DENQ
DEContact
Home / Our approach

From target state to implementation of AI agents.

We work on three levels: target state and priorities with management, process analysis with the departments, and implementation in your systems. The KPIs from operation then feed back into decision-making.

Two galloping horses in motion blur, crossed by a blue line of light
THREE LEVELS, ONE APPROACH
Strategic levelTARGET PICTURE
Process levelOPERATING MODEL
System levelIMPLEMENTATION
Strategic level01

Target state and use case portfolio

Together with management, we define the target state for AI use and derive a prioritised use case portfolio from it. The basis is business value, feasibility, risk profile and strategic relevance.

RESULTS
Target state for AI use
Aligned with company goals, budget and organisation.
Prioritised use case portfolio
Assessed by business value, feasibility, risk profile and strategic relevance.
Business cases and roadmap
With clear priorities, sequence and responsibilities.
Process level02

Current-state analysis and target design

We map the relevant processes in a structured way, together with the process owners and the employees who carry them out every day. In doing so, we make implicit experience explicit.

RESULTS
Structured current-state analysis
Triggers, data flows, decision points, manual handoffs between systems, waiting times and responsibilities.
Target process
Developed with the people who own the process today.
Classification per process step
Basis for choosing the right solution approach at system level.
System level03

The right solution approach for every process step.

We then implement agents, workflows and interfaces in your existing system landscape, including a role and permission concept, approval processes and logging. Where a process does not require an agent, we neither recommend nor implement one.

SOLUTION APPROACHSUITABLE WHEN …CLASSIFICATION
Rule-based automationClear rules, stable inputs, recurring processes.Deterministic and most economical
Data integrationInformation is spread across several systems.Creates the basis for every further level
Language modelUnstructured content needs to be read, summarised or written.Transformation without its own decision
Agent-based decisionAmbiguous information, multiple sources, choice between courses of action.Only with defined rights and approvals
Human decisionFinancial, legal or reputational consequences.Deliberately stays with people
Feedback loop04

KPIs from ongoing operation determine prioritisation, autonomy levels and scaling.

Our agent systems record processing volume, lead time, correction rate and exceptions. The Agent Board reviews them regularly and decides on further use on that basis.

Target stateImplementationMeasurementSteering
VALUE TRACKING
Working together05

From the first engagement to handover to your organisation.

PHASE 01

AI Operating Review

Target state, prioritised use cases, assessment of existing initiatives and governance framework in two to three weeks.

PHASE 02

Implementation with approvals

Implementation of the prioritised use cases, integration into the system landscape and start at a low autonomy level.

PHASE 03

Transition to day-to-day operation

Embedding in roles and processes, setting up the Agent Board and monitoring through the KPI framework.

PHASE 04

Scaling and handover

Reuse of the architecture for further use cases and gradual handover of management to your organisation.

The entry point is the AI Operating Review.

Prioritised use cases, business cases and a governance framework in two to three weeks.

View AI Operating Review