Glare uses a small set of words to explain how teams make product and design decisions. This glossary gives each word a short, simple meaning. That way teams, leaders and AI can all use the same words. Some words link to a full definition. It has examples and the rules Ray follows.
AI Agent
An AI system that works toward a goal. It picks the steps and uses Skills to get the work done. Glare's agent is Ray.
AI Skill
A set of steps that AI and teams can use again and again. A prompt tells AI what to do one time. A Skill shows how to do it the same way every time.
Skills help teams:
-
Frame decisions
-
Make user needs clear
-
Connect to business goals
-
Shape signals
-
Make sense of results
AI Workflow
How prompts, Skills, Ray, the Glare MCP and the team work together. It takes the team from a question to a decision. A simple AI workflow looks like this:
-
A prompt comes up with an idea
-
A Skill gives the work a shape
-
Ray runs the steps and saves what it finds
-
Answers from users create a signal
-
The team makes the decision
Ask
A question the team needs to answer. It is how work starts in Glare. Anyone can ask one. Written as: Should we [action] for [audience] to [goal]? See the full definition.
Assumption
Something the team thinks is true because of some data, but nobody has checked yet. Written as: Based on [data], we assume [conclusion]. A hunch may have no data at all. An assumption always points to some. See the full definition.
Audience
Who the evidence comes from. Glare uses four kinds: the Project Team, Stakeholders, Customers and Participants. Audience is the second block in Define. See the full definition.
Benchmark
A number you compare against. It tells you if a result is good or bad. Benchmarks help in Comparing in Focus and in Mapping in Lead. See the full definition.
Bet
A promise to make a product or marketing change that will affect how the business works. Bets is the last block in Focus. A bet is the biggest kind of decision. Written as: We bet [change] will move [goal], affecting [workflows]; measured by [metric]. See the full definition.
Business Goal
One of 20 common goals a business works toward. They come in five groups of four: Growth, Retention, User Experience, Efficiency and Operations. Leaders pick which goals matter most. Business Goals is the first block in Lead. See the full definition.
Business Intent
What the business wants to get done right now, and why. It ties to one business goal. It helps decide which questions are worth asking. See the full definition.
Business Outcome
A change you can measure in how the business does after a bet ships. Examples are sales, spending, customers leaving or cost. See the full definition.
Collecting
The Define block for learning with an open mind, before anyone has a hunch. The team gathers what is out there. They decide how much to trust each source. They mark what is missing. Then they hand a Source Map to Situations. No hunch yet: Collecting. A hunch about a concept: Testing.
Comparing
The Focus block where you run the same test again while one thing changes. Then you look at the difference. A/B tests, visual iterations, before-and-after tests and benchmarks are ways to do it.
Concept
An idea the team can show people and test. Written as: [Name]: [what it does] for [audience]. Concepts is the first block in Measure. The concept comes first. Then hunches are written about it. See the full definition.
Constraint
A firm limit on what the team can try. It might be a budget, a deadline, a technology, a rule or a brand rule. Concepts have to fit inside it. See the full definition.
Decision
A choice the team makes, at any step. A bet is the biggest kind. Written as: We decided to [choice] because [reason]; owner [person]. In Glare, signals back up decisions. See the full definition.
Decision Map
The part of Glare that shows how decisions get stronger. It has four facets. Each facet has four blocks, so there are 16 blocks in all. The last block in each facet is where the work comes together:
-
Define ends in Situations
-
Measure ends in Findings
-
Focus ends in Bets
-
Lead ends in Results
Define
The first facet of the Decision Map. It is judged by credibility. Define finds out what matters through User Needs, Audience, Collecting and Situations.
Design Assessment
A way to see how well design decisions move through a company. The Design Assessment looks at five areas:
Design Impact
The change a design decision causes. Design impact is easier to see when a decision connects to:
-
A user need
-
A UX metric
-
A design signal
-
A product or business outcome
Design Review
The time when the team looks at the work, talks about it and tries to decide. In Glare, Design Review is where signals keep the talk on track so decisions get made.
Design Signal
Evidence that has been read and given a meaning, so it can move a decision forward. A signal says what the evidence shows. It says if it supports or goes against what the team was testing. Each signal is rated strong, moderate or weak. Written as: [Evidence] shows [interpretation]; supports or contradicts [test]. A signal always rests on evidence. So a hunch is never a signal. See the full definition.
Evidence, Expectations and Proof
The three layers the Glare MCP keeps behind every decision:
-
Evidence: what the team knew, and where it came from
-
Expectations: what the team expected, like hunches, assumptions, named situations and business priorities
-
Proof: how well an expectation held up once users showed the answer, growing over time
Findings
Everything that comes back from one round of testing. That includes the concept, the hunches and questions, what was seen and measured, what it means and what to do next. Findings start with the call on the hunch. Then comes the design signal. Findings is the last block in Measure. See the full definition.
Focus
The third facet of the Decision Map. It is judged by desirability. Focus runs one test again while one thing changes. Then it places a bet. It does this through Initiatives, Variables, Comparing and Bets.
Glare MCP
How people and AI reach Glare's evidence, expectations and proof. It brings back what the team already has. Each piece shows where it came from, so a decision can show its reasons.
Glarify
The shared way Glare turns any result into something one reader can use. It shows up in the place they already look. It works through six choices in order: evidence strength, purpose, audience, content shape, presentation approach and destination. It never makes weak evidence look strong.
