Swiftask

Swiftask is a no-code platform for building AI agents, which its pages pitch as "The operating system of your enterprise AI." It is aimed at business teams and the IT or data-protection people who oversee them. The company's guide calls it a French platform, the footer shows "Hosted in Europe" and "France", and lists a "Scale Up Excellence Winner French Tech Nantes 2026" badge on the home page. The product is offered in French and English. The home page headline reads "Agentic AI & AI Governance for Enterprises." The workspace comes with pinned, ready-to-use agents such as Marc the developer, Julie the business coach, Antoine the marketing expert, Thomas the data analyst, Sophie the web copywriter and Alexandrine for UX and UI, plus a library of pre-built agent templates.

What you build and how

The core object is the agent. The company's guide, published September 10, 2025, walks through 6 steps: initialization, visual configuration, behavioral instructions, skills, data integration, and deployment. You open the Agents section, click "Create an agent", name it (the guide's examples are "Support Assistant" and "Product Expert"), pick the "Blank template" for maximum customization, and confirm. The home page FAQ states the position directly: "Swiftask is a 100% no-code platform. You configure your agents through an intuitive visual interface", with "no coding required." The guide times a basic agent at 2 minutes, an agent with data at 5 minutes and a complex multi-skill agent at 5-10 minutes. The home page says the first agent "is operational in less than 10 minutes."

Skills are the abilities you attach. The guide groups them as communication (automated email, Gmail and Outlook management, LinkedIn publishing), productivity (Excel, PDF and DOCX export, Google Drive, Python automation) and integrations (custom REST APIs, SQL databases, web scraping). The home page adds that agents connect to business tools such as Slack, HubSpot, Notion and Google Drive, and says "3,000+ integrations available."

Knowledge comes from your own material. The guide lists PDF, DOCX, PowerPoint, Excel, CSV, JSON and Markdown files, YouTube videos, web pages, SQL databases and REST APIs, with connected sources including Google Drive, Dropbox, OneDrive, SharePoint and internal systems through an API. The home page labels this "Knowledge Powered by RAG", which means answers are retrieved from your documents, and says answers come back with "citations and sources."

On the model side, the site says you get "80+ LLMs in one place, without being locked into a single provider." The guide's model list names GPT-4o, o3-mini and GPT-4.1 from OpenAI, Claude 4 Sonnet, Claude 3.7 Sonnet and Claude 3.5 Haiku from Anthropic, Gemini 2.5 Pro and Gemini Flash from Google, Mistral Large, Medium and Codestral, and Grok 3 from xAI. The home page's AI chat hub shows OpenAI, Claude, Gemini and Mistral answering side by side, and the FAQ calls Swiftask multi-LLM and AI-agnostic, so you choose the model for each task.

Publishing for a chatbot is a widget. The guide's final step turns on long-term memory for conversation continuity, lets you customize the widget's colors, position and welcome messages, then has you "Copy the integration code and add it to your site", a single script tag. A "widget inbox" lets you supervise conversations in real time, and the analytics section lists message logs, response time by query type and automatic resolution rate. The guide also mentions A/B testing of configurations and comparing agent versions.

Beyond chat widgets, the home page lists what agents can do after a conversation: transfer to another agent in a multi-agent setup, update a CRM when a lead qualifies, notify Slack or Teams, create a task in an ERP or CRM, and produce slides, reports, meeting summaries and action points. The FAQ says the platform combines custom agents with built-in tools for every employee: a meeting assistant, slide generation and multi-model chat. The footer offers downloads for Mac, Windows, Linux, a Chrome extension, Android and iPhone.

The evidence

The guide's pricing answer is the only price on the pages. It describes a freemium approach with a free plan for "Complete platform testing", a Starter plan at $6 a month for 1 user and standard models, and Professional at a custom quote for the complete team and all premium models. The guide closes by inviting readers to create a free account.

Governance is the home page's strongest section. It lists cost and credit control with AI budgets per team or department, tracked in real time with alerts before limits are exceeded; control over which models each person can reach; sharing of agents, prompts and workflows across teams; audit logs of every AI interaction showing who used what, when, with which model and at what cost; and granular roles and permissions with "Native SSO/SAML integration." It also says "We guarantee that LLMs will never train on your private data" and that the platform is hosted in Europe with strict GDPR compliance.

The ISO 27001 wording differs between sections: the data privacy block says "ISO 27001 certified (in progress)", while the FAQ says the platform is "ISO 27001 certified." The guide adds French hosting with encrypted data, and the footer links to a data processing agreement and a subprocessors page.

The guide claims "Average ROI observed: 300-500% in 6 months" and cost reductions of 30 to 60%. These are the company's own figures with no study attached. It also describes a 3-phase support program that ends in "Assisted Creation", where the team helps develop your first agents, trains your team and supports production deployment.

The home page's customer stories give concrete builds. Next Decision, a data and AI consultancy, qualifies every incoming lead automatically, scores opportunities from 1 to 5 stars and updates Zoho CRM without human intervention. Frennly, a digital agency, put a technical conversational agent on a client's website that answers complex questions from the CTICM document database. Millésima cut translation of multilingual product sheets from 20 min to less than 1 min and builds a newsletter through a multi-agent setup of manager, writer and layout agents. At Hâpy Services, 10 human employees are supported by 12 specialized AI agents, and Data Inceptio's slide agent, Agent Picasso, generates 16 slides in under 5 minutes following brand guidelines.

Who it suits and who it does not

Swiftask suits a European company that wants many teams building agents under one contract, with budgets per department, audit trails, SSO and a GDPR and European-hosting story. Support, sales, HR and finance teams can build for themselves, and the home page says they can do it "without relying on IT or a developer." A department that wants a website assistant fed by its own documents has a short path: create, connect sources, paste a script tag.

It is less natural for a solo founder who wants a visual flow canvas with branching nodes. The pages describe configuration, instructions and skills rather than a node editor. The Starter plan's 1 user and the quote-only Professional tier mean a small team has to talk to sales. Compliance-minded buyers will want the ISO 27001 status settled, since the home page uses both "certified" and "in progress".