Questions, answered plainly.

Shapes is an AI workflow agency. We design, build, and run AI agents and automations on n8n, connected to the tools a business already uses, with people in control where it matters.

No. Start with the problem and the process you want to improve. We assess the options and recommend an approach on a free call.

Yes. Every workflow is scoped around your process, systems, and rules. The approach, deliverables, cost, and timeline are agreed before work begins.

Yes. Shapes is remote-first and has delivered projects for clients in 14 countries across Europe, the United States, Asia, and Oceania.

Book a free call or describe the workflow you want to improve. You do not need a technical specification to start the conversation.

Each project is quoted after we understand the work: the workflow, the systems involved, and the level of testing and support needed. You receive a written proposal before anything starts.

A single workflow typically takes two to six weeks from agreed scope to launch, depending on the systems, integrations, and testing involved.

A short, fixed-fee review of your processes. We map how the work is done, score each candidate workflow, and give you a prioritised roadmap and a scope for the first build. It is yours to use with any team.

n8n is open, connects to almost any system, supports AI agents natively, and can run in your own account. That means you own your workflows and data, with predictable costs.

n8n Cloud is the quickest to run. Self-hosting in your own cloud account gives more control over data location and cost at volume. We recommend one based on your situation and set it up.

You do. We build inside your n8n instance, and every workflow is documented so your team or another partner can maintain it.

Yes. We hand over with documentation and, if you want, training, so your team can make everyday changes safely.

Yes. We rebuild automations in n8n, run them alongside the old ones until results match, then switch over in planned steps without downtime.

Whichever fits the task and your data rules: OpenAI, Anthropic Claude, Google Gemini, or open-source models. We confirm each provider's data terms before use.

Only where you agree it should. Sensitive, uncertain, or high-impact steps are designed with a human checkpoint and a clear owner.

In your n8n instance and the systems you already use. With self-hosting, you choose the region, for example within the EU.

We use AI providers and settings that do not train on your data by default, and confirm each provider's data handling terms before use.

Yes. A data processing agreement is available on request for any project that handles personal data.

Every project ends with documentation and a handover. Managed support plans cover monitoring, fixes when tools change, and steady improvements.

Every workflow reports errors to a named person. On a support plan, we monitor and fix issues within agreed response times.