Sopact is a technology based social enterprise committed to helping organizations measure impact by directly involving their stakeholders.
Copyright 2015-2026 © sopact. All rights reserved.
Confidential by design — VOCA reporting as one query.
Victim services case management software is a platform that holds one confidential record per survivor across the whole service arc — crisis intake, needs and safety assessment, safety and service plan, advocacy case notes, services delivered, and outcome follow-up — so advocates, supervisors, and directors get answers without stitching spreadsheets together, under access controls aligned with VAWA and VOCA confidentiality requirements. It is also called domestic violence case management software, DV/SA case management software, or victim advocacy software. The newest generation adds intelligence to the record itself: AI reads each advocacy note on arrival, scores it against the safety plan with citations, flags escalating risk before the next contact, and turns follow-up into outcome evidence — so the software answers whether the survivor is actually safer, not just what services were logged. It is the separate, dedicated system DV and sexual assault programs use instead of a shared HMIS database, which VAWA prohibits for survivor personally identifying information.
Used by: domestic violence shelters and emergency-housing programs · sexual assault advocacy and SART/SANE-connected programs · legal advocacy (protective orders, court accompaniment) · human trafficking survivor services · child and teen victim support · multi-program victim service centers coordinating one survivor across teams.
Not because the software stopped working — because storing the survivor and logging the service, confidentially, became table stakes. Victim-services and DV systems — Osnium, EmpowerDB, Collaborate, Apricot/Bonterra ETO, ClientTrack — earned their positions honestly: they got casework out of the paper file, kept survivor data out of shared databases, and gave programs a defensible, separate system of record that ties the VOCA report out. If your problem was confidential collection — hundreds of survivors, a handful of advocates, a VOCA deadline — that generation solved it.
But the strengths hardened into weaknesses. The richest evidence those systems hold — the advocacy notes, the safety plans — dies after collection: narrative piles up in a system nobody reads across, and the analytics describe what was delivered, not whether the survivor got safer. A program lead running services across sites put the pattern plainly: they could track "outputs and dollars and volunteer hours," but "the biggest pain point is that we can't tell a cohesive story across all of it. Each thing works fine in a silo."
The cost of not reading is concrete in this vertical, because the signal that matters is a survivor whose risk is rising between contacts. As one practitioner described it, case notes end up "just sitting around in the systems… by the time they find out, you already failed." The work that decides whether a program keeps people safe has moved to the two ends the record-keeping tools never owned: reading every advocacy note on arrival, and proving the safety outcome months and years later on the same survivor.
None of this requires ripping out your incumbent — and this vertical gates hardest on trust, so the honest frame matters. Sopact provides AES-256 encryption at rest, TLS 1.3 in transit, field-level role-based access, SSO/MFA, full audit logging, and no AI training-data retention. But Sopact is not currently HIPAA-certified and offers no Business Associate Agreement, and VAWA confidentiality must be evaluated against your own program's rules — consult counsel before storing protected information. The sentence we hear on almost every call: "We're not gonna leave our system, but we're open to an AND." Keep the confidential system of record; add the layer that reads what it produces. (Comparing named platforms outright? Start with case management software.)
The stake, stated honestly: funders have already changed the question from "how many survivors did you serve" to "did their safety and stability improve, and can you show it." If you are signing a configuration build today, ask which question it will answer when it goes live.
Case intelligence is reliable answers from your survivor data — in minutes, not months — without ever pooling identifiable data into a shared database. Everything a survivor touches is treated as data: the confidential intake, the safety assessment, the validated screen, the safety plan, every advocacy note, the 90-day and year-later follow-up. All of it lands on one persistent survivor record — one ID, one story, under your own field-level access controls — so the same person looks like the same person across shelter, legal advocacy, and counseling, without double-counting in the funder report.
The part that changes daily work is the Assistant. Caseload analysis, screen scoring, and open-text note analysis are unified into one chat-based function: ask a question, get a defensible answer with citations to the underlying records. No prompt engineering, no dashboard hunting, no waiting for the one person who knows where the export lives. A program is never one user — advocates, supervisors, the director, funders, and survivors themselves all need different views of the same record — and a chat interface empowers each of them at their own permission level. Sensitive fields can be excluded from AI processing entirely, and analysis can run on anonymized IDs.
