Six expert minds. One unified answer.
Economic mobility decisions are multidimensional. A salary increase might improve your credit score, trigger a benefits cliff, and mean nothing without childcare access. MobilityLens gives you six expert perspectives simultaneously — grounded in real practitioner knowledge, not generic AI opinions.
Practitioner intelligence, encoded as AI
Each AI persona is trained on structured expert interviews — real social workers, financial coaches, workforce development specialists, and policy analysts who work directly with ALICE-threshold households.
Their expertise is indexed as RAG knowledge (not generic LLM training), so when the Case Manager persona says “what works in Oakland for families at 150% of the ALICE threshold,” it’s drawing from interviews with people who have done exactly that work — with source attribution.
- Citation-grounded — every claim traces to a document
- Bias-transparent — each persona discloses its blind spots
- Debate-capable — personas challenge each other's recommendations
Multi-perspective analysis
“Should I take the $18/hr job offer?”
Run cliff calculation first — check SNAP + childcare loss.
Consider childcare costs — net take-home may drop.
Anchoring bias detected — $18 vs your current $15 feels large.
General AI gives general answers. This isn’t general.
ChatGPT is a powerful general-purpose tool. MobilityLens Personas are purpose-built for economic mobility decisions — with structural constraints that general AI cannot replicate.
| Capability | General AI | MobilityLens Personas |
|---|---|---|
| Knowledge source | General training data — no updates after cut-off date | Practitioner interviews + peer-reviewed research, RAG-indexed and current |
| Citations | None — claims cannot be independently verified | Every factual claim linked to a source document |
| Blind spots | Hidden — the AI appears confident even when uncertain | Explicitly disclosed per persona ("What I'm not seeing") |
| Location awareness | National averages only — no ZIP-level data | County-specific ALICE thresholds, HUD FMR, and local programs for your ZIP |
| Role consistency | Changes based on prompt framing — easy to manipulate | Deterministic role assignment — no prompt drift, no creative improvisation |
| Benefits cliff detection | Not computed — no access to current SNAP or Medicaid eligibility tables | Automatic — every income recommendation is checked for cliff risk before surfacing |
Meet the six perspectives
Each persona is a distinct expert lens grounded in practitioner expertise and authoritative data sources.
Financial Advisor
Credit, savings, and income strategy
Analyzes your financial capital score, identifies benefits cliff risk, and builds a sequenced savings-and-credit roadmap based on your ALICE survival budget baseline.
What I’m NOT seeing: Does not model emotional resilience or social capital leverage.
Case Manager
Social services and program navigation
Thinks like a trained social worker: surfaces the right programs, flags referral gaps, and identifies risk indicators from your situation.
What I’m NOT seeing: Does not prioritize financial optimization or market timing.
Behavioral Economist
Decision patterns and behavioral design
Identifies cognitive biases and decision traps slowing your progress, then suggests commitment devices and micro-goal structures.
What I’m NOT seeing: Does not account for structural barriers that no individual action can overcome.
Policymaker
Systemic context and structural factors
Frames your situation in local policy context: housing affordability, minimum wage, childcare subsidy eligibility, and transit investment.
What I’m NOT seeing: Does not give individual-level tactical recommendations.
Employer
Workforce perspective and hiring signals
Reads your Human Capital profile like a hiring manager: credential gaps, soft-skill signals, role-fit matches, and upskilling paths with the highest ROI.
What I’m NOT seeing: Does not model emotional wellbeing or work-life sustainability.
Community Member
Lived-experience and social capital
Grounds recommendations in neighborhood realities: who has navigated this situation, what resources actually get used, and where social capital fills institutional gaps.
What I’m NOT seeing: Does not quantify financial impact or policy constraints.
A case manager, a client, and a decision that looked obvious.
Maria is a case manager at a family services nonprofit in Memphis. Her client — a single mother of two — was offered a $19/hr job, up from $14/hr. It looked like a clear win. Here’s what the Persona Engine surfaced in under 60 seconds.
At $19/hr ($39,520/yr), this household crosses the SNAP eligibility cutoff. Annual SNAP loss: ~$4,200. Net household income gain after benefits loss: approximately $1,120 — not the expected $10,400.
The new job requires a 45-minute commute. Combined with existing childcare hours, the school pickup window conflicts with the shift end time. Three subsidized childcare programs in ZIP 38118 accept emergency enrollment.
$19 vs $14 anchors attention to the pay difference. The real comparison is net monthly take-home — which increases by only $93 in year one. Framing risk: the gain feels large, the cliff is invisible.
Tennessee's childcare subsidy expansion (2023) covers households at up to 85% SMI. This client qualifies. Enrolling before the job start date changes the net monthly gain from +$93 to +$433.
What happened next: Maria ran the benefits calculation, enrolled the client in the childcare subsidy before the job start date, and negotiated a schedule to cover the school pickup. A decision that would have cost the family $9,000/year became genuinely beneficial — instead of a hidden trap.
Sources: Tennessee DHS SNAP Eligibility Table 2024 · ALICE Shelby County 2023 · Tennessee Child Care Certificate Program guidelines.
Debate mode: when experts disagree
When you submit a decision, every persona receives the same context and generates an independent analysis. Where they agree, you get confidence. Where they disagree — you see exactly why.
The Financial Advisor and the Policymaker may disagree on whether a wage increase is advisable. The Behavioral Economist may flag an anchoring bias the Case Manager missed. Every tension is surfaced, not smoothed over.
Each persona’s “What I’m not seeing” disclosure ensures you understand the limits of every perspective before acting on any of them.
Request flow
- 1
User asks a question or presents a decision
e.g., "Should I take the $18/hr job offer?"
- 2
Context assembled: EMS scores, ZIP data, research
Profile, household composition, and location-filtered DataHub documents retrieved.
- 3
Each persona receives its fixed role definition
No prompt injection, no creative drift — deterministic role assignment.
- 4
Parallel AI calls with citation enforcement
Each persona generates its analysis grounded in its assigned data slice.
- 5
Responses surface with blind-spot disclosure
Each persona output includes a "What I'm not seeing" — limits of its perspective.
The Expert Persona Library
Behind each AI persona is a curated library of structured practitioner interviews — social workers, financial coaches, workforce development specialists, and policy analysts who have worked directly with ALICE-threshold households. Their knowledge is encoded as RAG-indexed structured knowledge, not as generic LLM training data.
RAG-indexed
Practitioner interview library
Citation-enforced
Every claim has a source
Bias-disclosed
Each persona reveals its limits
Ready to see all six perspectives?
The AI Persona Engine is available on all paid tiers. Create a free account to explore the platform.