Scoring by design,
not by intuition.
HUMANOVA publishes its full impact readiness methodology. Every point awarded has an explicit rule. Every AI report is decision support — not a verdict.
HUMANOVA's impact screening platform
CAHI (Centre for AI-Assisted Human Impact) is HUMANOVA's primary tool for evaluating social impact projects. Applicants submit structured proposals. A scoring engine and Claude AI produce an impact readiness report. Human reviewers make the final call.
Transparency as a principle
Applicants deserve to know exactly what they're being evaluated on. HUMANOVA's scoring rules are deterministic and public — not a black box. If the methodology is wrong, anyone can challenge it.
How the score is built
Each dimension is scored 0–100 and weighted into a final readiness score.
Does the project clearly identify a real need, a specific audience, and measurable improvement in human, animal, or ecological wellbeing?
Is the execution plan realistic? Does the team have the budget, time, resources, and partners to deliver what they're proposing?
Does the project bring something new — a novel method, technology, delivery model, or combination — that justifies HUMANOVA's active involvement?
What is the likelihood that this project fails to deliver, and how fragile is the proposal? Lower risk means a higher score.
Does the project contain enough structured data to support Social Return on Investment analysis later — costs, outcomes, beneficiaries, duration?
Current scoring criteria
24 criteria across 5 dimensions. Rules are deterministic — the same input always produces the same score.
Clear problem
How precisely the project describes the social, health, or inclusion need it addresses.
Defined target group
Whether the beneficiary audience is specific, reachable, and relevant to HUMANOVA's mission.
Scale of need
Estimated reach and urgency — both the number of people affected and where.
Measurable outcomes
Presence of concrete outcomes and success metrics that could be tracked and evaluated.
HUMANOVA mission alignment
Alignment with health, performance, wellbeing, longevity, education, research, and inclusion.
Realistic budget
The budget appears proportional to scope and not financially fragile.
Realistic timeline
The timeline appears practical for the proposed activities — neither too compressed nor open-ended.
Available skills and resources
Evidence that the team understands the people, tools, and capabilities required.
Partner support
Named delivery partners or institutional allies signal that dependencies are being managed.
Risk mitigation
The team anticipates execution risks and describes active mitigation actions.
Novelty
How distinctive or original the proposal appears relative to existing programs.
Technology or methodology
Use of new methods, science, digital tools, biometrics, or evidence-based delivery models.
Scalability
Potential to expand, replicate, or transfer the intervention to new contexts.
Differentiation
Clarity on what makes this solution different from existing options in the same space.
Budget risk
Likelihood that financial scope or cost-per-beneficiary could undermine delivery.
Timeline risk
Likelihood that timing assumptions could compromise execution quality.
Operational risk
Likelihood that execution complexity or capability gaps could block delivery.
Dependency risk
Exposure to fragile partners or single points of failure in the delivery chain.
Evidence risk
Likelihood that weak measurement plans will reduce confidence in impact claims.
Project cost
Cost inputs are clear enough to serve as the investment figure for later SROI work.
Measurable outcomes
Outcomes are defined clearly enough to assign future monetary value proxies.
Beneficiary count
Reach estimates are specific enough to support per-person valuation.
Duration of benefit
The team indicates how long benefits may last — a key input for SROI modeling.
Attribution & deadweight
The team acknowledges attribution, deadweight, or proxy assumptions — hallmarks of SROI literacy.
AI assists. Humans decide.
The screening process is a pipeline — not an algorithm that makes decisions autonomously.
Applicant submits
The applicant completes a structured form across 7 sections: identity, description, goals, resources, impact, risk, and innovation. Completeness drives the score.
Engine scores
A deterministic scoring engine evaluates the submission across 5 weighted dimensions. No machine learning — every point awarded has an explicit rule and explanation.
Claude reviews
HUMANOVA uses Claude (Anthropic) to produce a structured screening report: project summary, strengths, weaknesses, missing information, and suggested reviewer questions.
Human decides
A HUMANOVA reviewer reads the AI report as decision support — not as a verdict. They can request revision, escalate to committee, promote to pilot candidate, or reject with rationale.
Outcome logged
Every decision is recorded with the reviewer ID, previous status, new status, and written rationale. Applicants receive structured feedback they can act on.
Ready to submit your project?
Use the structured form to build a HUMANOVA-ready submission. Your score updates live as you write.