Reimea

The AI control platform

Put AI agents to work.
Stay in control of every action.

Reimea gives regulated organisations ready-made AI agents and the controls to run them safely. Guardian watches every agent and service, every action passes one control layer, and a named person signs off what matters. Deploy AI agents. Keep learning under control.

See Guardian

Azure & AWS deployment options. Built around your governance.

REIMEA / CONNECTED INTELLIGENCE

Illustrative data flows between agents, knowledge and controls.

Agent activityillustrative
    chain head#000184 · 3c07e91a
    • 1 · Deploy20 agent-powered products for voice, service, knowledge, care and clinical work.
    • 2 · ControlIdentity, policy, one AI gateway and named approval around every agent action.
    • 3 · OverseeGuardian shows what every agent is doing and where a person needs to look.
    • 4 · ProveA tamper-evident record of who acted, on what, and who approved it.

    The problem

    AI agents are easy to start. Hard to sign off.

    Most regulated organisations already have AI pilots. Few reach production, because nobody can show who the agent acted for, what it was allowed to do, what it knew and who approved the result.

    01

    Every tool brings its own controls

    Each AI product arrives with its own sign-in, model settings and logs. Security and governance teams have to assess every one from scratch.

    02

    Nobody owns what the AI knows

    Policies change and answers drift. No one can show which version of the guidance an agent relied on, or who approved it.

    03

    Approval is an afterthought

    "Human in the loop" becomes clicking OK on a screen nobody reads. When something goes wrong, there is no clear decision record.

    04

    Data goes where you can't see it

    Prompts and records flow to vendor clouds and model providers you don't control, outside your keys and your audit.

    The evidence

    Uncontrolled agents are failing in production. Expensively.

    40%+

    of agentic AI projects will be cancelled by 2027

    Gartner cites escalating cost, unclear value and inadequate risk controls.

    Gartner, Jun 2025

    54

    AI agent incidents per enterprise in a year

    On average across 2,000 technology leaders; 17% were high severity, led by data exposure and cascading failures.

    IBM Institute for Business Value, 2026

    97%

    of organisations with AI breaches lacked AI access controls

    63% had no AI governance policy at all. Shadow AI added about $670k to breach costs.

    IBM Cost of a Data Breach, Jul 2025

    80%

    of executives had an AI incident reach the board

    Only 37% say their governance scales to agentic AI. Survey of 3,221 EMEA executives, UK included.

    IBM, Oct 2026

    30%

    of realistic office tasks completed by the best agent tested

    Carnegie Mellon's TheAgentCompany benchmark. The rest failed or were left unfinished.

    Carnegie Mellon University, 2025

    35%

    success rate for agents in multi-turn business tasks

    Down from about 58% on single-step tasks, in Salesforce's CRMArena-Pro tests.

    Salesforce AI Research, Jun 2025

    50%+

    of GenAI projects abandoned after proof of concept

    Weak risk controls, poor data, rising cost and unclear value among the causes.

    Gartner, Jan 2026

    Figures are from the sources linked; survey figures are self-reported and benchmark results reflect the agents tested at the time. They describe the market, not Reimea customers.

    The market

    Agents are going mainstream. Control is the missing piece.

    Spending on AI is exploding and agents are moving into everyday business software. Every one of those agents needs identity, permissions, oversight and evidence. That is the market Reimea is built for.

    40%

    of enterprise apps will feature AI agents by 2026

    Up from less than 5% in 2025.

    Gartner, Aug 2025

    $1bn+

    AI governance platform spend by 2030

    From $492m in 2026, as AI rules spread to 75% of the world’s economies.

    Gartner, Feb 2026

    Why it matters: most vendors sell either the agents or the governance. Reimea sells both in one platform, so the controls arrive with the agent instead of months later.

    Meet Guardian

    The control room for your AI agents.

    Guardian gives supervisors one live view of every service, every agent and every person handling the work. It shows what is happening, raises the moments that need a human, and puts your approved policy beside the conversation, so people can step in before a small issue becomes an incident.

    • See everythingCalls, waits, agent activity and service targets across every service, in one supervisor view.
    • Know where to lookAlerts and conversation cues arrive with a reason and the evidence behind them.
    • Policy at handApproved, versioned policy passages appear beside the conversation. Nothing unapproved is ever shown.
    • Your rulesWatchwords and thresholds you configure, per service, without retraining any model.
    • People decideGuardian advises. A named supervisor reviews, acts and is recorded.
    • Proof afterwardsQuality review and a tamper-evident record show who saw what, and what they did.
    meridian · guardian · illustrative
    Meridian Guardian supervisor view with service activity and review cues
    Human decidesCues with evidenceOne view of every agent

    The control layer

    How control works on every agent action.

    Reimea’s design brings identity, model policy, human approval and evidence together. The reviewed build has task and route coverage to validate before production use.

