Human-controlled AI agents for professional services

AI agents built, operated and owned by accountable people.

Babagana Zannah leads the work day to day. Mach Lilies keeps AI workflow scoping, build and monitoring connected, so you know who owns delivery before anything goes live.

Babagana Zannah, Team lead at Mach Lilies
Babagana Zannah Team lead · workflow mapping, build and oversight
One line of responsibility

Call → map → build → monitor

The work stays connected from the first fit decision to the live workflow.

  1. Call

    A builder tests whether the chase loop is practical and names what should stay human-led.

  2. Map

    The workflow, approval rules, exceptions and expected outputs are agreed in writing.

  3. Build

    Named people build and test the pilot. The work is not passed to a hidden delivery team.

  4. Monitor

    The people responsible for delivery stay close to the workflow once it is running.

The ownership record

Names beside every operating responsibility.

This is the format agreed before access. It is not a claim that an unnamed team or round-the-clock rota already exists.

Fit and workflow lead
Babagana Zannah · the named day-to-day lead shown on this page.
Build and test owner
A named person recorded in the proposal before system or sample access is granted.
Monitoring and incident contact
Named in the operating plan, together with the actual monitoring window, contact route and escalation cover.

Holiday cover, absence cover, specialist involvement and incident availability are contract-specific. Ask to see those names and terms; do not infer them from “one accountable lead”.

Quiet Quarter support uses hey@machlilies.com. Monitoring windows, response targets, third-party or tool costs, holiday cover and service levels are proposal-specific and recorded in signed terms.

Proposal-specific service levels · no public SLA

No public response-time or availability SLA is approved. Monitoring windows, response targets, third-party/tool costs and service levels are proposal-specific and governed by signed terms.

Named accountability

Babagana leads the work

Babagana has spent his career leading engineering teams and, in recent years, putting AI to work inside one of the UK’s largest companies. He also founded Africa Code Academy, which trains new software engineers.

At Mach Lilies he maps workflows, builds helpers and tests them before launch. Before work begins, you will know who is involved in your build and who remains responsible once it is live.

Check Babagana Zannah’s LinkedIn profile

How responsibility shows up

  • Builders stay involved. The people who scope the workflow remain connected to build, testing and monitoring.
  • Names before access. You know who is working on the helper before anything goes live.
  • Limits said plainly. If a workflow is not practical or worth automating, we say so before you commit.

The plain-English promise

  • If you have to re-read a sentence, we rewrite it. That rule covers this website, our emails, and the weekly summary your helper sends you.
  • Jargon gets translated or binned. When a technical word is genuinely useful, we explain it once, kindly. Otherwise it goes.
  • We’re honest about the limits. We’ll tell you what AI can’t do, and when it isn’t worth the money. If a helper isn’t right for your practice yet, we say so.

The money side, briefly

Every quote is fixed and agreed in writing before anything starts — never a day rate, never a surprise invoice. The current offer and scope are on the pricing page.

Start with a chase-loop assessment

No preparation needed. Tell us about your chase loop, and we’ll tell you — honestly — whether Quiet Quarter can take it off your team’s hands.

Book an AI workflow assessment