Hunch
A guess about what is true or what will work. It may have no data behind it, and that is fine. Written as: We think [belief] because [intuition].
In Measure, a hunch is written with the questions that would prove it right or wrong. Then it is tested. A hunch has a perspective. External means it is about users. Internal means a part of the business holds it. A hunch is never evidence. A test decides if it is supported or contradicted. See the full definition.
Initiative
A body of work the team takes on after testing shows it is worth doing. Written as: Improve [situation] for [audience] by [timeframe]. Initiatives is the first block in Focus. See the full definition.
Input
Any raw material that comes into Glare. It could be a report, a deck, a transcript, a data export, a survey file or an email. Every other record links back to its input. See the full definition.
Intuition
What the team notices from experience and good judgment. Glare does not throw out intuition. It gives intuition a shape, starting with a hunch. Then teams can test it against user needs, metrics and decisions.
Lead
The fourth facet of the Decision Map. It is judged by viability. Lead follows bets to the results leaders track. It does this through Business Goals, Workflows, Mapping and Results.
Mapping
The Lead block that links design KPIs to product KPIs, and then to business KPIs. That way a change in the experience can be traced to use, people staying and sales.
Measure
The second facet of the Decision Map. It is judged by usefulness. Measure runs one test to get an answer. It does this through Concepts, Hunches, Testing and Findings.
Method
A named way of working, like journey mapping, A/B testing or RICE. In Glare, methods are ways of working inside a block. Comparison designs go in Comparing. Test types go in Testing. Ways to rank options go in Bets. See the full definition.
Observation
Something the team saw or heard during testing. It is written down just as it happened, without saying why. Written as: [Who or how many] [did or said] [what]. See the full definition.
Opportunity
A gap between what users need and how well they get it today. Written as: [Audience] needs [need], but today [current state]. See the full definition.
Participant Input
Answers from people who try a concept, flow, message or experience. Participants are one of the four audiences. Their evidence counts for more than the project team's or stakeholders' when a signal is rated.
Product Outcome
A change you can measure in how the product does after a bet ships. It shows if more people use what was built, and keep using it. See the full definition.
Prompt
A one-time instruction you give an AI tool. Prompts are good for new ideas, new ways to ask a question, new versions or quick exploring. They are flexible, but they do not last. Skills make that work repeatable.
Ray
Glare's AI agent. Ray reads inputs and finds records like hunches, signals and workflow changes. It scores them and links them. It sends unclear ones to a review digest so the team can check them.
Recommendation
A next step that comes out of testing. It always points back to the evidence behind it. Written as: [Action] because [evidence]. See the full definition.
Result
What all the work delivered toward a business goal by the end of a set time. It is judged Met, Partly met or Missed. Results is the last block in Lead. See the full definition.
Risk
Something that could go wrong. It names a problem that has not happened yet, what it would hurt and what would cause it. Written as: [Risk] could [impact] if [condition]. See the full definition.
Signal Loop
The steps that repeat from a question to a decision:
-
Start in Define with the need, the audience and the situation
-
Write a hunch about a concept, with its questions
-
Test it and make the call in Findings
-
Compare while one thing changes, then place a bet
-
Follow the bet to results
-
Carry what held up, and what did not, into the next round
Situation
What the team agrees is happening: how people act and what that leads to. Written as: [Audience] is [behavior], which [effect]. Situations is the last block in Define. Only the team can accept one. See the full definition.
Source Map
What Collecting hands to Situations. It shows what evidence there is, how much to trust each source and what is missing.
Test
How the team checks if a hunch or assumption holds up. It says what is checked, with whom, how, and what result to expect. Written as: Test [hunch or assumption] with [audience] using [method]; expect [result] on [metric]. See the full definition.
Test Question
One clear question a test must answer to tell if a concept works. It is small and direct. You can answer it from what happens in a session. See the full definition.
Testing
The Measure block that puts a hunch up against real evidence. That could be a study, a standard survey or data the team already has. Each result is rated weaker, moderate or stronger evidence. It is saved with its source.
Tradeoff
What the team gains or gives up when it picks one direction over another. Signals make tradeoffs easier to see. They show how each option does on what matters.
User Need
What people need from the product to succeed. Glare uses 20 common needs in five groups: Basics, Trust, Personal, Impact and Feelings. User Needs is the first block in Define. Without a clear user need, signals drift toward the team's opinion or business-only goals. See the full definition.
User Outcome
A change you can measure in what people experience after a bet ships. It shows if the product got easier, clearer or more useful for the people it was built for. See the full definition.
UX Metric
A number that measures what people experience with a product. Each one is tagged three ways: its type, whether it predicts or confirms change, and how far up it reports (Design, Product or Business KPI). Common UX metrics include:
-
Comprehension
-
Usability
-
Confidence
-
Desirability
-
Trust
-
Completion
Variables
The Focus block that names the one thing that changes between tests. It could be a version, an audience, a competitor or a time period. Glare has 12 variables, from Iteration to Season.
Watch Item
Something the team agrees to keep an eye on after a decision. It has a clear line. If the number crosses that line, the team looks at the decision again. Written as: Watch [metric]; if it [threshold], reopen [decision]. See the full definition.
Workflow
One of the eight parts of a business that a change has to go through: Sales, Marketing, Product, Engineering, Strategy, Operations, Finance and Legal. A workflow record says how one part will work differently because of a bet. Workflows is the second block in Lead. See the full definition.
