Private by architecture
Keep knowledge, local inference and governance inside your network. Agree access, retention and operational controls with your team.
← FOR BUSINESS
powered by neurologic
appana CONNECTED INTELLIGENCE / FLAGSHIP
powered by neurologicYour expertise. On your infrastructure.
Private, on-premise AI shaped around your experts. Fine-tune specialist models on approved business knowledge and build agents around the work that matters, with no external model-token charges for locally hosted models.
ON-PREMISE / ENTERPRISE & SOVEREIGN
Your infrastructure. Your intelligence. Your control.
ACI Enterprise and ACI Sovereign are scoped for on-premise deployment, with locally hosted models and an offline operating configuration.
No external token bill
Locally hosted inference does not incur external model-token or API charges. Hardware, electricity, implementation and support still apply.
Room for more agents
Create agents around the workflows you need. Capacity is sized for your workloads, model choices and concurrent use.
Tuned to your work
Fine-tune specialists on approved examples and evaluate accuracy on unseen business tasks before expanding responsibility.
Offline operation requires local models and internal dependencies. Public model APIs, internet research and external SaaS connectors need connectivity when enabled.
THE PRODUCT
Expert knowledge becomes a specialist.
Our flagship enterprise product combines appana’s workflow design and integration approach with Neurologic’s specialist-learning technology. Customer scope, deployment and available capabilities are agreed for each engagement.
Specialist learning has been demonstrated in focused supplier evaluations. Capture and execution tools are in development; team guidance and workflow-variance tools are planned. General autonomous customer operation is not currently available.
FROM EXPERTISE TO EVIDENCE
One workflow. A deliberate path.
01
Understand expert work
Use approved examples to retain the context, decisions and outcomes behind a bounded workflow. Capture is designed to be visible and participant controlled.
02
Shape a specialist
Select work that follows the agreed rules, then specialise a model around the workflow and its exceptions.
03
Validate before expanding
Test against separate customer-held work. Review errors, escalation and outcomes before agreeing any greater responsibility.
PRODUCT CAPABILITIES
Different roles. Connected intelligence.
Expert capture
Visible, participant-controlled observation of expert work, with pause and local privacy handling as product requirements.
Team guidance
Read-only answers supported by relevant examples, with uncertainty made clear when the evidence is insufficient.
Workflow insights
Aggregate views of how work varies across a process, supporting improvement without individual productivity rankings.
Operations review
Review queues, workflow-specific approvals and records linking proposed actions to their evidence and outcomes.
CONTROL STAYS WITH YOU
Your knowledge. Your boundaries.
Your environment
An on-premise configuration keeps models, knowledge and agent execution inside your infrastructure. Offline operation uses local models and internal systems; external integrations are optional and need connectivity.
Your knowledge
The design keeps customer-instance data separate from other customers and foundation-model training.
Your authority
People define permitted actions, review uncertain cases and decide when responsibility may expand. Controls are agreed per workflow.
Start with a workflow worth proving.
- 01Define one workflow and its success measures
- 02Review data, access, privacy and operational readiness
- 03Agree the scope and validate on separate work
- 04Expand only when the evidence supports it
What is ACI Enterprise?
ACI stands for appana Connected Intelligence. ACI Enterprise is appana’s flagship white-label product powered by Neurologic’s Digital Human technology, offered through an appana-led enterprise engagement.
Is every capability available today?
No. The supplier has demonstrated specialist training in focused evaluations. Capture and execution tooling remain in development, while guidance and workflow-variance tools are planned. We agree available scope before any commitment.
Can it run without internet?
We scope an on-premise deployment with local models and internal systems. The offline configuration has no dependency on public model APIs. External SaaS integrations need connectivity and are agreed separately.
Does local inference have token charges?
No external model-token or API charges apply when inference runs on locally hosted models. Infrastructure, electricity, implementation, licensing and support costs remain part of the deployment.
How many agents can we create?
Build agents around as many business workflows as you need. We size the infrastructure for your model choices, workloads and concurrency; creating more agents still requires sufficient compute capacity.
Can a fine-tuned specialist beat a general-purpose model?
Fine-tuning targets your process, vocabulary and exceptions. Accuracy is tested on separate business tasks against agreed baselines, which can include Claude. A result on one workflow is not a claim of superiority across all tasks.
How do we start?
Bring one bounded workflow and the people who understand it. We will review readiness, define the pilot scope and success criteria, and agree responsibilities before implementation. No fixed production date or guaranteed savings are implied.
Technology sources and scope
Adapted from Neurologic’s public product, technology and trust materials, reviewed 23 October 2026. Diagrams illustrate the intended method and architecture. Supplier evaluations are not appana customer results or independently audited guarantees.
CUSTOMERS & PRIORITY USE CASES
Turn expert judgement into a specialist your organisation can build on.
Insurance and financial operationsEngineering and technical organisationsOrder, service and document-heavy operations
01 / Insurance operations
Claims intake and coverage review
Connect policy documents, intake records and linked cases. Surface conflicting evidence and prepare the next review step for an authorised claims handler.
What to measure: Source selection, missing checks, reviewer corrections and handling time.
Explore the workflow02 / Medical-device distributors and service operations
Order-to-account decisions
Combine purchase-order fields with site, department and account history. Suggest the account or route, with supporting evidence and an exception queue.
What to measure: Account-match accuracy, abstention quality and review effort.
Explore the workflow03 / Engineering and R&D teams
Engineering hypothesis and verification
Use approved expert attempts to propose focused changes, run the agreed test or simulator and learn from verified outcomes.
What to measure: Constraints met, valid proposals and engineering review time.
Explore the workflow04 / Large operational and service teams
Specialist knowledge assistance
Answer bounded questions using approved procedures and examples. Keep source references visible and escalate when the evidence is insufficient.
What to measure: Evidence quality, answer acceptance and escalation accuracy.
Explore the workflow05 / Finance and procurement teams
Document-to-workflow coordination
Extract information, compare it with approved records and prepare the next task across existing systems. People retain approvals for commitments and postings.
What to measure: Field accuracy, exceptions resolved and repeated entry avoided.
Explore the workflow06 / Shared services and complex support
Case preparation and service hand-offs
Prepare a concise case history, identify missing evidence and route the next action to the responsible team.
What to measure: Case completeness, routing quality and time to prepare.
Explore the workflowPriority workflows to configure and validate. Supplier specialist training is demonstrated in bounded evaluations; capture and general execution tooling retain their published development status.
Explore more use casesSEE THE WORKFLOW
From a conflicting record to an evidence-led next step.
01
Read the case
Bring together an intake record, the authoritative policy and linked records.
02
Resolve the context
The specialist identifies a mismatch and cites the source that should govern the next check.
03
Prepare the action
Create a review note and a proposed task; retain uncertainty and missing information.
04
Keep the person in control
An authorised reviewer decides what may proceed and records the correction or outcome.