    Reimea control layer architecture Five agents on the left connect into the Reimea control layer in the centre, made of identity, policy, AI gateway, human approval and audit. The control layer connects to your systems on the right: telephony, case and record systems, service desk, knowledge sources and operational data. Agents Reimea control layer Your systems Voice agentAnswers, captures, hands over Service desk agentTriages and suggests owners Knowledge agentAnswers with cited sources Intelligence agentForecasts and monitors Your own agentsRoadmap: bring your own IdentityCustomer identity & accessverified Policy and entitlementsEdition, role and purpose of useallowed AI gatewayShared route for registered tasksuk-south Human approvalA named person decidesawaiting Audit ledgerAppend-only, hash-chained#000184 TelephonyCalls and transcripts Case and record systemsHandover and capture Service deskTickets and inventory Knowledge sourcesOwned, reviewed, published Operational dataImports and monitors
    Built in Meridian todayOn the roadmap Azure and planned AWS deployments, scoped to your environment.

    Identity

    Identity and access are scoped to the customer environment. The Azure baseline uses Microsoft Entra ID; AWS identity integration is defined for its deployment configuration.

    Policy

    Editions and roles decide which agents and modules each person can use.

    AI gateway

    Registered tasks use the AI gateway. The reviewed hosted configuration restricts provider selection; region and task coverage need testing.

    Approval

    Agents draft captures, suggestions and answers. People check and decide.

    Audit

    Application audit chains and structured logs support review. Durable storage and coverage need verification.

    Why Reimea

    Others give you agents or governance. Reimea gives you both.

    Buyers usually weigh four alternatives. Each solves part of the problem. Reimea is built to solve the whole of it, for organisations that have to answer to a regulator.

    Agents and controls in one product

    No separate governance project for every new agent. Add a product and it inherits the same identity, gateway, approval and audit.

    Your cloud, your keys

    A deployment per customer in your own Azure environment, with AWS planned. Your data stays inside your boundary and under your audit.

    Built for the people who sign it off

    Designed around the questions a CISO, SIRO and clinical safety officer ask, with the evidence laid out for DCB0129, DCB0160 and DSPT reviews.

    Honest about readiness

    We publish what is built and what still needs validation. You plan your assurance around facts, not a sales deck.

    How Reimea compares with the alternatives
    What you needReimeaGeneral AI assistantsSingle-purpose AI toolsAI governance softwareBuild it in-house
    Agents that do the operational work20 products across operations, care and clinical Yes Partly Yes No You build it
    One control layer across all your agentsIdentity, policy, gateway, approval and audit, shared Yes Partly No Partly You build it
    A named person approves consequential actionsBuilt into each workflow, not added afterwards Yes Varies Varies No You build it
    Controlled learning with owners and versionsStaged, reviewed, published; agents use approved versions Yes Partly No Partly You build it
    Runs in your own cloud environmentCustomer-specific Azure deployment; AWS planned Yes No Varies Varies Yes
    Designed for UK health and care assuranceEvidence for DCB0129, DCB0160, DSPT and DPIA reviews Yes No Varies No You build it
    Tamper-evident record of what agents didHash-chained application audit events Yes Partly Varies Partly You build it

    Categories describe typical products in each group, not specific suppliers. Check any supplier's current capabilities, ours included, against your own requirements. See the assurance page for what is built and what still needs validation.

    See it working

    Every product, one control layer.

    All 20 products and Knowledge Control, across operations, care, clinical and records, running on the same identity, approval and audit. Screens are from the Meridian product; clinical and Exchange views remain demonstrators. See the assurance page for validation boundaries.

    meridian · customer environment · illustrative

      Find a product

      Pick the work. See the products for it.

      Choose a group or search by name. For every product side by side, go to the full product range.

      6 of 21 products

      NEW / CARE COORDINATION

      Meridian Clinical Case Management

      One episode. Every handoff in view.

      Bring reviewed intake, encounters, follow-up tasks and case context into a human-owned episode timeline.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian AI Scribe

      Consultation words in. A clinician-reviewed note out.

      Turn a consented consultation transcript into editable structured fields and a clinician-reviewed note in the selected local case.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Virtual Receptionist

      An informed front door that knows when to hand over.

      Answer approved service questions, capture requests and prepare a clear handoff through a web or browser-voice reception experience.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Transitions

      A handoff is work. Give it an owner.

      Bring referral gaps, pathway milestones and home-care follow-up into one coordination queue with a visible next action.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Neighbourhood Capacity

      See the work. Test the assumptions.

      Compare open coordination tasks with locally entered team-capacity assumptions and make unconfigured workload visible.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Outcomes & Prevention

      Measure the process. Keep the meaning clear.

      Review referral flow, home-care participation and recorded coordination activity with their source and reporting window visible.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Care Pathways

      An agreed follow-up. A clear, owned timeline.

      Turn a care-team follow-up plan into owned milestones, due dates and visible human-recorded progress.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Access & Referrals

      From intake to the next human decision, without the gaps.

      Track referral completeness, ownership, contact attempts, waitlist status and recorded appointment offers in a visible worklist.

      Features, AI and workflow ↗ Brief (PDF)

      NEW / CARE COORDINATION

      Meridian Home

      Care at home, with an accountable team watching.

      Coordinate home-care episodes, patient check-ins and consented observations with a clear staff-review workflow.

      Features, AI and workflow ↗ Brief (PDF)

      CLINICAL

      Meridian Crisis Text

      Human-led text conversations with visible oversight.

