FAQ
Questions, answered
What Wellence does, how agents stay in check, and what happens to your data.
What is Wellence?+
Wellence puts AI agents to work alongside your people, team by team, and shows whether it's working. Every agent has a human owner, a goal, a spending cap and guardrails, and asks a person before anything is sent, posted or changed.
Who is it for?+
Leaders who've been asked to make AI work across their company, and the teams doing the work. Most companies start with one team and roll out from there.
We already build agents in Copilot or ChatGPT. Why Wellence?+
Building an agent is the easy part. Running them is hard: knowing which agents exist, who owns each one, what it costs, whether its work is any good, and which goals it moves. Wellence is where your people and their agents work together, and where that's managed and measured.
Do I need to be technical?+
No. You describe the job in plain English and Wellence builds the agent with you in about ten minutes: its goal, its steps, its guardrails and who does what. You approve it before it runs, and you can try it on a sample first.
Does it work with the AI we already have?+
Yes. Wellence works inside Claude: connect it once and you can run and build agents from your own Claude, on your own Claude plan, with the tools you've already connected. Companies can also use their own Anthropic, OpenAI or Gemini key, or run it in their own AWS Bedrock, Google Vertex AI or Azure OpenAI account, and require that only their own AI is used, with no fallback to Wellence's.
Can I run Wellence agents inside our own Claude?+
Yes. A Claude owner adds Wellence as a connector, then adds an agent as a skill for the team (or anyone adds it for themselves). The model and your Claude connectors run in your Claude. The agent's owner, approvals, run history and data rules stay in Wellence: each run starts and ends through the Wellence connector. Anything Claude does outside that connector isn't recorded in Wellence. Developers can also make it a Claude Managed Agents agent, which isn't covered by zero data retention.
Will agents do things without asking?+
No. Agents ask a named person before anything is sent, posted or changed, and each agent has an owner who is accountable for its work.
Can other tools start a workflow?+
Yes. Any app that can send a web request (Zapier, n8n, Make, a form) can send data to a workflow's own private link, signed with a secret or, for tools that can't sign, a simpler link. By default the workflow's owner is asked before it starts. It can also start from a new Typeform response, new rows in a Google Sheet, a new Gmail email or a new Pipedrive deal (Sheets and Gmail are checked every few minutes, so up to about 15 minutes late). What comes in is used as data, never as instructions. See Data and security.
Can a workflow change things in our other tools?+
Yes, if you add a step for it: add a Google Sheet row, update a HubSpot contact or deal, create a Notion page, or add a Pipedrive note or an Airtable record. It never deletes anything. Each change is shown first and waits for approval in the Inbox, unless the workflow's owner lets it save on its own, and every change is kept in a record.
How do we know it's working?+
Every run can be rated, and every agent has a score, a track record and a monthly cost. Leaders see use, quality and spend by team, and which goals agents are moving.
What are Wellence AI credits?+
Prepaid AI usage in Wellence, in pounds, for when the month's AI in your plan is used up. Admins buy them for the company, and team leads for their team, by card. They're for your workspace only, have no cash value and expire 12 months after purchase. Companies that only use their own AI never need them.
How do caps work?+
Agent runs typically cost cents. Set a monthly cap for the company, a team, a person or an agent, and see the spend as it happens. Each cap counts inside the ones above it. When a cap is reached, work pauses: ask for more in one click, or finish on a cheaper model. A cap can warn instead of stopping.
Is our data safe?+
Each company's data is kept separate, people see only their own team and what's shared with them, and admins can download or delete everything at any time. Agents run on Anthropic's Claude, which doesn't train its models on API data. See Data and security.
What can agents read in our tools?+
Your files, email and messages only as the person the agent is working for, through that person's own connection. Where a folder, file, sheet or channel is picked in its setup, it reads only that; anything else is refused and recorded. Work from someone's own files is shown only to them until they share it. See Trust and AI.
Can anyone with our email address join our workspace?+
Not unless you want them to. New people from your company's email domain wait until an admin lets them in. New companies start this way, and an admin can change it in Settings. People you invite join straight away.
Is Wellence GDPR compliant?+
Wellence Ltd is a UK company registered with the ICO. For the data you put into Wellence, you're the controller and we're your processor, under our Data Processing Agreement. Our GDPR page lists where data is stored, who processes it and how to use your rights.
How long does it take to get started?+
Minutes. Create your workspace with your work email, take the two-minute AI diagnostic to see where AI can help, then add a ready-made agent or build your own.
Can we start with one team?+
Yes, and we recommend it. Start with one team, see it work, then roll out to the rest of the company with the same owners, caps and guardrails.
How do I get help?+
In the app, click Get help: it answers straight away and passes anything it can't answer to Wellence Support. Or email support@wellence.ai.