When the analysis is done, it does not die in the chat: create shareable reports tailored to each audience — the supervisor's caseload view, the VOCA PMT report, the de-identified board summary — from the same underlying answer, each number traceable to the source advocacy note.
One proof point from the field. Open Play Foundation ran programs the way most funded organizations do — intake forms, follow-up reflections, stacks of narrative that never made it past the spreadsheet. When that work moved onto Sopact, the record could finally read itself: "Those statistics that we're now running on Sopact immediately showed me there's something significantly wrong … things like that, we would never have been able to do in the past." — Marco Botha, CEO, Open Play Foundation. A system of record tells you the service was delivered. An intelligent record tells you a survivor's safety is slipping in time to act.
The honest way to evaluate victim services case management software is against the survivor's arc, not the feature list. Below is the full cycle — six stages, each with what the software should do, the exact prompt to use, and what to expect back. Every prompt is copy-paste; the placeholders in brackets are yours to fill. Confidentiality is not a stage — it is the constraint on every stage, so field-level access and sensitive-field exclusion from AI are noted throughout.
Intake is where confidentiality is won or lost. Instead of free-text you will pay an advocate to decode later, the form is designed so every narrative field maps to your framework, and every survivor gets a persistent unique ID that follows them across programs — with a minimal-PII, anonymized-ID option for programs that need it. Eligibility, consent, and survivor-controlled release of information are captured up front; AI drafts the intake form from documents you already have.
Build a confidential victim-services intake form from this program description: [PROGRAM URL OR DOCUMENT]. Create minimal-PII structured fields, consent and survivor-controlled release-of-information fields, and narrative fields for presenting needs and immediate safety concerns mapped to our VOCA outcome rubric. Assign a persistent survivor ID and support an anonymized-ID option. Flag any field that collects more identifying information than the service requires.
Expected output. A ready-to-edit confidential intake form: minimal-PII fields, consent and release capture, mapped narrative prompts, and a persistent survivor ID assigned at first contact — with an anonymized-ID path.
Tips for reliable output. Assign the survivor ID at first contact, not at enrollment. Decide which fields are excluded from AI before you collect anything — the exclusion is easiest to enforce at form design.
The baseline is the reference every later wave is compared against. Validated screens (a lethality/danger assessment, or your own indicator) and the intake narrative land on the same confidential record, and the assessment is read the moment it arrives — needs, risk factors, and protective factors extracted and cited, not left in a folder until something goes wrong.
From this intake and safety assessment, extract the survivor's baseline needs, risk factors, and protective factors, each with the exact source sentence. Score the danger/lethality screen included, flag any immediate-safety or escalation language for advocate review, and note where the assessment is incomplete. Do not infer a diagnosis — report only what the text supports. Treat all fields marked sensitive as excluded from output.
Expected output. A structured baseline with per-item evidence, a scored danger screen, and a flagged list of immediate-risk language routed to a named advocate.
Tips for reliable output. Lock the baseline before services begin — a baseline captured on day one, even on a handful of survivors, proves the loop works before anything scales.
Every survivor gets a safety and service plan built from the assessment and mapped to your framework — goals that are observable, timelines that are real, and the outcome each service is meant to move. The plan becomes the thing advocacy notes are later read against, and access is limited to the assigned advocate.
Draft a safety and service plan from this assessment: [ASSESSMENT]. Map each identified need and risk factor to a measurable safety or stability goal, a service or referral, and the outcome indicator it should move, aligned to our VOCA outcome rubric. Write goals as observable statements an advocate can evidence, and flag any need with no service currently available.
Expected output. A safety and service plan with measurable goals, mapped services, outcome indicators, and a gap list where needs have no matching service.
Tips for reliable output. Name the outcome for every goal. A plan that can't say what safety looks like can't be evaluated later.
This is the stage record-keeping software cannot do. Every advocacy note is read as it lands, coded against the safety plan, with risk signals — a rising danger indicator, a missed safety-plan review, escalation language, disengagement — surfaced the week they appear instead of at a month-end sample. The narrative stays with the advocate; the structure is generated and tied back to the source sentence, and sensitive fields never enter the AI call.
