Boost.ai

The removal test leaves nothing behind. Boost.ai has built conversational artificial intelligence since it was founded in 2016 and has never sold a product underneath it: the virtual agent is what an institution buys, and understanding what a customer asked and producing the response is done by models throughout. The no code builder configures the agents rather than constituting a separate product. This is the assistant native shape rather than the channel platform shape, and it sits a grade above the conversation platforms in this category that added models to an existing communications spine.

A named mechanism exists and its internals are not described, which is the middle rung of this index's ladder. The Trust Layer is presented as the control surface around the virtual agents, with guardrails intended to prevent manipulation, keep responses within approved bounds and align behaviour to sector regulation.

The reference deployment supports it in practice, since DNB automates a stated 20 percent of customer service and routes the remainder to people, and its executive frames the objective as an equilibrium between human and machine rather than maximal automation. What is missing is the specification: no confidence threshold, no described trigger for handover to a person, no approval checkpoint and no account of what the guardrails evaluate before a response reaches a customer.

The Trust Layer is a safety control rather than a validation position, and this axis asks a different question. No accuracy figure, containment rate, escalation rate, error rate or benchmark is published for the agents, no artificial intelligence management system certification is held, and no validation documentation is offered to the institution. The security and privacy certifications are genuine and are credited on their own axes, but neither attests to how the models behave. An institution running virtual agents across 150 million conversations a year has a measurement surface that would support a published quality figure, and none exists.

The cleanest evidence showing of this session because it has the join that most vendors lack. Named customers include Nordea, DNB and Telenor, and DNB, one of the largest banks in the Nordics, is the subject of a published case study reporting 20 percent of customer service automated through the platform, with a named executive quoted on the balance between human and machine handling.

Independent evaluation is specific rather than gestured at: the company was named a Leader in the 2025 Gartner Magic Quadrant for Conversational AI Platforms, with a stated placement rather than mere participation, having previously appeared in the 2022 enterprise edition. Published scale is consistent across sources at more than 600 live agents across more than 450 organisations handling more than 150 million conversations annually. Nordic Capital has held the company since 2021, which puts a party with capital at risk behind the diligence.

The company names its safety architecture and, unusually, names the failure modes it is built against. The Trust Layer is described as fortifying virtual agents with checks and balances, with guardrails positioned as tamperproof and resistant to jailbreak attempts, and the published explanation states plainly that language models can be unpredictable, can generate biased responses, can hallucinate and can be exploited through prompt injection.

Vendors in this index rarely write down what can go wrong with their own product. Held off the top grade because the boundary questions are unaddressed: nothing states whether conversations from one of the 450 organisations inform models serving another, and nothing describes whether customer interactions are used for model improvement.

ISO 27701 certification is the distinguishing fact and very little in this index holds it. It is an audited privacy information management system rather than a policy statement, held since June 2022, scoped to the whole group including subsidiaries, with 46 of 49 controls implemented and that ratio published.

The privacy policy names the specific regimes it addresses, European data protection law, California privacy law as amended, and United Kingdom data protection law, and the legal entity and its company registration number are printed.

Held off the top grade because the operational documents a buyer would need are not public: no data processing agreement, subprocessor list or retention schedule was located outside the gated trust centre, and nothing describes how long conversation transcripts are held or what is done with them.

The most complete certification disclosure found anywhere in this index, and the reason is not the certificates but what accompanies them. Boost.ai holds ISO 27001 since May 2021 and ISO 27701 since June 2022, both scoped to the whole company including subsidiaries rather than to a product line, and it publishes the implementation ratio for each: 113 of 114 security controls and 46 of 49 privacy controls. Almost every vendor in this index that holds a certification publishes the badge.

This one publishes how complete the implementation is, including where it is not, which tells a buyer something a certificate alone never does. A trust centre is operated for the full compliance overview, and alignment with open web application security guidance is stated.

Recorded as unconfirmed rather than credited: a third party consultancy reports a service organisation control type two report, and it does not appear on the company's own security page, so under the index rule that a credential must appear in the vendor's own material it is not counted here.

Boost.ai is a technology supplier to regulated institutions and holds no financial licence, which is the correct posture and carries no penalty. Its positioning is built around regulated industries and its certifications are scoped accordingly, but under the standing index ruling technical conformity assessments do not read across as regulatory standing and are credited on the security and privacy axes instead.

No supervised regulator test, sandbox participation or programme admission was located, and no statement of position under the European artificial intelligence regime was found despite a European base and deployment inside banks and public bodies.

Bias is named as a risk the Trust Layer guardrails are built to address, which is more than the silence that is the norm on this axis, and it stops at naming. No fairness testing, differential outcome monitoring or impact assessment is published.

One gap is specific and measurable: the platform operates across Norwegian, Swedish, Danish, Finnish and English at minimum, and no per language accuracy or containment figure is published, so a customer served in a less represented language may be understood less reliably than one served in English. Public sector deployment sharpens this, because a citizen facing service agent that performs unevenly across languages affects access to a public entitlement rather than a commercial product.

No error rate, remediation commitment, liability position or correction path is published. The guardrail architecture is a control against a wrong answer being produced, which is credited on the safety and oversight axes, and it is not a route for someone who has already received one.

A customer told something incorrect about an account, a payment or an entitlement by a virtual agent has no stated way to have that interaction reviewed or corrected, and nothing describes whether a transcript is retained in a form that would let the institution reconstruct what was said.

No model provider, family, version or country of processing is disclosed. The published material discusses large language models in general terms as the technology underlying modern virtual agents and identifies their risks candidly, without stating whose models the platform runs or whether they are hosted by the company or called from an external provider.

For a vendor whose whole proposition is reliability in regulated settings, and whose customers include national banks and public bodies, the question of whose infrastructure processes a citizen's or customer's words is one a buyer would expect answered.

Nothing describing integration depth was located beyond a listing on one enterprise software marketplace. No core banking system, policy administration system, contact centre platform or data source is named as a supported integration, and no integration count is published.

Deployments at Nordea, DNB and Telenor make deep integration a practical certainty, so this is a disclosure finding rather than a capability one, but it is a conspicuous omission against peers in this category that publish counts in the hundreds and name the core platforms they connect to.

Private cloud deployment is offered alongside the standard hosted service, which is a genuine second model and more than most of this category provides. The European base and the group wide privacy certification give a buyer reasonable grounds to expect European processing.

Held off the top grade because nothing states it: no region list, no residency commitment, no single tenant description and no on premises path was located, so a bank in a jurisdiction with localisation requirements would be relying on inference from the company's domicile rather than on a published position.

No pricing, tier structure, billing basis or indicative range is published, and the route to a number is a demo request. This is the index norm and is measured against Sumsub, which publishes per verification rates on a public page.

The absence is more noticeable here than usual because the company publishes precise operational figures elsewhere, including certification control ratios and conversation volumes, so the reticence is specific to price rather than a general reluctance to publish numbers.

More than 450 organisations across four regulated sectors, banking, insurance, telecommunications and the public sector, which is genuine breadth of buyer type. Within financial services the coverage is narrower than the headline count suggests. The installed base is concentrated in the Nordics and wider Europe, and the workload is customer service and internal employee support rather than lending, servicing or back office process execution.

Two separate competitors, each writing on their own domain and therefore treated as leads rather than findings, make the same architectural observation, that the platform is positioned for conversational support rather than end to end operational automation, and the vendor's own material is consistent with that reading.