m00labs
Request a quote

Insights

Practical notes from the lab on getting AI to work inside real operations. No hype, no vendor theology — just the questions that decide whether an AI build earns its keep.

Request a quote
Technical explainers

Three figures behind the notes

Illustrative concept
Figure 1 · Five questions before automating Decision checklist 1. Does the workflow follow a rule — or does it need judgement? 2. How much time does it actually take, and how often? 3. Where does the data live, and how do you reach it? 4. What happens when it fails? 5. What does being wrong cost? Figure 2 · Hosted against open-source Hosted Open-source Lower operational load More control, more to run Fastest to a resultLeast to operate Runs on your termsCost and data control Vendor dependencyWhere hosted ends Operational effortWhere open-sourcestarts Figure 3 · Fixed scope against open scope Fixed scope Agreed deliverableA definition of done both sides accept Change is re-scopedNew work is quoted, not absorbed Open scope Moving deliverablePriorities shift, so finish lines move
Figure 1 — five questionsThe five questions, takenin order: does the workflowfollow a rule or need judgement;how much time does it take andhow often; where the data livesand how you reach it; whathappens when it fails; and whatbeing wrong costs.Figure 2 — hosted againstopen-sourceThe trade is operational loadagainst control: hosted servicesreach a result faster,open-source weights run onyour terms and ask more of youroperation.Figure 3 — fixed againstopen scopeFixed scope names thedeliverable up front andre-quotes new work; open scopekeeps the deliverable moving,which is why finish lines movewith it.
  1. Figure 1 — five questionsThe five questions, taken in order: does the workflow follow a rule or need judgement; how much time does it take and how often; where the data lives and how you reach it; what happens when it fails; and what being wrong costs.
  2. Figure 2 — hosted against open-sourceThe trade is operational load against control: hosted services reach a result faster, open-source weights run on your terms and ask more of your operation.
  3. Figure 3 — fixed against open scopeFixed scope names the deliverable up front and re-quotes new work; open scope keeps the deliverable moving, which is why finish lines move with it.
Illustrative concept. Three independent explainer figures for the notes on this site: the five questions to settle before automating a workflow, the hosted-versus-open-source trade, and fixed against open scope. They sit side by side as separate figures — figure 1 is not a prerequisite for figure 2 — and describe general engineering practice, not measured results from any engagement.

Notes

Automation

Five questions to ask before automating a workflow

Every automation conversation starts with a tool and ends with a workflow.

Read the note

Model choice

Choosing between hosted and open-source models

Model choice is a scoping decision, not a religion.

Read the note

Engagement design

Why fixed scope works for AI builds

AI work is easy to start and hard to finish, and the shape of the engagement has a lot to do with why.

Read the note

Put these notes to work.

Bring us a workflow. Whether it is worth automating is the first thing scoping answers, and hi@m00labs.com is the published email address.

Request a quote About the lab