Read this batch of advocacy notes: [NOTE BATCH]. For each survivor, summarize progress against the safety plan with citations, code the note against our outcome indicators, and flag risk signals — rising danger indicators, missed safety-plan reviews, escalation or disengagement language — with the exact source sentence. Exclude any field marked sensitive. Use the same method as last month so results are comparable.
Expected output. Per-survivor progress summaries with citations, coded outcome evidence, and a risk-flag list with sources — the day notes are written, not six weeks later.
Tips for reliable output. Route every risk flag to a named advocate with a deadline. A flag nobody owns is a survivor who slipped through.
Closure is not the end of the record. The 90-day, one-year, and longer follow-ups land on the same survivor ID as the intake, so the question every funder asks — is the survivor actually safer and more stable — has a reproducible answer instead of a year-end reconstruction. A survivor-controlled link lets them fill the one missing field without creating a duplicate record.
Compare the baseline assessment to the [90-day / 1-year] follow-up across [COHORT / PROGRAM]: which safety and stability outcomes moved, by how much, and with what confidence? Show change per indicator, note where the sample is too small to conclude, and pair every number with a representative advocacy-note quote. Treat this as change over time, not attribution.
Expected output. A baseline-to-follow-up outcome analysis with honest confidence bounds and a narrative quote behind each number — the longitudinal view a persistent ID makes possible.
Tips for reliable output. Capture follow-up channels and expectations at intake, respecting the survivor's safety and contact preferences. The longitudinal horizon is what separates an exit survey from a safety outcome.
Reports are questions, not formats. From the same accumulating survivor record, the caseload report, the supervisor view, the VOCA PMT, the VAWA/OVW submission, and the state coalition report are each one query — with the supporting advocacy note two clicks away — instead of a multi-week reassembly across intake, services, and follow-up systems. De-identify for the board without a separate export.
Aggregate this program's survivor records into a [VOCA PMT / coalition] outcome report: outcomes achieved against targets, coded advocacy-note themes ranked by frequency with representative quotes, demographic distribution, and survivors flagged as missing a required follow-up. Cite the source record for every number and quote. Produce one de-identified version for the board and one for the funder.
Expected output. A funder-ready outcome report generated as a query, every figure citing its source record — plus the "missing" list surfaced before the deadline asks.
Tips for reliable output. Lock the data dictionary before the first reporting cycle and version every change — comparability across years is the entire value. If your outcome framework needs an external anchor, align it to the funder's own guidance at OVC (ovc.ojp.gov) so metrics map to what VOCA asks.
Beyond table stakes — confidential intake, caseload views, services logging, VAWA-aligned access — four criteria actually separate tools: time to first live cycle (days vs. a quarter), whether AI reads advocacy notes on arrival or a supervisor still samples them by hand, whether configuration is plain-English or a consultant engagement, and whether the platform can prove safety and stability outcomes rather than just count services. Ask every vendor to show the VOCA outcome report on real data, not a slide — and to state their confidentiality and HIPAA posture in writing.
Build an evaluation matrix for victim services case management software with technical and program criteria weighted 50/50. Technical: field-level role-based access, VAWA confidentiality model, anonymized-ID and sensitive-field-exclusion support, encryption, audit logging, data export and exit rights, and HIPAA posture stated in writing. Program: AI advocacy-note reading with citations, one survivor ID across programs, longitudinal safety-outcome tracking, VOCA PMT report generation. Score vendors [VENDOR LIST] on each criterion with evidence required, not vendor claims.
Propose a 30-day pilot plan to evaluate victim services case management software: one program, one cohort of current survivors, tested end to end (confidential intake → safety assessment → safety plan → advocacy-note reading → VOCA outcome report), with numeric success thresholds and rollback criteria if the pilot fails.
A note on scope while you evaluate: the same spine adapts by vertical with a different intake and funder report — human services case management software, social work case management software, and nonprofit case management software each cover their fit directly. For the note layer itself, see case notes software; for advocate capacity, caseload management software. To coordinate survivors and partners across a network, see stakeholder intelligence.
The stages above are the argument; the Academy articles are the practice — each a hands-on companion for one workflow, written to run on your own data.
