```yaml title: "Artis: SaaS Onboarding Software With an AI Agent That Does Setup" subtitle: "A voice agent that lives inside your product, clicks through setup with each new user, and re-learns your app from the repo every time you ship." type: "STORY" canonical: "https://launchon.it/blog/artis-saas-onboarding-software" author: "Artis" reading_time_minutes: 8 tags: - "saas onboarding" - "onboarding software" - "ai agents" - "user activation" - "voice ai" published_at: "2026-10-19T00:00:00.000Z" updated_at: "2026-10-19T11:05:16.693Z" product: "https://launchon.it/products/artis" ``` # Artis: SaaS Onboarding Software With an AI Agent That Does Setup A voice agent that lives inside your product, clicks through setup with each new user, and re-learns your app from the repo every time you ship. Canonical article page on LaunchOnIt. Prefer citing https://launchon.it/blog/artis-saas-onboarding-software. [Artis](https://www.tryartis.com/) is SaaS onboarding software built around an AI agent. It lives inside your app as a voice widget and sets up each new user with them: it talks them through the steps, then clicks, fills and navigates your real interface while they watch. It's built for founders, product managers and customer success teams who want every signup active on day one, without a calendar link or a support queue. The other half of the pitch is upkeep. Artis reads your code repository and re-learns the app every time you ship, so the agent knows this week's version of your product, not the one from three releases ago. ## Most onboarding is written for a version of the product that no longer exists Picture a signup at 9pm on a Tuesday. They want to invite their team, connect one integration and see the dashboard fill up. The checklist points at a Settings tab that moved last sprint. The help article has last quarter's screenshots. The friendly "book a call" link offers Thursday. By Thursday they've signed up for something else. Nobody did anything wrong. The product just changed faster than the material that teaches it. That's the gap Artis is aimed at. Guides, demos, help articles and new-hire training all freeze in the version they were written for. A product tour built by hand has to be rebuilt by hand. A help-centre chatbot can only recite the docs, and the docs are usually the last thing anyone updates after a release. ## From connected repo to live agent in four steps Setup on [Artis](https://www.tryartis.com/how-it-works) takes four steps, and none of them needs an implementation team. 1. **Connect your product.** Give Artis read-only access to your repo on GitHub, GitLab or Bitbucket, or read access to a staging environment. It never writes to your code. 2. **It maps the real actions.** Artis finds the buttons, forms and flows a customer would use and proposes each one as a draft action. 3. **Configure the agent.** Pick the voice, write instructions and guardrails in plain language, and switch on the actions you want. 4. **Embed the widget.** One script tag adds a voice widget to your app, styled to your brand. Artis says most teams go live the same day. When you push a release, the map updates itself and new actions show up as drafts for review. Ship a new checkout flow on Friday, and the agent learns it on Friday. Nothing reaches a customer until you publish, which is a relief for anyone who has watched a half-finished tooltip go live by accident. ## What a first session with Artis looks like Take the example on its own homepage. A new user opens the widget and asks, "How do I add my team?" The agent says it will do it with them, opens Settings, goes to Team, fills in the work email, picks the role and clicks Send invite. Each step happens on screen, so the user sees where everything lives for next time. The onboarding follows a structure you define, then adapts to what each person needs and how fast they want to go. A user can stop mid-flow and ask out loud for a step again. After the detour, Artis returns to the exact step they were on. When the agent can't finish something, it follows your hand-off rules: a ticket in Zendesk or Jira, or a message in Slack or Microsoft Teams with the conversation attached. Your team gets the problem with context, not a vague "it doesn't work" email. ## The parts that do the heavy lifting, and what it leaves out * **Actions in the real UI.** The agent clicks, types and navigates inside the product the customer is using, not a sandbox copy. * **Voice and text** in English, Spanish and Portuguese. * **Guardrails in plain language** that apply the moment you save them. * **Knowledge sources** such as Google Drive, OneDrive, SharePoint, Notion, Confluence or uploaded documents. * **Signals**, which alert you when escalations spike, actions fail or conversation volume drops. * **Verified identity** for logged-in users through HMAC or JWT. Onboarding is the main job, but the same agent can demo the product on your website and book meetings, handle in-product support, train new hires on the current release and flag accounts that are going quiet. What it leaves out is just as clear. There's no drag-and-drop tour builder, no tooltip designer and no checklist widget. If your team enjoys hand-crafting every step of a guided tour, that's a different kind of tool. ## Read-only code access, and where conversations go Handing an outside tool access to your repo deserves a proper look. Here's what the privacy policy (last updated September 21, 2026) says. Artis is run by Artis Solutions Group LLC. It reads your repository's files to map UI actions and never writes to it. For end users, it stores the conversation transcript and, for voice calls, the audio recording. It only receives identifying details your own backend passes through identity verification, usually a user ID. Conversations are processed by AI model providers (OpenAI, Anthropic