01
Context
Problem, stakeholders, constraints, success criteria, and claim boundaries.
Project evidence & capability infrastructure
See how real work was actually built.
Howthy turns complex, AI-assisted projects into structured Build Records—showing what was attempted, who contributed, which decisions were made, and what the evidence actually supports.
✓ Built for authentic assessment, capability evidence, and AI-transparent project work.
Live evidence view Review ready
AI & DATA CAPSTONE · REVIEW READY
Capacity-Aware Predictive
Maintenance Triage
Risk rankingModel evaluationDecision support
CURRENT CLAIM
The model prioritises higher-risk cases within a fixed inspection capacity.
SourceArtifact-linked
ValidationTested
OutcomeDemonstration only
LATEST DECISION · D-08
Optimise ranking utility, not failure probability
The available labels support relative prioritisation. They do not support physical-failure probability.
AK
Reviewer note
Claim boundary is well supported.
A Build Record connects the complete evidence chain
01Problem
02Decision
03Contribution
04Evidence
05Test
06Review
✓Outcome
The trust gap
Polished output is no longer enough.
AI can now produce convincing reports, code, dashboards, prototypes, and portfolios. The final output alone no longer tells the whole story.
Can you see...
01Who performed the work?
02What did the person understand?
03What did AI contribute?
04Which decisions changed the direction?
05What was genuinely tested?
06Which claims remain unverified?
“As producing plausible work becomes easier, proving trustworthy work becomes more valuable.”
The core product
One record. The whole story of the build.
A Build Record is a structured, evolving evidence object—not a polished case study written after the fact.
02
Contribution
Individual responsibilities, human and AI-assisted work, external input, and review.
03
Decision trail
Assumptions, options, rejected alternatives, rationale, and later corrections.
04
Evidence trail
Repositories, notebooks, sources, designs, tests, feedback, and metrics.
05
Validation & limitations
What passed, failed, was simulated, reproduced, reviewed, or remains unsupported.
06
Outcome & reuse
Results, lessons, reusable assets, replication guidance, and next versions.
How it works
Evidence without the surveillance.
Capture meaningful project events from the tools people already use. Review material claims—not every keystroke.
Set milestones, evidence requirements, AI-use rules, contribution expectations, and assessment criteria.
Connect selected repositories, documents, notebooks, submissions, presentations, and project artefacts.
Faculty, mentors, peers, or practitioners inspect the claims and evidence that matter.
Create an academic record, learner capability record, and concise employer brief.
Built for every decision
Different stakeholders. One shared evidence base.
Each view answers the questions that matter—without losing the project’s underlying context.
Institutions
Strengthen authentic assessment, moderation, auditability, and employability evidence.
Explore this viewLearners
Build credible proof of capability, a reusable professional asset, and stronger interview stories.
Explore this viewReviewers & faculty
Review material claims instead of reconstructing a project from scattered files.
Explore this viewEmployers
Understand contribution, judgment, trade-offs, validation, and limitations.
Explore this viewAn honest evidence model
No single score. No false certainty.
Howthy separates source, validation, review, outcome, and contribution. Strong evidence is visible. So are the boundaries.
Evidence dimension
Supported claim
Prioritised higher-risk cases.
Unsupported claim
Reduced operating costs.
Source status✓Artifact-linked!Insufficient evidence
Validation✓Tested—Not performed
Review✓Faculty-reviewed—Not reviewed
Outcome✓Demonstration only—Unsupported
Contribution✓Supervisor-attested—Not established
A strong claim can be tested and faculty-reviewed while still being limited to a demonstration. Howthy keeps those distinctions clear.
A distinct layer
Complements the tools you already use.
Howthy does not replace learning, project, or credential systems. It connects the decisions and evidence they leave behind.
LMSPortfolioBadgeAI detectorHowthy
Manages course content●———○
Displays final work○●○—●
Captures decision history—○——●
Links claims to evidence—○○—●
Distinguishes contribution————●
Records limits & validation————●
Creates academic & employer views—○——●
Where we start
Evidence-rich applied projects.
Howthy begins with complex technical work where decisions, evidence, and claim boundaries matter most.
01
AI & machine learning
→02
Data science
→03
Business analytics
→04
Data engineering
→05
Software & digital product
→06
Corporate AI upskilling
→Designed to connect with the tools where project evidence already lives.
Planned connections are directional, not claims of current availability.
GitHubGitLabGoogle DriveMicrosoft 365MoodleCanvasNotebooksOpen Badges
Start focused
Make project capability visible and trustworthy.
Run a focused pilot with one programme, one cohort, and one applied-project workflow.