A manually assembled decision memo
- 120 minutes
- 5 handoffs
- 1 typical rewrite
- Reviewer confidence: 3/5
The inherited assumption was that leadership needed one manually assembled narrative before it could compare the options.
01 AI work design for people and teams
AI changed what
people can build.
Most work still follows the old path.
Premise & Practice helps people and teams choose the right work to change, rebuild it around what AI makes possible, and prove the new way in practice.
02WHERE YOU'RE STARTINGCHOOSE THE CLOSEST VERSION
Start where
the work is stuck.
You do not need to diagnose the offer before you arrive. Start with the situation you recognize. The right form of work follows from there.
Organizational starting point
Choose the work that has earned the right to become a pilot.
Map the real work, surface inherited assumptions, rank the opportunities, and narrow the field before a team spends time building the wrong thing.
03WHAT WE DOTHE WORK, NOT THE CAPABILITY LIST
Three ways to move
from premise to practice.
The form changes with the decision. The discipline does not: start with the outcome, keep human judgment visible, and make the result prove itself.
The Audit
Several AI ideas. One useful place to begin.
WHO IT'S FOR
See the full offer ↗Teams and leaders with pressure to act, but no defensible way to choose the first opportunity.
WHAT HAPPENS
- 01Map the recurring work
- 02Rank value, feasibility, and risk
- 03Define one bounded pilot
WHAT YOU LEAVE WITH
A prioritized opportunity map, pilot brief, and proof plan.
WHAT THIS IS NOT
Not a technology shopping list or a generic AI maturity score.
04THE PREMISE TO PRACTICE METHODMETHOD / V1.0
Keep the judgment visible
from the first question
through real use.
The method is the connective tissue across every offer. It turns a point of view into a repeatable way of working.
- 01ASSUMPTION
Find the old premise
Name the constraint, handoff, or belief that shaped the current work.
- 02REVISION
Rewrite the work
Design the path around what people and AI can do now.
- 03WORKING VERSION
Build close to reality
Make the smallest version that can survive real data and real boundaries.
- 04EVIDENCE
Prove it in practice
Watch what happens, keep what helps, and change what does not.
05PROOF OF PRACTICEINSPECT THE WORK
The work has to
hold up close.
Follow one example from the inherited workflow through the working system and into the evidence review. See what changed, what failed, what still belongs to a person, and what the evidence supports.
SYNTHETIC TEACHING CASE
Method and artifacts are real. Results are illustrative, not client outcomes.
A source-linked decision brief
- Facts separated from estimates
- Conflicting evidence preserved
- Human approval remains visible
- Unsupported claims stop the system
A program lead verifies the evidence and completeness. Leadership still chooses the option.
The evidence did not earn a victory lap
- Adoption was uneven
- Exceptions were omitted
- Time savings were unmeasured
- ROI could not be supported
The review produced three defensible choices: extend, narrow and repair, or pause.
ARTIFACT / 01Workflow Definition
The before-state, outcome, and inherited assumption.
ARTIFACT / 02Working System Brief
Inputs, boundaries, human approval, quality bar, and stop condition.
ARTIFACT / 03Evidence Record
Real uses, corrections, failures, uncertainty, and the next decision.
FOUNDER NOTE / RUSTY WALTERS
Close enough to the work to know where the tidy story breaks.
WHY PREMISE & PRACTICE EXISTS
Access became common.
Judgment did not.
Rusty works at the intersection of AI, work design, data, and adoption. He built Premise & Practice to help nontechnical people use AI in consequential work without surrendering judgment or accountability.
The work starts before the prompt: choose the right problem, make the human decision visible, build a bounded system, and prove what holds up.
More about Rusty and the work →WORKING PREMISE→BETTER PRACTICE
What part of your work
still assumes the old rules?
Bring the real version. We will work out whether it belongs in an Audit, a Build Day, an advisory conversation, or somewhere else entirely.
Start a conversation ↗