The short answer
About a week for the audit, then two to four weeks before the first workflow is live in your environment. A full first engagement usually runs six to eight weeks end to end. Building one workflow at a time rather than delivering everything at the end is what keeps that first live date early.
The timeline, stage by stage
Week one — the audit. Sitting with the people doing the work and counting where the hours actually go. This produces a written picture of what is worth automating, what is not, and what each piece would cost. It is deliberately short. An audit that takes a month is usually a consultancy filling time.
Weeks two to five — the first workflow live. Not a demo, not a prototype in a sandbox: the real thing, in your environment, being used by the people whose job it is. Two weeks when the process is clean and the systems have APIs. Four when neither is true.
Weeks five to eight — the rest, then handover. Remaining workflows go live one at a time, then documentation, training and a named owner on your side who has actually changed the system while there was still someone there to ask.
Why one workflow at a time
The alternative is a three-month build that arrives all at once, and it fails in a specific and predictable way: you discover in month three that a decision made in week two was wrong, and by then everything sits on top of it.
Shipping one workflow at a time means you see something real early enough to change direction while changing direction is still cheap. It also means that if the engagement stops for any reason — budget, a change of priority, a company emergency — you are left with something that works rather than something that is 70% built.
What actually makes it slower
Almost never the building. In practice the delays are:
- Access. Waiting on credentials, a licence, or someone in IT who has three other priorities. This is the single most common cause of slippage and it is entirely avoidable by sorting it in week one.
- Nobody owning the decision. A workflow with four stakeholders and no decision-maker will sit still while everyone waits to see what everyone else thinks.
- An undocumented process. If the process only exists in one person’s head, the first job is getting it out of their head, and that person still has their day job.
- Discovering the process is contested. Two departments doing the same thing differently, each believing theirs is the standard, is a management question before it is a technical one.
How to make it faster
Pick the person who will own it before the work starts, and give them the time. Sort access in the first week. Choose the process that annoys everyone rather than the one that is technically most interesting — momentum on an unloved workflow is worth more than elegance on an obscure one.
And start with one process. A business that automates one workflow properly in six weeks is far ahead of one that is still scoping five of them in month four.
Last updated 18 August 2026 by Hugo, founder of Hugo Signal.
Bring me the process that wastes the most time.
A first call takes thirty minutes. You will leave it knowing whether automation is worth doing at all in your case.