WealthSchema vs. Howso (formerly Diveplane) — fintech-content synthetic data vs. interpretable-ML privacy synthesis
Howso (rebranded from Diveplane in 2024) is the synthetic-data vendor with one of the most distinctive technical positions in the market — they use an interpretable, instance-based ML approach (Reactor / GenAI) rather than the GAN / diffusion models common in the rest of the space. Their privacy story is constructive (the synthesis process is described as 'similar but not the same as' the source data, with explainability per-record). WealthSchema operates from a different starting point — archetype-driven generation against public-aggregate references with no customer-data input — but shares Howso's emphasis on interpretability and audit-defensibility. Both are mature, both are credible, and they appeal to overlapping but distinct audiences.
The two options
Howso (formerly Diveplane)
Synthetic-data platform built on interpretable instance-based ML. Trains on customer data, produces synthetic outputs with per-record explainability, with strong adoption in regulated industries (finance, insurance, healthcare) where audit-defensibility matters.
- Interpretable synthesis — per-record provenance is traceable, which is unusual in the synthetic-data space and meaningful for regulated industries
- Strong privacy framing — the 'similar but not the same' approach has a constructive defensibility that doesn't rely on differential-privacy mathematics
- Audit-defensibility — the explainability matches what model-risk-management programs (SR 11-7, equivalents in insurance / healthcare) want to see
- Active research output — the team publishes on synthetic-data quality and interpretability
- Mature platform with deployment in major financial institutions
- Requires customer data as input — pre-launch fintechs can't use it; teams unwilling to expose production data face friction
- Generic across regulated industries — fintech-specific edge cases (lot-level basis, IRMAA brackets, K-1 cascade, AG 49-A illustrations) aren't pre-built; the customer's training data has to contain them
- Customer-data dependency means corpus refresh requires ongoing customer-data ingest rather than annual rule-tracking refresh
- Pricing is typically enterprise-platform-deployment rather than per-corpus; the operational model differs from product-content products
Choose Howso when: (1) you have substantial customer data and the value of synthesis is partly the privacy-preserving aspect of using that data more responsibly; (2) you operate in a regulated industry where audit-defensibility through interpretability is meaningful; (3) you want a platform deployment rather than a content product; (4) your synthetic-data needs are broader than fintech-vertical content (you may also need synthesis for healthcare, insurance, or other domains).
WealthSchema
Archetype-driven synthetic financial data with public-aggregate calibration, 31 fintech-vertical bundles, no customer-data input, regulator-grade per-bundle documentation.
- Fintech-vertical depth — pre-built coverage of lot-level basis, IRMAA brackets, K-1 cascade, RMD timing, AG 49-A illustrations, QSBS, multi-state tax
- No customer-data input required — corpora are generated against public sources only
- Constructive privacy by construction — no real-person provenance to begin with
- Audit-defensible documentation per bundle — calibration sources cited, edge cases documented
- Refresh tracks regulatory changes (SECURE 2.0, AG 49-A revisions) annually, not customer-data drift
- Less interpretability per-record — Howso's instance-based approach gives per-record provenance that archetype-driven generation doesn't replicate exactly (though archetype membership is itself a form of provenance)
- Vertical (fintech) focus — not a horizontal synthetic-data platform
- Fixed bundle structure — non-bundle use cases need custom engagement
Choose WealthSchema when: (1) your synthetic-data need is fintech-vertical content rather than horizontal privacy synthesis; (2) you don't have or don't want to use customer data; (3) you need US-fintech-specific edge cases pre-built; (4) you want a content product with bundle-shaped delivery rather than a platform deployment.
Decision framework
Both products score well on audit-defensibility and constructive privacy framing — that's actually a meaningful overlap relative to the rest of the synthetic-data market. The distinction comes down to two questions:
Do you have customer data, and is using it more responsibly part of the value? If yes, Howso's privacy-preserving synthesis from real data is the more direct fit. The interpretability gives you per-record provenance for the synthesis decisions, which is meaningful when explaining to regulators or model-risk teams.
Do you need fintech-specific edge cases that aren't reliably present in customer training data? If yes, WealthSchema's archetype-driven generation is the more direct fit. Lot-level basis with wash-sale awareness, IRMAA bracket transitions, AG 49-A IUL illustrations, K-1 cascade with QBI calculations — these are pre-built rather than dependent on customer data containing them.
The two products coexist in some sophisticated teams. Howso for the customer-data privacy synthesis layer; WealthSchema for the fintech-content engine validation layer. They map to different parts of the synthetic-data problem.
Bottom line
Howso is the right answer for institutions wanting interpretable privacy-preserving synthesis from customer data with strong audit-defensibility. WealthSchema is the right answer for fintech engineering teams needing pre-built fintech-vertical content for engine validation. Both are credible; the buyer profiles overlap less than the marketing positioning suggests.
FAQ
Can Howso and WealthSchema work together?+
Yes — they map to different parts of the problem. Howso for customer-data privacy synthesis (where the goal is reducing exposure of real customer information), WealthSchema for fintech-content generation (where the goal is engine validation against named edge cases). Sophisticated teams use both.
How does Howso's interpretability compare to WealthSchema's documentation?+
Different forms of interpretability. Howso's is per-record (you can trace why a specific synthetic record looks the way it does). WealthSchema's is per-bundle (you can trace why the corpus's distributions are calibrated to the values they are). Both are useful; they answer different audit questions.
Does Howso have pre-built fintech scenarios?+
Some examples exist; they're general financial-services tabular data rather than the specialized fintech-content depth (lot-level basis, IRMAA brackets, AG 49-A illustrations) WealthSchema is built around. The two address different parts of the synthetic-data spectrum.
What about pricing?+
Howso is typically enterprise-platform-deployment pricing; WealthSchema is per-bundle. Total cost depends on team size and usage; neither is dominantly cheaper for typical buyers in their respective profiles.
Why did Diveplane rebrand to Howso?+
The rebrand happened in 2024. The underlying technology and team are the same; the new brand reflects the productization journey from research-led to commercial. References to 'Diveplane' in older documentation refer to the same company.