Agent IQ

Models carry a substantial share of the work, with self service handling more than 80 percent of incoming conversations, real time insight generation on what customers are discussing, conversation analytics driving proactive outreach, translation across more than 100 languages, and an orchestration framework integration extending the agentic layer.

Against that, the platform underneath is a relationship banking communication system spanning asynchronous messaging, video, voice, co browsing and the banker assignment mechanism, and that would remain valuable without the models because the company's own argument is that persistent human relationships beat transactional bots. This sits where the product intends to sit, between the two.

The customer, not the institution and not the vendor, controls the handover, which is the strongest position on this axis in the conversational lane. A person can speak with a human agent at any point during an engagement, unconditionally, rather than after satisfying an escalation rule, and the banker carousel goes further by letting them choose which human, browsing brief biographies and photographs to select a personal banker.

That inverts the usual design, making the person more visible rather than less. The company states the principle directly, that it augments human bankers rather than replacing them, and positions persistent relationships against transactional bots. High automation coexists with it at more than 80 percent of conversations, and the escape hatch is unconditional.

The published figures measure containment and cost rather than correctness. More than 80 percent of conversations handled by self service says how many the system kept, not how many it answered correctly, and a 29 percent call centre reduction measures displaced volume, both of which can rise while accuracy falls.

The unconditional human escape hatch limits the downside in a way a contained conversation elsewhere would not, since a customer who is not helped can simply leave the automation. No accuracy, resolution quality or error analysis is published.

The distribution arrangement is the strongest in this lane: a listed digital banking platform provider made the company a strategic alliance partner, empowering its own sales organisation to resell the product to its bank and credit union customers, who can contract directly with that provider when purchasing, after many successful implementations through its innovation marketplace. That is an incumbent selling a startup's product as its own.

A university credit union is named as a customer, and further partnerships cover a digital banking provider, an identity verification specialist and a major technology company's agent orchestration framework. Total funding reaches 37 million dollars across four rounds, including strategic capital from a community bank focused fund. Outcomes are quantified at 29 percent lower call centre volume and more than 80 percent of conversations handled by self service, alongside an industry award in 2024.

No data boundary statement was located. Self service quality improves with exposure to more banking conversations, and the platform serves many community institutions competing for the same customers in overlapping regions, so what is learned at one institution has direct value to another.

Nothing states whether conversation content or resolution patterns inform models serving other customers, whether an institution can decline to contribute, or what happens to conversation histories when a contract ends.

No data protection agreement, retention schedule, subprocessor list or deletion commitment was located. The platform holds ongoing conversation histories between named bankers and identified account holders, which is the point of a persistent relationship model, and now also processes identity verification through an integrated provider. Nothing published states how long conversation records persist, whether they follow a customer when a banker leaves, or how the identity data captured at verification is handled.

Security is named as a design focus for community institutions and no attestation, certification, trust centre or enumerated framework was located. Eleven years of operation, embedding inside a listed banking platform and a reselling arrangement with that provider all imply assessment has been passed repeatedly at institutional standard, and none of the resulting assurance is published.

No supervisor, statute or rule is named. Automated customer communication at regulated institutions engages expectations around fair treatment, complaint handling and record keeping of customer interactions, and the recent addition of identity verification brings customer identification requirements into scope. None of that framework is identified, and no accessibility or language access guidance is cited despite translation being a headline capability.

Two mechanisms address the exclusion pattern recorded repeatedly across this index, that automated channels serve worst the people most dependent on them. Real time translation across more than 100 languages tackles the language barrier at far greater breadth than the dedicated vendors in this space, and the unconditional route to a human means anyone the automation fails, for any reason including accent, phrasing, disability or complexity, reaches a person immediately rather than after exhausting a decision tree.

Together those are structural rather than aspirational. What is absent is measurement: no translation quality by language, no analysis of who escalates and why, and no outcome data across customer groups.

No commercial guarantee was located, and the customer's position is unusually strong because recourse is a product feature rather than a policy. A person can leave the automation for a human at any moment without meeting a condition, and can choose which named banker to speak to, so the remedy for a system that misunderstands them is immediate and in their own hands. That is more than any other conversational vendor in this index offers the end customer. What is missing is the institutional side: nothing describes correction where the platform gives wrong information, or what a bank is owed when it does.

Two technical dependencies are named and dated, an identity verification provider integrated into the platform in late 2025 and a major technology company's agent orchestration framework adopted in early 2025, so a buyer can see where parts of the capability originate and assess that fourth party exposure. Distribution partners are similarly identified. What is not disclosed is the model layer itself, with no provider named for the conversational or translation components, and no subprocessor list or hosting arrangement located.

Four named integrations cover the layers that matter for this buyer. The platform embeds inside a major listed digital banking platform, having come through its innovation marketplace before becoming a strategic alliance partner and a reselling arrangement, which places it where community institutions already run their digital channel.

A second digital banking provider is named, an identity verification specialist was integrated in 2025 to bring document based verification into the conversation, and a major technology company's agent orchestration framework was adopted in the same period. Deployment in as little as four weeks with a light resource lift confirms the integration is genuinely shallow to install.

No hosting provider, region selection, residency commitment or private deployment option was located. Exposure is domestic and therefore simpler than for international vendors, and the platform holds ongoing customer conversation records on behalf of supervised institutions, so where that content resides is a question their vendor management process will raise.

No pricing, packaging or basis of charge was located. Two adoption signals are published: deployment in as little as four weeks with a light resource lift, which matters for institutions without implementation teams, and the option to contract through an established digital banking provider rather than directly, which simplifies procurement for that provider's existing customers. Neither states cost, and nothing indicates whether charge scales with account holders, conversations, bankers or as a platform fee.

Buyers are community banks and credit unions, with the platform serving retail banking, business and commercial banking and wealth management inside them, so one deployment reaches several lines. Channel coverage is genuinely complete, spanning asynchronous messaging, live chat, video, voice, co browsing, text message, social media and self service, which matters because a relationship that persists across channels is the entire proposition. Real time translation across more than 100 languages extends reach considerably. The limits are geographic and structural: this is a United States product for smaller institutions specifically.