Making expertise accessible when it matters
SLAN is built around a simple belief: expertise is earned through research, practice, and mistakes.
Our goal is to make that expertise available at the moment of decision, replacing guesswork with grounded guidance.
Why we built it
Concepts are taught. Application is where people get stuck.
Under constraints, “I understand it” becomes “I don’t know what to do next.” The missing layer is structured judgment: assumptions, tradeoffs, and checks.
Time pressure
You don’t have time to re-learn the theory. You need the shortest path to a defensible output.
Example: “What checks do I run before I decide?”
Imperfect or missing data
You must act without perfect information, so assumptions and risks must be made explicit.
Example: “What must be true for this recommendation?”
Unclear incentives
Stakeholders optimize for different things; tradeoffs need to be surfaced early.
Example: “Who wins/loses with this option?”
Ambiguous expectations
People disagree on what ‘good’ looks like. Structure makes success criteria visible.
Example: “What does ‘complete’ mean here?”
Principles
What we believe in
Boundaries
What we are not
- Not a generic internet-wide chatbot.
- Not a tool designed to replace instructors or experts.
- Not a shortcut for answer dumping.
- Not a source of ungrounded, generic advice.
- Not a decision-maker you can delegate to.
- Not a black box where you can't see the logic.
- Builds structured paths with checks + completion criteria (not answer dumping).
- Grounds guidance in your materials and teaching intent (not generic internet AI).
- Makes assumptions and tradeoffs explicit so outputs are defensible (not vague advice).
- Keeps humans accountable: supports decisions, doesn't make them (not delegatable).
- Makes the logic visible in steps you can review (not a black box).
- Supports governance for proprietary content and cohorts (access control, IP boundaries).
Team
Who’s behind SLAN
Selena Tabbara
Ex-AWS Professional Services (London). Shipped production AI systems across forecasting, anomaly detection, and GenAI workflows with customer teams.
At AWS, I learned that most "best practice" advice sounds like common sense, yet teams still get stuck executing it. Workshops didn't fix that. Coaching them through the decision process a few times did. SLAN turns that coaching into repeatable, structured guidance.
Read the longer version
I built SLAN after seeing the same failure mode everywhere: people understand concepts in theory, then reality shows up: time pressure, incomplete data, unclear incentives... and then they freeze.
At AWS, I helped customers ship production-grade AI that delivered business outcomes, not just prototypes. But the most important lesson wasn't technical: telling teams to "identify the right use case" or "find the right data sources" rarely changed behavior, even when packaged as a one-day workshop.
What worked was guiding them through the process repeatedly until the steps became obvious and repeatable. SLAN is built around that idea: expertise becomes usable when it's structured into a path you can follow, not a recommendation you're supposed to magically execute.
- Structured reasoning paths: steps, checks, completion criteria
- Decision quality under real constraints (not idealized conditions)
- Governance + IP boundaries for proprietary content and cohorts
Want to see if SLAN fits your course or academy?
We’ll scope it to your materials, your governance constraints, and your learners.