Bytefo// a two-person workshopStart a project

Two builders · AI automations · Our own products

We automate the boring parts, then build the fun ones

Bytefo is two AI enthusiasts who wire automation into the work nobody wants to do — and ship our own products on the side. Small team, hands on the keyboard, no hand-offs.

Start a projectRead the blog
  • Automations
  • AI tools
  • Integrations
  • Products
AI automationsCustom AI toolsWorkflow designAgentsIntegrationsData plumbingPrototypesInternal appsVoice & chatOur own products

Capabilities

What we build with AI

Three things, done properly. No fifty-item service menu — we would rather be genuinely good at a short list.

Automations

We find the task your team repeats every week, then wire AI and your existing tools around it until it runs itself.

  • Inbox, invoice and ticket triage
  • Lead research and enrichment
  • Reporting that writes itself
  • Handover docs, not black boxes

Custom AI tools

Small, sharp apps and internal assistants built around how you already work — not another platform you have to migrate to.

  • Chat and voice assistants on your own data
  • Internal dashboards and admin tools
  • Document and knowledge search
  • Prototypes in days, not quarters

Our own products

We are building products of our own. Everything we learn shipping them goes straight back into client work.

  • First product in the workshop now
  • Built in public, documented on the blog
  • Small scope, shipped early
  • Same two people on the tools

The bench

Pick the job that eats your week

What did you do this week that you have done before? Choose the one that sounds familiar and see the shape we would build — including the part that stays with a person.

Choose the repeating job you want to see built

Job 01 · Intake

Intake that needs sorting

InEmail, forms and tickets. Messy, inconsistent, written by humans.

  1. Readmessy input into structured fields
  2. Decidea rule picks the destination
  3. Enrichyour CRM, your data
  4. Deliverthe tool they already open

Anything uncertaingoes to a human queue with context attached. The automation handles 90% and hands you the tricky 10% with a summary.

What the model does
Reads the messy input and turns it into structured fields. That is the job it is genuinely excellent at.
What the rules do
Ordinary deterministic code decides what happens next. When something goes wrong you can read the rules and point at the line — try doing that with a prompt.
Inside your tools
The automation runs where the work already happens, not in a new dashboard somebody has to remember to open.
Email us about this oneOpens your mail app — nothing is sent from here

Job 02 · Documents

Documents that need reading

InContracts, invoices, PDFs, and the knowledge base nobody has indexed.

  1. Readthe fields you actually care about
  2. Deciderules, not a prompt
  3. Enrichthe sentence it came from
  4. Deliversearchable, where you work

Every extracted fieldkeeps the source sentence attached, so a person can check it in seconds instead of re-reading the document.

What the model does
Reads the document and pulls out the fields you care about, however the document is laid out.
What the rules do
If the output format lives in one person’s head, that is a documentation problem first, and a code problem second. So we write it down in code, where it can be tested.
No drifting copy
If the answer lives in your CRM, the automation reads your CRM. It does not keep a hopeful copy of your data that drifts out of sync by week three.
Email us about this oneOpens your mail app — nothing is sent from here

Job 03 · Updates

Updates that need writing

InThe Monday report somebody rebuilds by hand, the status update, the summary.

  1. Readwhat actually happened
  2. Decidea schedule and a template
  3. Enrichthe context that matters
  4. Delivera draft, for a person

Nothing sends itselfA person reads and approves every outgoing update before it leaves the building.

What the model does
Reads what happened and writes the first pass — the version you edit instead of the version you write.
What the rules do
The model reads; the rules decide. What goes out, to whom, and when, is decided by code you can read afterwards.
The person doing the work
The people who own the process are watching while it is built. A report that nobody trusts is worse than no report at all.
Email us about this oneOpens your mail app — nothing is sent from here

Job 04 · Follow-ups

Follow-ups that need remembering

InLeads, quotes, and promises made in a call three weeks ago.

  1. Readthe notes and the CRM
  2. Decidea schedule and a rule
  3. Enrichwhat was actually promised
  4. Delivera nudge, drafted

Nothing sends itselfThe machine remembers and drafts; a person decides whether it is the right moment to send.

What the model does
Reads the notes and the CRM to find what was promised, to whom, and how long ago.
What the rules do
Cron, webhooks and a queue decide when to nudge — before frameworks. Reaching for a platform first is how you end up maintaining a platform.
Uncertain is a state
Anything the machine is not sure about is drafted for a person to send, never sent blind. That is the difference between a system people trust and a system people unplug.
Email us about this oneOpens your mail app — nothing is sent from here

How we would know it worked

Not a benchmark

Hours back per week

Measured before and after, from the people doing the work.

Touch rate

How often a human has to intervene. This is the maintenance bill, and it should fall over the first month.

Time to answer “why did it do that?”

If nobody can explain a decision in a couple of minutes, the automation is not finished.

Process

How a project actually goes

Boring on purpose: short loops, working software early, nothing hidden behind a status report.

  1. Tell us the annoying bit

    One call or a long email. We want the process that makes your team groan — the spreadsheet, the copy-paste, the "who is doing this today?".

  2. We map it and pick the fight

    You get a short written plan: what we automate, what we deliberately leave alone, what it costs and what it saves. No jargon wall.

  3. Working version, fast

    Days, not months. You use it while it is still rough, we fix what actually hurts, and we throw away the parts that did not earn their place.

  4. Handover and tune-up

    You own the thing: documentation, access, and a walkthrough with your team. We stay around for the first weeks of real use.

The workshop

Our own products, out in the open

Client work pays the bills; products are the long game. Here is what is on the bench right now.

Product 01

In progress

An AI toolkit aimed at the operational busywork small teams drown in. Same automations we build for clients, packaged so you can run them yourself.

Naming and launch notes will land on the blog first.

Follow along

Your project

Slot open

We take on a small number of builds at a time so each one gets real attention. That means there is usually room for one more.

Automation, custom AI tool, or something we have not thought of yet.

Talk to us

About

Why the goat

A company of two AI enthusiasts. Between us we handle the whole thing — the automation, the interface, the integrations and the unglamorous glue.

A goat standing on a steep, uneven slope

Bytefo is byte plus baifo — the Canarian word for a young goat. Small, curious, sure-footed on ground that looks impossible to everyone else. We could not think of a better description of what we do with AI, so we kept it.

Small is the feature

You work with the two people who build your automation. No account manager, no telephone game, no discovery phase that produces a slide deck.

Boring beats clever

The best automation is one nobody notices. We optimise for things that keep working after the novelty wears off — and we document them so they survive without us.

Ship, then sharpen

Working software in days tells you more than a plan in weeks. We get something real in front of you early, then improve it with evidence.

Say hello

Got a process that eats your week?

Tell us what it is. We will tell you honestly whether AI can fix it — and if it cannot, we will say so instead of selling you something.

What you get back

  • A real reply from one of the two of us.
  • An honest read on whether AI helps here.
  • If it does, a plan with days in it instead of phases.
Email us
hello@bytefo.com