invitations work in share a Space with a link.

Works well with

Barside is free on Windows, Linux and Mac. For a team that wants quick answers in its Space, the Space helper uses the owner's own AI to answer members, and the Slack alternative guide compares Spaces with a full workplace chat tool.