CENTRL vs Termina (2026)
The decision is what you believe: testimony or telemetry. CENTRL processes testimony, questionnaires a manager answers about itself, pre populated from an answer library, verified, scored, benchmarked at question level and ranked by machine assigned severity, across a platform serving pension funds, sovereign funds, consultants, managers and custody teams on one spine. Termina processes telemetry, the target's own transaction records at a scale no deal team reads by hand, more than 120 million transactions in one published case, scanned within an hour of notification and returned inside 24 hours as a quality of growth read, customer acquisition cost, payback period, retention and growth composition benchmarked against prior deals. Each method inherits its material's weakness. Testimony can be curated, and CENTRL's assistant now helps write it on the responding side of the same platform that scores it on the allocating side, with what crosses between the two sides unanswered. Telemetry cannot be curated, and it can be misread: a company outside the distribution of deals the platform has previously observed is measured against norms that may not apply, nothing signals when a comparison cohort is thin, and a scan showing deterioration is what a principal takes to a board. The assurance surfaces are this pocket's two odd Bs. Termina maintains a real trust centre with a responsible disclosure programme and names no framework behind it. CENTRL publishes algorithm level controls and quarterly key rotation while the only attestation on its security page belongs to its cloud provider, cited to a standard retired nine years ago. Both draw power from accumulation, answer libraries at one, a pool of thousands of observed deals at the other, with contribution, thresholds and opt outs unstated at both, and Termina adds a wrinkle it names itself, an operating entity asserting separation from affiliated investors without describing the mechanism.
- You are managing a relationship, not running a scan. Questionnaire construction, manager portals, remediation tracking, benchmarking and time series analytics across a standing programme of oversight, on both sides of the exchange.
- Your institution types are the breadth. Pension funds, sovereign funds, endowments, consultants, managers, service providers and custody teams overseeing correspondent banks, one platform across all of them.
- Agents arrive governed. The agentic layer ships with curated workflows, prebuilt integrations and permissioning controls determining what an agent may reach.
- You want evidence the target cannot curate. Transaction level records scanned at deal timescales, more than 120 million transactions in one published case, begun within an hour of notification and returned as a quality of growth read inside a day.
- The humility is published. Every page states the platform does not substitute the subscriber's own analysis, and delivery is described as models and operators together, people disclosed rather than implied.
- The methodology surface is real. Platform documentation, a glossary defining every metric, a standing trust centre and a responsible disclosure programme.
This comparison is published by AI FinTech Index, an independent research platform that publishes independent ratings of AI vendors for financial services. CENTRL and Termina are each graded against the same capability taxonomy, from each vendor's own public materials and the regulatory record, under the AI FinTech Index verification standard. No vendor pays for placement, and no vendor has reviewed this page. How this evidence is graded
Plain facts
| CENTRL | Termina | |
|---|---|---|
| Primary category | Capital Markets & Research AI | Capital Markets & Research AI |
| Founded | 2015 | 2023 |
| Headquarters | Mountain View, California, United States | Menlo Park, California, United States |
| Website | www.centrl.ai | www.termina.ai |
Side by Side
| Axis | C CENTRL |
T Termina |
|---|---|---|
| AI Centrality | ||
| Autonomy and Oversight Model | ||
| Model Risk Management and Transparency | ||
| Operational and Outcome Evidence | ||
| AI Safety and Data Stewardship | ||
| GLBA and Data Privacy Posture | ||
| Security Certifications and Trust Center | ||
| Regulatory Status and Licensure | ||
| AI Governance and Bias Disclosure | ||
| AI Liability and Recourse | ||
| Model Supply Chain Disclosure | ||
| Core Systems and Integration Depth | ||
| Deployment Model and Data Residency | ||
| Commercial Transparency | ||
| Institution and Segment Coverage |
The short version of each
CENTRL
CENTRL processes testimony, questionnaires a manager answers about itself, pre populated from an answer library, verified, scored, benchmarked at question level and ranked by machine assigned severity, across a platform serving pension funds, sovereign funds, consultants, managers and custody teams on one spine. The AI FinTech Index records the method's inherited weakness as the diligence question: testimony can be curated, and the platform's assistant now helps write it on the responding side of the same system that scores it on the allocating side, with what crosses between the two sides unanswered. The index records the evidence and assurance gaps: sixteen quantified case studies with no names, accuracy claimed against general purpose tools and measured nowhere on a tool that pre fills fiduciary answers, and a security page whose only attestation belongs to its cloud provider under a standard retired nine years ago.
Source: AI FinTech Index, 2026
Termina
Termina processes telemetry, the target's own transaction records at a scale no deal team reads by hand, more than 120 million transactions in one published case, scanned within an hour of notification and returned inside 24 hours as a quality of growth read, customer acquisition cost, payback, retention and growth composition benchmarked against prior deals. The AI FinTech Index records the method's own weakness as the black box: a target is measured against cohorts assembled from whatever deals the platform previously observed, nothing describes cohort composition, nothing signals when a comparison set is thin, and the published cases are anonymous self reported anecdotes rather than a hit rate, so a scan a principal takes to a board arrives without a published probability of being wrong. The index also records a real trust centre with no framework named behind it, accumulation terms unstated for the pool of prior deals, and separation from affiliated investors asserted without a described mechanism.
Source: AI FinTech Index, 2026
Common questions
What is the basic choice between CENTRL and Termina?
Testimony or telemetry. CENTRL processes what a manager says about itself, questionnaires scored and ranked by machine severity, while Termina processes what the target's transaction records show, more than 120 million transactions scanned in one published case. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified August 23, 2026. No vendor pays for placement.
What weakness does each method inherit?
Testimony can be curated, and CENTRL's assistant now helps write responses on the same platform that scores them, with the cross side boundary unanswered. Telemetry cannot be curated but can be misread, since Termina's benchmarks come from cohorts nobody outside can examine, with no signal when a comparison set is thin. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified August 23, 2026. No vendor pays for placement.
How do the assurance surfaces compare?
The AI FinTech Index records them as the pocket's two odd Bs: Termina runs a real trust centre with responsible disclosure and names no framework behind it, while CENTRL's security page carries only its cloud provider's attestation under a standard retired nine years ago.
What terms are unstated at both?
Contribution terms for the accumulation each runs on, answer libraries at CENTRL, the pool of observed deals at Termina, plus Termina's asserted but unmechanised separation from affiliated investors. Graded by AI FinTech Index against the same capability axes from each vendor's own published materials, verified August 23, 2026. No vendor pays for placement.
Related comparisons
Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Wealth & Advisory AI page.
Termina's benchmark is its black box: a target is measured against cohorts assembled from whatever deals the platform previously observed, nothing describes cohort composition or comparability, nothing signals when a comparison set is thin, and the published cases are single self reported anecdotes rather than a hit rate, so a scan that a principal takes to a board arrives without a published probability of being wrong.
Its accumulation needs terms, the pool of prior deals is the selling point with contribution thresholds and opt outs unstated, and the corporate separation it asserts from affiliated investors is stated rather than mechanised, with models and operators meaning people see a target's data under controls nobody has published, and nothing on the fate of 120 million records when a bidder loses.
CENTRL's parallel items: accuracy claimed against general purpose tools and measured nowhere on a tool that pre fills fiduciary answers, the cross side boundary between scoring managers and helping them respond unanswered, severity assignment ungoverned in public material, and a security page whose only attestation belongs to its cloud provider under a standard retired nine years ago. Neither publishes a price, Termina's turnkey report catalogue implying one exists, a named framework certification, or a residency commitment.