How to Rebuild Customer Onboarding: Zero-Touch for Self-Serve, Concierge for Enterprise

Many SaaS companies run one onboarding process for every new account. That usually means self-serve customers get emails written for enterprise buyers, and enterprise customers wait in the same queue as everyone else for a kickoff call. Splitting onboarding into two paths fixes both: an automated, zero-touch path for accounts that can get started alone, and a concierge path where a person leads, supported by automation in the background.

This guide covers how to decide who goes where, what each path needs, and how to build the workflows.

Step 1: Define the routing rule

Start with the signal that decides the path. Keep it simple enough that sales, success and finance agree on it. Common options:

• Plan or contract type. Self-serve checkout goes zero-touch; a signed order form goes concierge.
• Seats or contract value. Above an agreed threshold, concierge.
• Requirements. SSO, a security review, data migration or a custom integration push an account to concierge regardless of size.

Write the rule down and put it in one place the workflow reads, such as a field on the account record set at signup or contract signature. Do not let the onboarding workflow infer it from five different fields; that is how accounts end up on both paths, or neither.

Also decide what happens in between. An account on a self-serve plan that invites many users in the first week, or asks for SSO, is a candidate to move to concierge. Build that upgrade path in from the start.

Step 2: Define activation, per path

You cannot automate onboarding toward a goal you have not named. Pick the one or two actions that show an account is getting value, for example “connected a data source and ran a first workflow” or “invited a teammate and completed a first project”.

The activation event may differ by path. A self-serve user may be activated after one successful run; an enterprise account may need integrations live and a first team trained. Track each as an event in your product analytics or database so the workflows can react to it.

Step 3: Build the zero-touch path

The goal is that a self-serve account can reach activation without talking to anyone, and that someone notices when it does not.

Trigger: account created with the self-serve flag.

Workflow outline:

1. Welcome message sent immediately, with one next step, not a tour of every feature.
2. Wait for the activation event, with a timeout. If activation happens, send a short confirmation and the next most useful step, then exit.
3. If a setup step stalls, send help specific to where the user stopped: an article for the integration they started but did not finish, rather than a generic “need help?” email. This needs product events for each setup step, not just signup and activation.
4. If there is no activity after the timeout, send one plain-text check-in from a real person’s address, and create a low-priority task for the success team so replies are answered.
5. Exit on activation, cancellation or a move to concierge, so nobody gets onboarding emails after they no longer apply.

Things that make this path work in practice:

• Behaviour-based branching beats time-based drips. A message tied to what the user did, or did not do, is relevant; a day-3 email sent to someone who finished setup on day 1 is noise.
• Fewer messages. Each one should exist because a specific stall needs it.
• In-product guidance where possible. An empty-state prompt inside the app often does more than an email about the same step.

Step 4: Build the concierge path

Here a person owns onboarding, and automation handles the logistics so that person spends their time on the customer.

Trigger: contract signed, or an account moved over from the self-serve path.

Workflow outline:

1. Assign an onboarding owner based on segment, region and current load, and notify them with the deal context from the CRM: what was sold, who the champion is, what the customer said they need.
2. Create the onboarding project from a template: kickoff, technical setup, integrations, training and a review point. Assign tasks to the right internal people (solutions engineer for SSO, success manager for training).
3. Send the customer a kickoff scheduling link from the owner’s address, with a short list of what to prepare.
4. Track milestones as product events arrive (SSO configured, first integration connected, users invited) and update the project automatically, so the owner sees progress without asking.
5. Escalate stalls internally. If a milestone is overdue, notify the owner and their lead before the customer has to chase.
6. Hand off to ongoing success at activation, with a summary of what was set up and any open issues.

The test for this path is whether the onboarding owner could run several accounts in parallel without losing track. If they keep a spreadsheet next to the workflow, the workflow is missing a step.

Step 5: Handle the move between paths

Accounts change. Build explicit transitions:

• Self-serve to concierge. Triggered by an upgrade, a large seat increase or a request for an enterprise feature. Stop the zero-touch sequence, assign an owner and create the onboarding project with the steps already completed marked done.
• Concierge customers adding self-serve workspaces. Decide whether a new workspace inside an enterprise account gets zero-touch messages. Usually it should not, or the champion gets emails meant for a trial user.

Step 6: Measure each path separately

Track, per path:

• Share of accounts reaching activation, and how long it takes
• Where accounts stall (which setup step)
• For concierge: time from contract to kickoff, and from kickoff to activation
• How many self-serve accounts move to concierge, and why

Averages across both paths hide the problems. A slow concierge path and a leaky zero-touch path can average out to something that looks fine.

Trade-offs and common mistakes

• Over-automating concierge. Enterprise customers notice when the “personal” kickoff email is a template with a merge field. Automate the scheduling and project setup, and let the owner write the messages.
• Under-staffing zero-touch. “Zero-touch” still needs someone to answer replies and review stalled accounts each week.
• Building before instrumenting. Without product events for each setup step, the zero-touch path falls back to time-based emails. Add the events first.
• One giant workflow. Keep each path, and the transition between them, as separate workflows that share sub-workflows for common steps like notifications. They are easier to test and change.

Where to start

If you are rebuilding from a single onboarding flow, start with the routing rule and the activation definition. Then build the zero-touch path with just the welcome message, the activation wait and the no-activity check-in. Add stall-specific help and the concierge project automation once you can see, in your own data, where accounts actually get stuck.