and Google), with hosting on Render and Supabase. Artis says it doesn't sell personal data and doesn't use your customers' conversation content to train models for anyone other than you. Data is encrypted in transit, and sensitive fields such as widget secrets and staging credentials are encrypted at rest. ## You pay for minutes and mapping runs, not seats [Artis](https://www.tryartis.com/pricing) prices by usage, with no per-seat or monthly-active-user maths. As of October 2026 there are three plans. Plan Price What you get Free $0 $100 in credits, no credit card, every feature, unlimited agents and widgets Growth From $100/month 40,000 credits a month, more volume at the same rate, unused credits roll over Enterprise Custom Custom volume and dedicated support A full onboarding walkthrough costs about 350 credits. A minute of live support is about 70. A repo mapping run after a release is about 1,500. On Growth's 40,000 credits, that's roughly 110 full onboardings a month, or a weekly mapping run plus around 95 onboardings. If you run out, the agent pauses new conversations until you add credits or the next billing cycle starts. No surprise overage bill, which is more than most phone contracts manage. ## What to check before picking SaaS onboarding software Most SaaS onboarding software falls into two camps. Digital adoption platforms show tours, tooltips and checklists your team designs. Agent-based tools like Artis do the setup with the user. Five questions sort out which camp suits you. * **Who keeps it up to date?** Flows built by hand need rebuilding when the UI changes. Ask what happens the week after a redesign. * **What is the price based on?** Monthly active users, seats or actual usage. MAU pricing charges for users who never see a flow. * **What access does it need?** A script tag only, your code, or a staging environment. Check the privacy policy before you connect anything. * **What happens when it gets stuck?** Look for clear hand-off rules and tickets that carry the conversation. * **Which languages does it speak?** Match it to where your signups actually come from. ## How Artis compares with Appcues, Pendo, Userpilot and WalkMe Tool Pricing (October 2026) Approach Artis Free with $100 in credits, Growth from $100/month, usage-based Voice agent that clicks and fills the real UI, re-learned from the repo on each release Appcues No list price; based on MAUs, installations and published experiences No-code flows, tooltips, checklists and announcements built by your team Userpilot From $299/month for up to 2,000 MAUs; Growth from $849/month No-code tours and checklists plus product analytics Pendo Free way in; paid plans quoted on MAU volume Product analytics with in-app guides WalkMe Not published, quoted per deal Enterprise digital adoption, owned by SAP since 2024 [Appcues](https://www.appcues.com/) is the best-known name here. It suits product teams that want to design flows themselves, A/B test them and watch adoption numbers. The trade-off is upkeep: every flow is something your team maintains when the UI changes. Userpilot and Pendo sit in the same camp with more weight on analytics. WalkMe targets large companies rolling out software across many apps. The difference comes down to pointing versus doing. If you want precise control over every tooltip, the established tools are made for that. If you'd rather onboarding update itself when you ship, and you're happy for an agent to act in your app, Artis is built that way. ## Who should give it a go, and who shouldn't Artis makes most sense for a SaaS team that ships often, has a setup flow with real steps (invites, integrations, configuration) and loses signups in the first session. It also suits small teams without a customer success hire, where "book an onboarding call" really means the founder's evenings. It's a weaker fit if there's no repo or staging environment it can read, if setup is a single screen, or if your policies rule out an agent acting inside the app. Teams that mainly want analytics with a few banners will be happier with something like Pendo. Users outside English, Spanish and Portuguese are worth checking first. The quickest test is on the onboarding page of [Artis](https://www.tryartis.com/use-cases/onboarding): talk to the live agent yourself, then point it at your own product with the free credits. ## Straight answers ### What is SaaS onboarding? SaaS onboarding is everything that gets a new user from signup to their first real result in the product: account setup, settings, integrations and inviting the team. Software for it usually comes as in-app tours and checklists, or, like Artis, as an agent that does the setup steps with the user. ### What is an AI onboarding agent? An AI onboarding agent is software that guides new users through setting up a product and answers their questions as they go. Some only chat. Artis also acts: it clicks, fills in forms and navigates the real interface with the user, by voice or text. ### Does Artis need access to my code? It needs read-only access to a repo on GitHub, GitLab or Bitbucket, or read access to a staging environment. It never writes to your code, and nothing reaches customers until you publish the agent. ### How much does Artis cost per onboarding? About 350 credits. The Growth plan gives 40,000 credits for $100 a month, so a full onboarding walkthrough works out to under a dollar. The Free plan's $100 in credits covers a trial on your own product without a card. ## Read next - [Visit the Artis: SaaS Onboarding Software With an AI Agent That Does Setup product page](https://launchon.it/products/artis) - [All posts on the LaunchOnIt blog](https://launchon.it/blog) --- Markdown twin generated from the same data as https://launchon.it/blog/artis-saas-onboarding-software. HTML version: https://launchon.it/blog/artis-saas-onboarding-software