They notice the wrong thing
Your evidence may be present, but another cue dominates attention.
Applied research and consulting
Why good information fails.
Violet helps campaign, product, AI, change, and service teams diagnose why an important message or journey is ignored, misunderstood, distrusted, or abandoned—even when the information is accurate.
Bring one consequential message or journey. Violet will show what people may notice, what they need to understand, where the next step becomes difficult, and what to change or test.
Start with a short description. Add detail only if it helps.
Interactive receiver demonstration
The same announcement creates different questions for each audience.
Message:
Customer
- Will this affect me unfairly?
- Can I appeal?
- Will a person review the decision?
Executive
- Will this reduce cost?
- What is the operational risk?
- Will adoption succeed?
Engineer
- How accurate is it?
- How is drift monitored?
- What happens when it fails?
Regulator
- Who is accountable?
- Is it explainable?
- Is there evidence of fairness?
Choosing a role changes which questions are easiest to see. Real people also bring their own experience, goals, trust, and circumstances, so these are starting hypotheses—not predictions about every person.
What Violet changes
The message says “efficiency.” The customer needs to know who decides—and how to challenge an error.
Original statement
We are introducing an AI system to review customer applications and improve processing efficiency.
What the sender intends
Applications will be reviewed faster and more consistently.
What the customer may ask
Will a machine make the final decision? Who is accountable if it is wrong? Can I ask for a human review?
One stronger version
AI will help our reviewers sort applications; it will not make the final eligibility decision. A trained reviewer remains accountable. If your application is declined, you can request a human reconsideration using the link in your decision notice.
Why it is stronger
- Name what the system does—and does not do
- Identify the accountable human decision-maker
- Put appeal and recourse beside the announcement
- Answer the receiver’s risk before promoting efficiency
Communication failure
“We sent it” is not the same as “they could act on it.”
They interpret it through a different prior
Trust, role, history, and current goals influence what the message means.
They never reach the intended action
The receiver may understand the words without making the transition you expected.
Violet identifies these risks before they become expensive.
Deliverable
Attention diagnosis
Shows which cues are likely to dominate attention, what gets overlooked, and where the message may be too busy or too thin.
Deliverable
Interpretation risk map
Surfaces the likely meanings a receiver can form, including the reasons those meanings differ from the sender's intention.
Deliverable
Intervention and redesign directions
Recommends the smallest justified change—whether to the message, interface, sequence, context, or process—plus a practical way to test it.
How the audit works
Diagnose the path before making the message louder.
01
Bring one failed transition
Bring the message, interface capture, video, presentation, resume and role, or journey evidence—then name the next step that is not happening.
02
Map the receiver, field, and path
Violet identifies the active job, likely receiver states, the surrounding location, people, institution, timing and constraints, and the largest unsupported jump. A preceding induction cue is checked only when one exists.
03
Audit the 5S cues
Spotlite, Substance, Source, Strain, and Stake are assessed for this receiver and transition.
04
Redesign and test
You receive a practical redesign, competing explanations, and a measurable next-step recommendation.
This is not generic copy editing. The audit examines the message in relation to the receiver, context, prior beliefs, trust, goals, and stakes.
Violet in live application
See the framework used before the result is known.
Applied Projects follow real work through diagnosis, a recorded prediction, intervention, measurement, model update, and synthesis. Earlier predictions remain visible, and null or negative findings stay in the record.
01 Observe
Separate submitted facts from what the material directly shows.
02 Predict
Name the receiver, transition, rival, and result that would weaken the explanation.
03 Test and update
Measure outcomes, mediators, guardrails, implementation, and model change.
Before and after · Entire process
Redesign the path, not only the sentence.
Illustrative AI feature onboarding
01Entry
Before
A banner says “Unlock productivity with AI” before the user has identified a job the feature can help with.
After
The entry asks what the user is trying to produce, then shows one relevant example and the evidence needed to begin.
02Trust
Before
A permission modal asks for broad data access using generic language about improving the experience.
After
The interface names the exact data used, what the model will do, what it will not do, retention, and who remains accountable.
03Setup
Before
Seven required fields appear before the user has experienced value or knows which answers materially change the output.
After
The flow asks for the minimum useful input, infers cautiously, and reveals optional context only when it can improve the result.
04First result
Before
The system returns a confident recommendation with no visible evidence, uncertainty, or alternative path.
After
The result connects claims to source evidence, shows uncertainty, and makes correction or rejection easy.
05Next action
Before
A generic “Activate” button asks for commitment before the user understands consequences or recourse.
After
The action states what will happen next, what can still be changed, and how success and guardrails will be measured.
More before and after
See the method applied to other material.
Work with Violet
Not every problem fits inside one message.
If the failure spans a journey, product, policy, dashboard, onboarding flow, or organizational change, start with a consultation. Violet can scope an Attention Audit, Transition Experiment, or the minimum measurement work needed to test the problem.
Research foundation
A practical framework with claims that can be tested—and fail.
Receiver state + Message cues + Context
→
What receives attention
→
How it is interpreted
→
What becomes more likely next
The framework separates exposure, attention, comprehension, belief, action, and benefit. A click is not treated as proof of understanding.
5S audit
Each cue is evaluated relative to a receiver, context, and target transition—not scored as a universal message property.
- Spotlite
- Substance
- Source
- Strain
- Stake
Practical guidance
Start with the problem closest to yours.
Why Good Messages Fail: The Receiver Is Part of the Message
Correct information does not automatically create understanding. The receiver’s goals, expectations, knowledge, and trust shape what becomes important—and what happens next.
Read the guide →How to Communicate an AI Rollout to Employees
A credible AI rollout message answers the receiver's employment, workflow, data, and accountability questions before asking for enthusiasm or adoption.
Read the guide →Digital Health Engagement Is a Journey Design Problem
Digital health platforms can integrate more care, benefits, and guidance while making the next patient or member decision harder. The product challenge is designing a trustworthy path from information to informed action.
Read the guide →One useful next step
Bring the message that is not producing action.
Paste it, attach it, or simply describe the failure. Violet can begin with incomplete context and ask for what matters next.