      A synthetic text-conversation workspace separating operator replies, consented AI simulation and Guardian oversight.

      Features, AI and workflow ↗ Brief (PDF)

      CLINICAL

      Meridian Clinical Symptom Checker

      Structured caller context for trained review.

      Explore a voice-led assisted-assessment conversation with reported symptoms, pathway trace and human-review considerations.

      Features, AI and workflow ↗ Brief (PDF)

      CLINICAL

      Meridian Risk Guardian

      Selected evidence. Transparent review signals.

      Bring deliberately selected conversation and synthetic-record context together for a human-reviewed local signal assessment.

      Features, AI and workflow ↗ Brief (PDF)

      OPERATIONS

      Meridian Voice

      Answer every call. Capture it properly. Hand over with confidence.

      Agent-led voice interaction, versioned scripts, transcript capture and a human-reviewed handover.

      Features, AI and workflow ↗ Brief (PDF)

      OPERATIONS

      Meridian Listen

      Hear what people actually said, with the evidence beside every signal.

      Conversation understanding with topics, concerns, evidence and candidate open questions for a handler.

      Features, AI and workflow ↗ Brief (PDF)

      OPERATIONS

      Meridian Service Desk

      Every request owned, on target and resolved with the right answer.

      Service requests, assignment, targets, inventory and reviewed knowledge-assisted suggestions.

      Features, AI and workflow ↗ Brief (PDF)

      OPERATIONS

      Meridian Knowledge

      Answers your people can check, from sources you approved.

      Scoped passage retrieval, inspectable references and bounded answer assistance.

      Features, AI and workflow ↗ Brief (PDF)

      OPERATIONS

      Meridian Intelligence

      Operational intelligence with the assumptions left in.

      Operational data, measures, reports, forecasts and monitored conditions with their assumptions visible.

      Features, AI and workflow ↗ Brief (PDF)

      CLINICAL

      Meridian Wellness

      Guided reflection, with clear limits and a route to real help.

      Bounded, guided reflection and wellbeing conversations in a clinical demonstrator.

      Features, AI and workflow ↗ Brief (PDF)

      RECORDS

      Reimea Exchange

      The source matters. So does what is missing.

      A source-attributed shared-care-record design, with visible provenance and missing-source context.

      Features, AI and workflow ↗ Brief (PDF)

      GOVERNANCE

      Knowledge Control

      Govern what your AI agents are allowed to know.

      Content ownership, staging, approval, versions, dates and publication scope around the knowledge AI uses.

      Features, AI and workflow ↗ Brief (PDF)

      See all 20 products and Knowledge Control

      How you start

      Start small. Decide on evidence.

      One workflow, a clear owner and a scorecard agreed before the result. No patient data needed to begin.

      1. Step 1 · 45 minutes

        Briefing

        Your workflow, a walk-through of the relevant product and its limits, and whether a discovery engagement makes sense.

      2. Step 2 · Discovery

        Scope one service

        A workflow map, approval points, data flows and the evidence needed before live use, with baseline measures agreed up front.

      3. Step 3 · Bounded pilot

        Measure and decide

        Handling time, capture completeness, correction rate and evidence quality against the scorecard, then a clear stop or go.

      Questions

      What buyers ask us first.

      What is an AI agent control layer?

      It is the set of controls between AI agents and the systems they act on: who the agent acts for, which models it may call, which actions need a person to approve, and a record of everything it did. Reimea Meridian provides that layer for voice, service desk, knowledge and intelligence work.

      How does Reimea keep people in control of AI agents?

      Agents draft and suggest, and a named person reviews and decides. Registered AI tasks use a shared gateway. Identity and permissions are scoped to the customer environment; the Azure baseline uses Microsoft Entra ID. The reviewed build includes application audit chains; completeness and durable retention need verification.

      Where does Meridian run, and where is our data stored?

      Deployment options cover Azure and planned AWS. The Azure baseline describes customer-specific resources, including identity, keys, storage and a database. AWS services and identity integration will be defined for its configuration. Agree the hosting region, model processing locations, data flows and retention before deployment.

      Which AI models does Meridian use?

      Registered model tasks use Meridian’s AI gateway. Hosted provider restrictions exist in the reviewed configuration; each task, model SKU and processing region still need deployment verification.

      Is Meridian approved for clinical use?

      Clinical and patient-facing modules are demonstrators. Production use requires an agreed intended purpose, appropriate safety evidence, security and identity validation, approved connections where relevant, and deployment-owner approval.

      What is the difference between Meridian and Exchange?

      Meridian runs voice-led service operations and the agents within them. Exchange is a source-attributed shared care record, currently in development. They are separate products, and a Meridian login never gives access to Exchange records.

      Put agents to work. Keep the controls.

      A 45-minute briefing for your technology, security and governance leads. We cover the control layer, Azure or AWS deployment planning and an honest view of what is ready.

      Request a briefing

      Tell us what you want to improve. We aim to reply within one working day to arrange a suitable time.

      Prefer email? [email protected]. A briefing request does not confirm an appointment.

      Open the pilot scorecard · Download Excel scorecard