THE SHORT VERSION
- The bottleneck is almost never the build — it is DNS access, A2P registration and decision latency.
- Infrastructure comes before assets: domains, authentication and phone registration first, funnels second.
- Two review gates, at day 7 and day 13, keep a fourteen-day build from becoming a six-week one.
- Training that is not recorded is training that leaves when the person does.
Fourteen business days is not an aggressive timeline for setting up a GoHighLevel agency. It only sounds aggressive because most agencies take six weeks, and they take six weeks for reasons that have almost nothing to do with the platform.
This is the sequence that fits in fourteen days. Whether you run it yourself or hand it to a white-glove onboarding team, the order is the part that matters.
Day 0 — Discovery (two hours, once)
Everything downstream is decided here, so do not do it in fifteen-minute increments over a fortnight.
What has to come out of it:
- Your services, as a list. Each one becomes a pipeline or a stage, and if you cannot list them cleanly you do not have a process to automate yet.
- Your delivery process per service. What happens after someone signs, in order, with who does it.
- Your reporting needs. What you send clients monthly and what you look at weekly. These become dashboards.
- The three things that currently break. Almost always: leads not followed up, onboarding chased by email, and renewals missed.
- Domains you control and who holds the DNS.
Days 1–2 — Blueprint before build
Write the plan down before touching the account. Pipelines and their stages. Custom fields and their types. The tag taxonomy with its prefixes. Which snapshots go where. Which automations exist and what triggers them.
This document is boring and it is the reason the build does not need redoing. An agency that skips it discovers in week three that its tag scheme cannot distinguish a lead source from a lifecycle stage, and everything built on top has to be revisited.
Days 3–5 — Infrastructure first
This is the order that saves the fortnight. Do not build funnels yet.
- Agency branding and white-label — login domain, colours, logo, mobile app branding.
- Domains and DNS — agency domain, client-facing subdomains, and the sending domain, which should be a subdomain of your main domain rather than the domain itself.
- Email authentication — SPF, DKIM and DMARC, verified rather than assumed. Then a warm-up schedule, because a cold domain sending a thousand emails on day one is a domain you will be replacing.
- Phone and A2P — provision numbers, submit brand registration, submit campaign registration. Start this on day 3 because approval is not on your schedule.
Everything after this depends on this being right. Anything built before it has to be revisited once the sending identity exists.
Days 6–9 — Assets and snapshots
Now build. Install the snapshots that match your verticals, rewire them to your agency’s standards, and build the pieces the snapshots do not cover — your own prospect pipeline, proposal flow and onboarding sequence. If you are starting from Agency Core OS, most of that spine already exists and this phase is customisation rather than construction.
Build in a staging sub-account. Not because the platform demands it, but because a half-built workflow that fires on a real contact is an embarrassing way to learn this lesson.
Day 7 — Review gate one
Thirty minutes. Walk the blueprint against what exists. The only question that matters: is anything here going to need rebuilding later? Structural problems found on day 7 cost an afternoon. The same problems found on day 13 cost the timeline.
Days 10–12 — Roles, permissions and reporting
- User roles. A VA should not be able to delete a workflow. Build the restricted permission set now, before there are eleven users.
- Dashboards. The client-facing monthly report and the internal weekly view.
- Notification routing. Who gets told what, and through which channel. Default settings notify everybody about everything, which trains everyone to ignore notifications.
Day 13 — Review gate two, then test
Run every workflow end to end with real test data. Submit forms. Receive messages. Book slots. Move stages. Watch automations fire.
Write down what failed, fix it, and run it again. A workflow that has never been executed is not built, it is drafted.
Day 14 — Training and handover
Two sessions, both recorded:
- Daily operators — how to work the pipeline, respond to conversations, and book. Forty-five minutes.
- Admin — how to change a template, add a user, edit a workflow safely, and what not to touch. Sixty minutes.
Plus a written runbook: what each pipeline is for, what each workflow does, which custom values control what, and what to check first when something looks wrong.
Training that is not recorded leaves when the person does. This is the cheapest insurance in the whole project.
Where the fortnight actually goes wrong
In practice, three things cause almost every overrun:
- DNS access. Nobody knows who holds the registrar login. Ask on day 0.
- A2P registration. Started late, then blamed for the delay. Start it on day 3.
- Decision latency. A build waiting four days for an answer about pipeline names is not a fourteen-day build. Two named review gates fix this better than any amount of chasing.
None of those are platform problems, which is why hiring someone to “do GoHighLevel faster” only helps if they also run the sequence. If you want the whole thing run for you on a fixed timeline, that is exactly what white-glove onboarding is, and agency account setup is the infrastructure half on its own.