hc-digitalsysteme
Approach

One process without hand-over losses – from discovery to operations.

Strategy and delivery form one continuum here: the same team, continuous accountability, measurable interim results. Below is every step – with duration, participants and what is on the table at the end.

Abstract illustration: six ascending paper planes connected by one continuous line with node points.
The process

Six steps, one accountable team.

  1. Discovery & assessment

    Interviews with newsroom, product, IT and sales; review of systems and data readiness. First we understand your company – then we talk about models.

    Duration: 1–2 weeks Participants: Your departments, our consulting team Outcome: Maturity picture and topic map

  2. Strategy & roadmap

    Score and prioritise use cases, calculate business cases, define governance guardrails. In the end your leadership decides on numbers.

    Duration: 2–4 weeks Participants: Your leadership & experts, our consulting team Outcome: Prioritised roadmap with business case per use case

  3. Prototype / MVP

    The most important use case becomes real: with real content, in a protected environment, against measurable acceptance criteria. Your newsroom tests from week two.

    Duration: 4–6 weeks Participants: Pilot newsroom, our development team Outcome: Working prototype plus measurements and a scaling recommendation

  4. Implementation & integration

    The prototype becomes a system: integration into CMS and editorial processes, scaling to real throughput, security and privacy sign-offs.

    Duration: 6–12 weeks, depending on scope Participants: Your IT & departments, our development team Outcome: Production system in daily editorial use

  5. Operations & MLOps

    Monitoring of quality and cost, model updates, further development. An AI system only stays good if it is observed and adjusted.

    Duration: ongoing Participants: Your operations or ours, in a support model Outcome: Stable operations with quality and cost reporting

  6. Enablement

    Training for newsroom and leadership, guidelines, AI literacy under Art. 4 EU AI Act. So the system gets used – not just exists.

    Duration: accompanying, formats from half a day Participants: Your teams, our consulting team Outcome: A company that carries AI forward on its own
Principles

What you can hold us to in every step.

Agile, but binding

Short cycles and visible progress – with fixed milestones, budgets and acceptance criteria.

Measurable ROI

Every use case has a metric before it gets built. What cannot be measured does not get scaled.

Privacy by design

Data protection is an architecture decision: EU hosting, data minimisation, DPAs and documented data flows from day one.

No black box

You understand at all times what the system does, what it costs and where its limits are. Code, documentation and knowledge belong to you.

Time & collaboration

Realistic timeframes, three collaboration models.

A full project from first discovery to production system typically takes 3 to 6 months. For the entry, pick the model that fits your company's budget situation and risk appetite:

Model 1

Project

Defined scope, defined price: from strategy to accepted system. The right model when goal and frame are clear.

Model 2

Retainer

Continuous development and operations with a fixed monthly allocation – for companies building AI as a lasting capability.

Model 3

4-week pilot

The low-risk entry: one use case, fixed price, measurable criteria, then an honest scaling recommendation. Details on the case studies page.

FAQ

Frequently asked questions about our approach

How long does a full project take, from idea to operations?

Realistically 3 to 6 months: a few weeks of strategy, 4 to 6 weeks of prototyping, then scaling and integration. Companies with clear priorities and available experts are faster; complex system landscapes take longer. After discovery we give you a reliable estimate.

Do we have to start with consulting, or can we go straight to development?

Both work. If your use case is clearly prioritised, we start directly with the prototype. If several ideas compete or the business case is missing, a few weeks of strategy save months of misdirected development later.

How much effort lands on our team?

Planned and limited: discovery needs interview slots with your experts, the prototype needs a pilot newsroom that tests and gives feedback – typically a few hours per week and person. Without this involvement we do not deliver, because adoption does not happen by hand-over.

Which collaboration models do you offer?

Three: a classic project with defined scope, a retainer for continuous development and operations, or the 4-week pilot as a low-risk entry at a fixed price. Many clients start with the pilot and then move into project or retainer.

What if a prototype misses its criteria?

Then we say so – with an analysis of why: data readiness, model limits, process. You receive all results and an honest recommendation on whether improving is worthwhile or the budget achieves more elsewhere. A cleanly ended pilot is cheaper than any half-heartedly scaled system.

The first step is a conversation.

Tell us where your company stands – we will tell you which entry pays off: consulting, pilot, or straight into delivery.

Book an intro call Go to AI consulting