Service · Strategy & Discovery
AI Roadmap
From business goals to a sequenced AI roadmap, with target architecture, operating model, and governance in one picture.
AI initiatives rarely fail on technology. They fail because nobody decided which business goals they serve, who operates them, and how success is measured. We build a roadmap with you that starts from the business goal and carries through to architecture, operations, and governance, for cloud, sovereign cloud, or on-premises.
The situation
Most organisations that come to us do not have an AI problem. They have a sorting problem. Pilots are running in several teams, a few of them work surprisingly well, and almost none of them are in production. Nobody can say with confidence what the initiatives cost together, what they contribute to the business result, or who answers for it when an agent makes the wrong call at three in the morning.
Pressure builds from two directions at once. Business units want the tools they already use privately. Security, data protection, and audit want to know the rules under which that happens. A roadmap that serves only one of those two sides gets either ignored or blocked.
How we work
Ambition and value logic
We start from your business strategy, not from the technology stack. With leadership and business units we establish which goals AI should support over the next twelve to twenty-four months, and how the contribution will be measured.
Deliverable
Agreed ambition with value logic and success metrics
Baseline assessment
An honest read of the current state across data, platform, processes, capabilities, and governance. We review what is already running, what of it will hold up, and where the real bottlenecks sit.
Deliverable
Maturity picture with named bottlenecks and credible quick wins
Opportunity portfolio
We map candidates along the value chain and separate cases that grow revenue and decision quality from cases that reduce effort, cycle time, and error rates. The two are funded differently, measured differently, and governed differently.
Deliverable
Prioritised portfolio scored on impact, feasibility, and risk
Target architecture and operating model
The reference architecture for your context: model access and gateway, identities for non-human actors, logging, cost attribution, reuse. Plus the question that gets expensive when asked late: who builds, who runs, and who is accountable for an agent when it misbehaves.
Deliverable
Target architecture, stack decisions, and an operating model with named roles
Sequencing and roadmap
Waves, not a wish list. We order the portfolio by dependencies, the foundation work it presumes, and what can realistically be funded, so early results carry the later ones and the organisation grows into it.
Deliverable
Sequenced roadmap with milestones, effort estimates, and decision points
Fields of action
Knowledge work and assistance
AI where the daily work happens. Research, documents, minutes, analysis, ticket and request handling. The field that earns adoption and turns users into a source of ideas.
Process and workflow automation
Agents inside the operational chain rather than beside it. Data pipelines, quality gates, monitoring, service and reconciliation processes. The field the durable business case comes from.
Platform and tech stack
Reference architecture, model access, and integration with the systems you already run. Where it runs is a decision, not a given: hyperscaler, sovereign cloud in Switzerland, or on-premises for everything that cannot leave the building.
Governance, security, and compliance
Risk classes, approval tiers and human-in-the-loop thresholds, access and identity, traceability and logging. Aligned to the revised Swiss FADP and the EU AI Act, and embedded in your existing governance rather than parked next to it.
Shared building blocks
Templates, curated skills and prompts, evaluation and guardrail frameworks, delivery paths. So twelve teams do not solve the same problem twelve times and pay for it twelve times.
Enablement and adoption
Training, roles, ways of working, and adoption measurement. The part without which a roadmap stays a presentation.
What you get
- Prioritised use-case portfolio, split by growth impact and efficiency impact
- Target architecture including stack decisions for cloud, sovereign cloud, or on-premises
- Operating model with roles, accountabilities, and approval paths
- Governance framework with risk classes and human-in-the-loop thresholds
- Sequenced roadmap with milestones, effort estimates, and decision points
- A defensible basis for budget and steering-committee decisions
Engagement
- Duration
- Scoped to fit, from a compact format for a single business unit to an organisation-wide engagement.
- Getting started
- A short scoping session that establishes scope, participants, and effort.
- Who is involved
-
Leadership and business units for ambition and prioritisation
IT, architecture, and operations for the baseline and target architecture
Security, data protection, and compliance for the governance framework
- Format
-
A working format of workshops, interviews, and review of the existing landscape
Continuous interim results rather than one big reveal at the end
Deliverables in your formats, usable in steering and budget processes
On site in Switzerland or remote
Frequently asked
How is this different from a strategy deck?
What you end up with is a sequenced roadmap with a target architecture, operating model and governance that a team can pick up the following Monday. A presentation is a by-product, not the deliverable.
How long does a roadmap stay valid?
It is cut for roughly twelve months, with a checkpoint each quarter. It is built to be updated rather than rewritten.
Who owns the code, models, and results in the end?
You do. Code, configuration, and documentation transfer to you in full, in your repositories and your environment. We build so your team can carry on without us.
Where do our data and models run?
In your environment. Your own cloud, your own data centre, or on-premises, in a Swiss or European region on request. Data leaves your environment only for the providers you explicitly approve.
Do you work remotely or on-site?
Both. We are based in Zürich and work on-site across Switzerland; the day-to-day of the engagement works remotely.
Can we start with a small, bounded step?
Yes, and that is the norm. The first slice is deliberately small so something usable exists early and the decision about the next step rests on experience rather than on a plan.