Honest boundaries, because the fastest way to a failed implementation is buying the wrong category.
Not a shared HMIS, a CRM, or an EHR. A shared HMIS pools identifiable client data — which VAWA prohibits for DV/SA survivors; victim services software is the separate, dedicated record that meets that requirement. A CRM (Salesforce, Blackbaud) tracks donors; an EHR (Epic, Cerner) tracks clinical encounters and billing. Victim services software tracks survivors through a confidential service-and-outcome lifecycle, with the advocacy note and outcome evidence at the center.
Not the 24/7 crisis hotline. Real-time crisis dispatch is not what Sopact is built for. Sopact is for the advocacy-and-outcome arc that follows — intake, services, follow-up — not the live hotline itself.
Not for every compliance regime. Sopact provides AES-256 encryption, TLS 1.3, field-level role-based access, SSO/MFA, and full audit logging, with AI under enterprise SLAs and no training-data retention — but Sopact is not currently HIPAA-certified or covered by a Business Associate Agreement. VAWA confidentiality and any HIPAA, FERPA, or 42 CFR Part 2 obligations must be evaluated against your own compliance program; confirm scope in writing, and consult counsel, before storing protected information.
Victim services case management software is a platform that holds one confidential record per survivor across the service arc — crisis intake, needs and safety assessment, safety plan, advocacy notes, services delivered, and outcome follow-up — so advocates and directors get answers without a spreadsheet merge, under access controls aligned with VAWA confidentiality. Also called domestic violence or DV/SA case management software. The newest generation adds AI that reads each advocacy note on arrival and turns follow-up into cited safety-outcome evidence. It is the separate, dedicated system DV and sexual assault programs use instead of a shared HMIS database.
Yes. Outcome-focused victim services software should produce the reports a funder asks for as a single query off the same records: the VOCA Performance Measurement Tool (PMT), VAWA/OVW reports, and state coalition or STOP-grant reports. The difference is that each number carries the supporting advocacy note, so the report is defensible rather than a year-end reconstruction. Aligning your rubric to OVC guidance keeps the metrics mapped to what VOCA asks.
VAWA confidentiality provisions restrict victim-service providers from entering survivor personally identifying information into shared databases such as HMIS, and require survivor-controlled release of information. Sopact supports this with field-level role-based access, AES-256 encryption at rest, TLS 1.3 in transit, SSO/MFA, full audit logging, white-label survivor-facing forms, and a record that is yours alone — not a shared comparable database. Sensitive fields can be excluded from AI processing entirely. Evaluate the model against your program's specific VAWA rules, and consult counsel before implementation.
Sopact is not currently HIPAA-certified and does not offer a Business Associate Agreement. It provides AES-256 encryption, TLS 1.3, field-level access, SSO/MFA, and full audit logging with no AI training-data retention — but if any part of your workflow handles Protected Health Information under HIPAA, confirm in writing whether that workflow falls inside or outside the HIPAA boundary before implementation, and evaluate these controls against your compliance program. Treat HIPAA scope as gating, not an afterthought.
Yes. Intake supports a minimal-PII, anonymized-ID option, and any field marked sensitive can be excluded from AI processing entirely — analysis runs on the anonymized ID and the non-sensitive fields. This is how a program keeps the reading layer useful without ever exposing identifying survivor data to the model or pooling it into a shared database.
Yes for both — they are the core case. A DV shelter runs confidential intake, safety planning, shelter stay, and stable-housing outcomes reported to VOCA without entering survivor PII into a shared HMIS. A sexual assault program runs hospital accompaniment, advocacy, and follow-up across a sensitive arc where the qualitative note carries the evidence and the survivor controls what is released. Both keep continuity from crisis intake to safety outcome on one confidential record. Multi-program centers running shelter, legal advocacy, and counseling on the same survivor use one confidential ID so the funder report stops double-counting.
Two months, one contained use case — one program, one confidential intake form, one cohort of survivors you already serve. You bring last year's advocacy notes; the pilot shows you the coded, cited version of your own caseload, ending with a demonstrated VOCA outcome export. If the outcome answers aren't defensible in front of your board or funder, don't continue. Scope a 2-month pilot →