deiadora · relational pathways for AI
AI resolves conversations too early.
Someone expresses doubt, and the model returns an answer — at exactly the moment the conversation was starting to move. Quantum Magic is an API that gives AI systems a structured map of how constructive conversations actually progress — pathways like doubtful → curious → confident — so a response can support the next step instead of forcing a conclusion.
It doesn't detect emotions, and it doesn't simulate empathy. It recognizes where a conversation is, and keeps constructive movement available.
the first 13 pathways are open source · the API is in development
the observation
You've seen this happen. A conversation with a current AI assistant — reconstructed from a pattern that repeats across models:
Person
I've been thinking about leaving my job. I don't know. Everyone says the market is terrible right now.
Assistant
That's a big decision. Here are five factors to weigh before leaving a job in a difficult market: 1. Your financial runway. 2. Your industry's hiring cycle. 3 …
The answer is competent. It's also the end of the conversation. The person wasn't asking for a framework — they were starting to work out what they actually think. The model resolved the exchange at the exact moment it was beginning to move. And the pattern is consistent: the more uncertain the person, the faster the system reaches for a solution.
the pattern
Constructive conversations move along recognizable pathways.
Here's the same opening line with a good friend — or a good manager, or a good therapist. Notice what the conversation does instead of ending:
1 · Doubtful
"I've been thinking about leaving my job. I don't know."
The friend doesn't answer. She asks what's underneath the doubt.
2 · Curious
"Honestly? I want to know if I can do more than this role."
The doubt has become a question worth exploring. Nothing was solved — something opened.
3 · Confident
"I'm going to start looking — on my terms, not the market's."
Nobody handed over a decision. The conversation carried the person from doubt to confidence, through curiosity.
That movement — doubtful → curious → confident — is a pathway, and it repeats. So do others: fearful → aware → courageous. Angry → mindful → compassionate. Lonely → self-sufficient → communal. Since 2013, across individuals, teams, and organizations, I've documented 264 of them. Each one is an arc: a starting state, the step that opens movement, and the constructive form the conversation can reach. The first 13 are open source — you can read them now.
same message · two responses
a standard response
"That's a big decision. Here's a framework: assess your financial runway, update your materials, set a timeline …"
Closes the conversation. The doubt is still there — it just has homework now.
a pathway-aware response
"It sounds like the doubt isn't really about the market. What's pulling at you about leaving?"
Keeps the movement available. The conversation can reach curiosity — and confidence is reachable from there.
The second response isn't kinder, and it isn't therapy. It's informed by structure: knowing that doubt resolves constructively through curiosity — not through a delivered conclusion.
the api · quantum magic
The pathways, structured so an AI system can follow them.
Quantum Magic is the 264 documented arcs as a structured, versioned API. A system using it can recognize which pathway a conversation is in, identify the constructive next step, and evaluate whether a response preserved movement or foreclosed it.
one arc, simplified
{
"arc": "doubtful-curious-confident",
"origin": "doubtful",
"opening": "curious",
"constructive_form": "confident",
"markers": ["hedged intent", "externalized blockers", "…"],
"response_guidance": "hold the question open; do not resolve"
}
What it is not.
It doesn't detect or classify emotions. It doesn't score sentiment. It doesn't simulate empathy or care. It isn't a safety or moderation layer. It models how conversations move — nothing about the person, everything about the movement.
A system prompt can say "don't rush to solutions," but it gives a model no structure — no way to know which pathway a conversation is in, what the next step is, or whether a response preserved it. A structured map is inspectable, versioned, and evaluable — the same reason evaluation criteria outperform good intentions.
API in development · the first 13 arcs are open source now → · API updates →
track record
Deiadora Blanche has spent her career building content systems for organizations that communicate at scale. She proposed, built, and launched Airbnb's first internal AI model for its help center — still in active use — and stood up Coursera's content design system, adopted globally across four product verticals. She taught AI content strategy to hundreds of practitioners at Fortune 100 and 500 organizations, and is a published researcher on language-model evaluation (The Transmutation Gap, Apart Research, 2026). The pathway research behind the API spans thirteen years of field work across individuals, teams, and organizations.
Airbnb
First internal AI model, help center
Proposed, built, tested, and launched — strategy, training data, evaluation criteria, A/B testing, rollout.
first internal AI model · still in active use
Coursera
Content design system
Built and led the design system and quality standards for cross-vertical teams — learners, enterprise, institutions, government.
four product verticals · still in use at scale
UX Content Collective
AI content strategy curriculum
Taught content operations and AI content strategy to practitioners at Fortune 100 and 500 organizations, and authored two modules on Dynamic Content Strategy for the Advanced UX Content Product course (launched June 2025).
hundreds of practitioners trained · 2022–2026
Apart Research
Published evaluation research
The Transmutation Gap: Cross-Lingual Coherence Evaluation in Large Language Models (2026). Peer research evaluating language-model behaviour across languages.
model-behaviour evaluation · published
from deiadora
I spent seven years building the content systems that shape how organizations talk to people at scale — Airbnb's help center, Coursera's design system, curricula for Fortune 100 and 500 teams. Then AI started taking over those conversations, and I watched it handle them badly in a consistent way: it answers too soon. Someone begins to move through doubt or fear, and the model closes the conversation with a solution. Since 2013 — long before the product work — I had been documenting how constructive conversations actually progress: the recognizable pathways by which doubt becomes confidence and fear becomes courage. Quantum Magic is those pathways, structured so AI systems can follow them. The conversations were already moving. I wrote down where they go.
— Deiadora Blanche
the wider world
The pathway research didn't start as a product, and it doesn't end as one. Where to go, by what you're looking for:
open source
i am deiadora
The first 13 arcs, free to read and use.
read the arcs →
the api
quantummagic.ai
The 264 arcs as a structured API. In development.
follow the build →
research
the quantum ceo
The Field Research series — five volumes behind the arcs.
read the research →
teaching
techtuition.ai
Advisory, builds, and team training for content teams in the AI era.
begin the workshop →
mentorship
rebel evolve
1:1 and team sessions built on the same research.
enquire →
speaking
deiadora.live
Podcast appearances, keynotes, and team sessions.
book the speaker →
fiction
deiadorebel.life
Five novellas — the field notes that came first.
explore the books →
music
deiadorebel
Five albums; each song paired with a poem from the novellas.
listen →
community
drbl exchange
Needs meet offers. Invite-based, built on mutuality.
say hello →
why it matters
AI now sits inside millions of difficult conversations a day — support queues, classrooms, health questions, three-a.m. doubts. If those systems can recognize the pathways conversations move along, they can support the movement instead of cutting it short. That's what this is for: healthier human interactions, at the scale AI already operates.
Start where you like.
Explore the 13 open pathways. Read the research. Get updates when the API opens. Or email directly — every conversation here starts peer-to-peer. Also available for senior leadership, strategy, and research roles in human-centred AI, AI content strategy, and relational intelligence.