Vestigo
Most businesses automate the wrong things.
AI automations and agents for 5-to-50-person businesses. We design the systems, deploy them into accounts you own, and stay to watch them run. Your team gets more capable, not smaller.
02 · The premise
The bottleneck was never your team.
Ten people reading ten thousand records is a staffing problem. One system reading them continuously is a design problem. Most businesses keep hiring at the first one because nobody has looked hard at the second.
03 · What we build
Four shapes of problem worth automating.
Read → Decide
Two days → twenty minutes
Quotes, estimates and applications get read, priced against your rules, and drafted. A person approves or overrides, then spends the hours won on work worth winning.
Answer → Escalate
Every lead answered inside the hour
Inbound leads and support messages answered in your tone, escalated by confidence score. The threshold is a setting your team owns and can change any day.
Chase → Collect
Eleven emails → one form
Onboarding documents and unpaid invoices are collected, validated, filed and chased. Exceptions route to a person by design, because that's where judgement is needed.
Assemble → Review
Rebuilt by hand → built by the system
The Monday report and the month-end reconciliation, assembled automatically. Your analyst reviews the anomalies, which was always the part that needed them.
04 · See it work
Don't take our word for it. Use it.
Answer three questions and get a first read on what's worth automating in your business, including what isn't. No email, nothing stored.
Demo · your automation read
60 seconds · no email · nothing stored
01 · Your business
02 · Team size
03 · What eats the most time? Pick up to three
Your read
Answer the three questions
Curated from real builds. The Full Read personalises this to your exact stack.
On the call we do this on your actual processes, live.
05 · Approach
One way in. Three steps.
01 Discovery call
Thirty minutes
A conversation, not a pitch.
Thirty minutes to understand how your business runs: where work stalls, what gets repeated, and which of those problems AI can genuinely solve. If the honest answer is that automation won't help you yet, we'll say so.
You get: a clear read on what's worth automating. No cost, no deck.
02 The Full Read
Two weeks
You see the whole plan before anything gets built.
We map the process end to end and write a proposal: what gets built, what it connects to, what it deliberately won't do, and the result you should expect. Fixed scope, fixed price, fixed dates, all approved by you before work starts.
You get: a written proposal you can act on, with or without us.
03 Build
Three to six weeks
Built, deployed, and watched.
We build the agents, test them against your real cases, and deploy into the systems your team already uses. Then we monitor how they perform, tune what needs tuning, and train your people to run them.
You get: working systems in your accounts, documented and handed over. You own the code and the accounts.
06 · Questions
Before you ask.
What does it cost?
Every engagement is fixed scope, agreed in writing before work starts. No hourly billing, no creep. The discovery call is free, and the Full Read fee credits against a build booked within 60 days. We'll give you numbers on the call, once we know what you need.
Where does our data go?
Language work runs through the business APIs of the major AI providers, and we pick the model that fits each job. Your databases, files and workflow runners sit in accounts you own. Nothing you send is used to train models, and we set zero-retention terms where a provider offers them. Confirmed in writing in the proposal.
What happens when the system gets something wrong?
It's designed to know when it might. Anything below the confidence threshold stops and waits for a person instead of guessing, every action is logged, and we watch the first weeks in production. When something does slip through, you have an audit trail and you have us.
We already tried ChatGPT and Zapier. It didn't stick.
Same. Off-the-shelf tools automate steps, and the work that eats your week is a process. We build around your edge cases, the part DIY skips, and give every system a human decision point so your team trusts it enough to use it.
Do we get locked into your tooling?
No. It runs in your accounts, you own the code and the workflows, and handover includes documentation written for whoever comes after us.
You have no case studies yet.
Correct. That is why the demo above tells you what not to automate, why pricing is fixed and agreed before work starts, and why the Full Read is a standalone deliverable you keep whether or not you hire us for the build.
What if our data is a mess?
It always is. Finding that out is part of what the Full Read is for. Messy data is a scoping question, not a disqualifier.
Is this going to replace our team?
These systems take the volume and hand the judgement back. Every build has a human decision point; a proposed system without one is the wrong system, and we will say so.
How fast?
The Full Read takes two weeks. Builds run three to six, depending on how many systems they touch.
Start with a conversation.
Thirty minutes, no deck. Tell us where the work piles up and we'll tell you whether there's anything here worth building.