The integration doesn't exist
Some software genuinely has no usable API. We find that in the audit, not in week three, and we tell you before you pay rather than building a workaround you did not ask for.
Process
Four phases.Nothing hiddenin any of them.
Most agencies describe their process in adjectives. This is the actual sequence, what we need from you at each point, and what we do when it goes wrong — which it occasionally does.
- ~14 days to live
- 3–4 hours of your time
- Staged rollout
01 / Week 1
Audit
We map every route a customer can take into your business and find where they currently fall out. You get the map whether or not you hire us — it is useful on its own, and it is how we decide what to build first.
A written map of every leak, ranked by what it costs you
What happens
- Every route a customer can take into your business, mapped end to end
- Test calls, test enquiries and test forms, submitted as a customer would
- Response times measured — after hours, at weekends, during your busiest hour
- Each leak ranked by what it plausibly costs you a month
What we need from you
- One 20-minute call
- Read access to your calendar and CRM, if you use one
- Permission to submit test enquiries
02 / Weeks 1–3
Build
We build the highest-value system first, configured to your business rather than a template. You review every message, every script, every escalation rule before anything goes near a customer.
A working system, approved by you line by line
What happens
- The highest-value system first — not the whole stack at once
- Scripts and messages written for your business, in your register
- Escalation rules drawn explicitly: what it handles, what it hands to a human
- Integration into the tools you already run, in accounts that belong to you
What we need from you
- Your prices, services and hours, accurately
- A review pass on every script before launch
- One person who can answer questions within a day
03 / Week 2–4
Launch
Deployed in stages alongside what you already do, so nothing breaks while it goes live. We watch the first days closely and fix what real customers surface, because real callers always do something the test calls did not.
Live, monitored, with a rollback if anything misbehaves
What happens
- Staged rollout alongside your current setup, never a hard cutover
- The first days monitored closely, because real customers always surface something tests did not
- A rollback path tested before go-live, not designed after an incident
- A short handover so your team knows exactly what the system does
What we need from you
- A go-live window that is not your busiest week
- Someone reachable during the first two days
04 / Ongoing
Tune
Every month we read real transcripts and conversations and improve the system from them. This is the part most agencies skip, and it is the difference between automation that decays and automation that compounds.
Monthly report, monthly improvements
What happens
- Real transcripts read every month — by a person, not a dashboard
- Anything handled badly rewritten and redeployed
- Reporting on volume, source and what the system actually recovered
- Changes as your prices, services and hours change
What we need from you
- Twenty minutes a month, if you want to be involved
- Telling us when something about the business changes
When it goes wrong
Becausesometimesit does.
Every agency's process page describes a project that goes perfectly. Here is what actually happens when one doesn't, and what we do about it.
Real callers break the script
They always do something the test calls did not. That is why launch is staged and why the first week is monitored. Early surprises get fixed in days, not at the next monthly cycle.
The system oversteps
If it answers something it should have escalated, the escalation rule was wrong, not the caller. We tighten the boundary and, where it matters, we call the customer ourselves.
It isn't worth it for you
Sometimes the audit shows the leak is smaller than the fee. We say so and refund anything unearned. We would rather lose the sale than have you cancel in month three.
Practicalities
The fivequestionswe always get.
How long until something is actually live?
A single system is usually live in about two weeks from the first call, and most of that is scheduling rather than engineering. A full backend across several systems runs six to ten weeks, staged so that useful things go live early instead of everything landing at the end.
How much of our time does this take?
Roughly three to four hours total across the build: one call for the audit, one review pass on the scripts, and availability during launch week. The single biggest cause of delay is waiting on a review from the client, not the build itself.
What if we don't like what you build?
You see every script and message before a customer does, so there is no reveal at the end to dislike. If the direction is wrong at review stage we rewrite it — that is what the review is for, and it is not billed as a change request.
Do you need access to our systems?
Read access during the audit, and admin access only to the specific tools a system touches. Everything is created in accounts you own, so revoking our access never takes the system down with it.
What happens if something breaks at 2am?
Systems are monitored and every launch has a tested rollback, so a misbehaving system can be switched off in seconds and your calls fall back to how they worked before. Nothing we build is allowed to be a single point of failure for your phones.
Phase one is a call and a map.
You keep the map whether or not you hire us. It is useful on its own, and it is the only honest way to decide what is worth building first.