amaiko vs Dust: Built Agents vs a Memory That Learns (2026)

Dust needs an operator. amaiko needs a chat.

amaiko arrives with its specialists already working: email, calendar, research, documents and the company systems you connect. You install it in Teams, say hello, and it starts. Nobody writes instructions, picks tools or wires up data sources first.

Dust’s pricing page calls Dust “the platform for AI Operators” and describes them: “They build it with judgment, deploy it, and run it for their whole team.” Everything an agent does is set up in Dust’s Agent Builder. Memory is a tool you add there, and so are schedules.

For your company that means a named person who designs, tests and maintains every agent, next to their actual job. With amaiko that role does not exist.

Every Dust message draws down a credit allowance

amaiko bills no credits. The plan includes unlimited tokens, so a long research task costs the same as a one-line question.

Dust meters every interaction in credits: tokens plus a fixed charge per tool action, three credits for each knowledge search or third-party tool call. A Pro seat includes 8,000 credits a month, and unused credits do not roll over. Dust’s seat documentation says it plainly: “When members reach their credit limit, they are blocked from sending messages.” On the Business plan there is no pay-as-you-go; the credit management page says admins buy workspace top-ups in advance. Automated usage, including triggered agents, draws only from that workspace pool, with “no free credit baseline”.

For your team that means someone watches a credit dashboard, and your heaviest users hit a wall mid-month, draw on top-ups you paid for in advance, or move to a seat that costs five times as much.

What amaiko learns belongs to the company

amaiko builds a persistent corporate memory. What your team works on, decides and knows is consolidated overnight into company knowledge, and it stays when someone leaves. amaiko teams spend 35% less time searching, and onboarding runs up to 57% faster.

Dust’s memory is personal. Its documentation says: “Memories are private to each user and only you can view and manage your memories with each agent.”

For your company that means knowledge in Dust leaves with the person who taught it. In amaiko it stays in the company.

In Teams: a colleague, not a bot you mention

amaiko runs in Microsoft Teams, in one chat, and reads across the Microsoft 365 workday: email, calendar, files, Word, Excel and PowerPoint documents, Teams chats, OneNote, To Do and SharePoint lists. The morning briefing is there before you ask.

Dust’s Teams integration is a bot. Your admin enables it in Dust, downloads the app and uploads it to Teams as a custom app. Agents “only have access to conversations where they are explicitly mentioned or linked”, and you call them as @dust +agentname.

For your team that means agents you have to summon by name. amaiko is already in the chat you work in.

Hosting: a country, not a region

amaiko is hosted in Germany on its own infrastructure, GDPR-compliant, with AI processing inside the EU. Details are on the security page.

Dust’s security page offers “Host in the EU or US”. It names a region, not a country.

For your data protection officer that means one more question for Dust: in which country the EU workspace is stored.

Pricing: a flat price against a credit meter

amaiko costs €41.58 per user per month billed annually (€499 per user per year), €49.90 billed monthly, for at least five seats. See pricing.

Dust’s Business plan is for teams of up to 100 people and charges per seat, all prices excluding VAT, on its pricing page:

• Pro: €24 a month billed yearly, €30 monthly, 8,000 credits per seat per month.
• Max: €120 a month billed yearly, €150 monthly, 40,000 credits per seat per month. Dust recommends it for “power users running complex automations, Deep research, or tool-heavy workflows regularly”.

Per seat, Dust Pro is cheaper than amaiko. That price holds only while nobody on the seat goes past 8,000 credits a month, and it does not include the workspace top-ups that scheduled and triggered agents run on. amaiko has no credits and unlimited tokens: the seat price is the whole bill.

Questions to ask Dust before you sign

Get the answers in writing. Ours is under each one.

1. Who in our company builds, tests and maintains the agents, and how many hours a month does that take? amaiko: Nobody. The specialists come ready and work from the first chat.
2. How many credits does a typical user in a company like ours use per month, and what happens on the day someone hits 8,000? amaiko: There are no credits. Unlimited tokens, one price per user.
3. What will our scheduled and triggered agents cost, given that they draw only on prepaid workspace credits? amaiko: The morning briefing and inbox triage are in the plan price. Meeting notes are the Meeting Notetaker add-on, €19.99 per user per month on top of the plan: €61.57 all-in, billed annually.
4. When an employee leaves, what happens to what the agents learned from them? amaiko: It stays in the company’s knowledge.
5. In which country is our EU workspace stored, and where do the models process our data? amaiko: Hosted in Germany on amaiko’s own infrastructure, AI processing inside the EU.
6. How many data-source connectors does the Business plan include, and what does the next one cost? amaiko: Microsoft 365 is one connector, included in the plan: email, calendar, files, Teams chats, OneNote, To Do and SharePoint.

See it for yourself

Book a demo and ask amaiko the questions your team asks every day, in your own Teams. If you are still comparing, our Dust alternatives list covers the rest of the field.