Kore.ai

The removal test leaves nothing. Kore.ai has been a conversational artificial intelligence company since its first product, and what it sells is an agent platform: the thing being bought is the capability to design, deploy, govern and optimise agents. AI for Banking is a pre-built set of banking agents rather than a workflow product with models attached, and there is no prior non model product underneath that would survive.

This is the assistant native shape rather than the channel platform shape, and it sits a full grade above the conversation platforms in this category that added models to an existing communications spine.

Immutable audit trails are stated to apply to every agent action, which is the strongest forensic position in this category and lets an institution reconstruct exactly what an agent did. Observability tooling surfaces agent performance, exception trends and risk indicators, developers can extend guardrails without disrupting the surrounding systems, and agents are described as escalating exceptions when human judgement is required.

Held at the middle grade on the same reasoning applied to other vendors in this session: an audit trail is a reconstruction capability after the fact rather than a control in the execution path. What triggers an escalation is not described, no confidence threshold or approval checkpoint is named, and the guardrails are configurable by the customer rather than specified by the vendor, so the strength of the control depends on the buyer rather than the product.

Governance is delivered as platform capability, which is what earns this grade: immutable audit trails on every agent action, observability into agent performance and exception trends, risk indicators surfaced to the operator, and tenant isolation separating one institution's agents from another's. An institution's model risk function could review what happened.

What it could not review is how well the agents work, because no accuracy figure, containment rate, escalation rate, error rate or validation summary is published. The absence worth naming specifically is the artificial intelligence management system standard: this vendor holds federal security authorisation, payment card certification, a service organisation control report and ISO 27001, and no AI specific certification at all, while two far smaller vendors elsewhere in this index hold one.

The gap between this vendor's scale and what it publishes about financial services outcomes is the single most conspicuous thing about its disclosure. Named customers with quantified results exist but sit outside banking operations: a European banking group describes taking human resources automation from a single region chatbot in 2020 to a multi jurisdiction strategy by 2025, and a pharmaceutical company reports a technology service desk handling 70 percent of requests.

The financial services customer stories are anonymised as a global bank and a major bank. One quantified banking figure appears without an institution attached, an executive describing 15 to 20 minutes returned to each financial adviser daily. Analyst evaluation is referenced for two conversational artificial intelligence categories without a stated placement, which is weaker than naming one. A vendor of this size publishing anonymous banking case studies is making a choice, and it is what keeps this off the top grade.

Tenant isolation is stated as a platform property, which answers the cross customer question architecturally where most vendors in this index leave it open, and real time tokenisation limits what sensitive content reaches the model layer at all. On premises and sovereign deployment give an institution the strongest available form of the same assurance. Held off the top grade for two reasons.

Nothing states whether customer interactions inform model or product improvement, which is the question tenant isolation does not answer. And the vendor operates an agent marketplace offering pre-built agents, templates and integrations, which introduces components an institution did not build into conversations with its own customers, with nothing published about what those components may access.

Real time tokenisation of personally identifying information is applied to every agent action, which is privacy by construction rather than privacy by policy and is the distinction this index has treated as genuinely gradeable. Tenant isolation, on premises and sovereign deployment options, stated European data protection compliance and health information privacy alignment all add to it, and an institution that needs conversation data never to leave its own estate has a supported path to that.

Held off the top grade because the public record stops at the assertions: no data processing agreement, subprocessor list or retention schedule was located outside the gated trust centre, and nothing describes what conversation content is retained or for how long once tokenisation has run.

The deepest security credential stack in this index. Kore.ai holds a service organisation control type two report, ISO 27001 and payment card industry data security standard certification, carries a Federal Risk and Authorization Management Program moderate authorisation, and states alignment with health information privacy rules, the health information trust framework and European data protection law.

A live trust centre operates at its own subdomain on a continuously monitored platform and is described as covering security practices, compliance programmes and governance. The federal authorisation is the standout: it is a government run assessment against a defined control baseline, and only a handful of vendors anywhere in this index have cleared one. Selling into government and healthcare alongside banking forces a bar that financial services only vendors never encounter, and the disclosure here reflects it.

