# From Launch to Handover: What You Actually Own When the Build Is Done

*Nikhai Jaysen · November 7, 2026*

> A voice AI agent or chatbot going live isn't the end of the engagement. Here is exactly what SpaceBlanket.AI hands over, and why most agencies skip it.

Source: https://www.spaceblanket.ai/blog/what-you-own-when-the-build-is-done

A voice AI agent or chatbot going live isn't the end of the project, even though most pitches treat it that way. Here is exactly what we hand over when a build is done: documentation, access, a change-request path, and the numbers actually worth watching.

A build going live gets treated as the finish line in almost every pitch deck, including ours when we started. In practice, the question we hear most often comes about a week after launch, and it's rarely about the agent's accuracy. It's some version of: if your team disappeared tomorrow, what would we still have? For a **voice AI agent**, a chatbot, or a workflow automation, the honest answer depends entirely on what gets handed over at the end of the build, and most AI automation agencies never make that part explicit.

## Why Handover Gets Skipped

It gets skipped for an ordinary reason: it isn't the exciting part. The demo is exciting. The go-live call is exciting. Writing down which Twilio number belongs to whom, or recording why a prompt handles a specific objection the way it does, feels like paperwork after weeks of building something that actually works. So it gets compressed into a five-minute screen share at the end of the project, and three months later nobody on the client's side remembers which account the knowledge base lives in.

We run an [automation audit before we scope anything](/blog/automation-audit-four-things-we-check), and one of the four checks in that audit is whether the client has someone who will own the system after launch. Handover is where that answer gets tested. An audit can confirm someone is willing to own it. It can't hand them anything to own if the agency never writes it down.

## What We Actually Hand Over

Four things, every time, regardless of whether the build is a [voice AI agent](/services/voice-ai-agents) or an [AI chatbot](/services/ai-chatbots):

- **The documentation.** Every prompt, every flow branch, and the reasoning behind each one, written so a non-technical team member can read it, not just the raw configuration file a developer would need.

- **Access and credentials.** The phone number, the API keys, the CRM connection, the hosting. These live in accounts the client controls from day one. We configure them during the build; we don't hold onto them afterward.

- **The change-request path.** A named way to ask for an edit to a flow, a script, or a trigger condition, without reopening a full project scope for a change that takes twenty minutes.

- **The numbers worth watching.** A short report of what to actually track week to week, call volume, escalation rate, the handful of metrics that tell you something is drifting, instead of every number the platform happens to produce.

## Where Handover Actually Breaks Down

The honest failure mode isn't technical. It's that no one on the client's team was ever named as the person who reads the documentation. We've seen clients ask us to keep holding the credentials "for convenience," which feels harmless in the moment and quietly turns a one-time build into a standing dependency. We've also watched the opposite happen: a team takes full ownership on day one, then [edits a live agent's prompt directly](/blog/updating-live-voice-ai-agent-without-breaking-it) without the documentation in front of them, and breaks a flow that had been tuned over weeks of real calls. Handover isn't just a folder of files. It's making sure a specific person knows what's in that folder and what to do with it.

This is also why the first month after launch matters more than the demo did. [What actually happens in those first 30 days](/blog/voice-ai-agent-post-launch-first-30-days) is largely a test of whether the handover held up: whether the right person is reading the logs, whether a change request actually gets used instead of ignored, whether the system is still anyone's job three weeks in.

## What This Changes

None of this is complicated to do. It just has to be treated as part of the build, not an afterthought once the invoice is paid. A client who owns their documentation, their credentials, and a clear way to request changes isn't dependent on the agency that built the system. They're running it, with us available when they need deeper changes, not because they have no other option.

If you're evaluating an AI automation agency and haven't asked what you'll actually walk away with, [get in touch with our team](/services/ai-consultation). We'll show you exactly what handover looks like before you sign anything, not after.

## What SpaceBlanket.AI hands over when an AI automation build goes live: the steps
1. **The documentation.** Every prompt, flow branch, and the reasoning behind each decision, written in plain language a non-technical team member can read, not just the raw configuration.
2. **Access and credentials.** The phone number, API keys, CRM connection, and hosting live in accounts the client controls. We configure them; we don't hold them.
3. **The change-request path.** A defined, named way to ask for an edit to a flow, a script, or a trigger, without opening a new project every time something small needs to change.
4. **The numbers worth watching.** A short report of what to actually track week to week, not every metric the platform can produce, so someone on the client's team knows what matters.

## Frequently Asked Questions

### What does SpaceBlanket.AI hand over when a build goes live?

Four things: documentation written in plain language, access and credentials held in the client's own accounts, a defined path for requesting changes, and a short report of the metrics actually worth tracking. Together these decide whether an automation stays useful after the agency stops being involved day to day.

### Who holds the API keys and phone numbers for a voice AI agent after launch?

The client does. We configure the telephony, the CRM connection, and the API keys during the build, but they live in accounts the client controls from day one. That way nothing about the agent's operation depends on our account staying active.

### What happens if a client wants to change something after launch?

They use the change-request path set up during handover: a named way to ask for an edit to a prompt, a chatbot script, or a WhatsApp trigger. It avoids both extremes, waiting on a full new project for a small edit, or a team member guessing at a live configuration with no documentation.

### Why do some AI automation builds stop working a few months after launch?

Usually because no one on the client's side was handed anything to maintain it with. The agent or chatbot still runs, but the knowledge base goes stale, no one reviews the call logs, and small drift goes unnoticed until a customer complains. A real handover is what prevents that.

## Related Service

### [Voice AI Agents](/services/voice-ai-agents)

Handle inbound calls, qualify prospects, and follow up with leads, using natural-sounding AI voice agents that never miss a call.

## Related Reading

### [What Happens After You Launch a Voice AI Agent: The First 30 Days](/blog/voice-ai-agent-post-launch-first-30-days)

Most of the conversation around voice AI agents focuses on the build. What doesn't get discussed is what happens after go-live, and that's where the real work begins.

### [What We Actually Do Before Writing a Single Line of AI Code](/blog/what-we-do-before-writing-ai-code)

The phase most teams underestimate isn't the build. It's everything before it. Workflow mapping, edge case documentation, and data contract design. Here's what we actually do before writing any AI code.

### [The Homework List We Send Before Any AI Build Begins](/blog/client-handover-checklist-ai-automation-build)

The contract is signed and the kickoff call is booked, but the build doesn't actually start until five specific things land in our inbox. Here's what we ask every client for before writing a single line of an agent's logic, and what slips when one of them is late.

## Related
- [All articles](https://www.spaceblanket.ai/blog)
- [Voice AI Agents](https://www.spaceblanket.ai/services/voice-ai-agents)
- [All AI automation services](https://www.spaceblanket.ai/services)
- [Contact](https://www.spaceblanket.ai/contact)