Kore.ai is a technology supplier and holds no financial licence, which is the correct posture and carries no penalty. What lifts it above the index norm is genuine formal admission rather than a stated position: a Federal Risk and Authorization Management Program moderate authorisation is a government operated assessment with a defined baseline and a sponsoring agency, and payment card industry certification is a second formal scheme.

Under the standing index ruling that enrolment in a programme earns this grade and only a supervised regulator test of the artificial intelligence product itself earns the top one, this sits firmly at the top of the middle grade. No financial supervisor has tested the banking agents.

No fairness testing, differential outcome monitoring or impact assessment was located. The exposure is concrete rather than theoretical because of where the agents operate: loan servicing, fraud resolution and card blocking are consumer facing decisions with material consequences, and a fraud resolution agent that treats some customers' disputes differently from others produces exactly the kind of uneven outcome this axis exists to detect.

Multi channel operation including interactive voice response adds an accessibility dimension, since voice recognition performance varies by accent, speech pattern and assistive technology use, and nothing addresses per channel or per population performance.

No error rate, remediation commitment, liability position or correction path is published. The agents operate at the point where a customer is told whether a disputed transaction will be reversed, whether a card can be replaced or what a loan account currently requires, so a wrong answer lands directly on a consumer who has no relationship with the vendor.

The marketplace adds the same unanswered question raised by other agent marketplaces in this index: when a pre-built or partner supplied agent gives a bank's customer wrong information, nothing states where responsibility sits between the component's author, the platform distributing it and the institution whose name appears on the conversation.

No model provider, family or version is named for the agents the vendor itself supplies, and the platform is positioned as model agnostic with developers able to integrate advanced models of their own choosing. Partial credit is due for one disclosure most of this index omits: the banking application is stated to run on a named cloud artificial intelligence platform, which tells a buyer where inference happens even though it does not say which models perform it. Deployment options including on premises give an institution a route to control the question itself, which is a different thing from the vendor answering it.

More than 300 integrations reach core banking systems, payment platforms, customer relationship management, policy administration, risk engines and data warehouses, and more than 40 voice and digital channels carry the conversation itself.

The distribution position adds to it materially: the banking application is packaged and sold through Microsoft's marketplace and runs on Azure AI Foundry alongside Dynamics 365 Contact Center, Teams and Copilot, so an institution already standardised on that stack adopts it without a separate integration programme.

Named enterprise connections extend beyond financial services to Salesforce, HubSpot, Jira and GitHub, which matters because agents operating in a bank touch internal engineering and service tooling as well as the core.

The strongest showing on an axis where 271 of the vendors in this index sit at the lowest grade. Four deployment models are offered rather than one: public cloud, sovereign regions, private cloud and on premises, with data residency stated by region. That means a bank in a jurisdiction with localisation requirements, a government client and a commercial customer wanting managed cloud can each be served without the vendor changing product.

On premises in particular is nearly extinct among the agent platforms in this index and it is the option that lets an institution keep member and customer conversation data entirely inside its own estate. Tenant isolation is stated separately as a platform property.

No pricing, tier structure, billing basis or indicative range is published on the vendor's own material, and the route to a number is a call with a sales team. This is the index norm and is measured against Sumsub, which publishes per verification rates on a public page.

The banking application is listed on a third party cloud marketplace, which is a channel where transactable pricing is conventional and would represent a straightforward route to disclosure, and nothing indicating a published rate was located there either.

Banking coverage spans retail, commercial and wealth management rather than one line, and the use cases named run across the institution: account support, card block and replacement, fraud resolution, loan servicing and internal employee support. Both banks and credit unions are served, and the footprint is global rather than domestic, with a named European banking group among the referenced deployments. Employee facing and customer facing workflows are both covered, which is unusual, since most vendors in this category pick one side.