The infrastructure layer.
We build the systems we then operate. Most of what we build solves a specific problem for a specific client and lives inside their architecture. Some of those pieces, when they prove useful across multiple contexts, graduate to products.
What we build.
Email and CRM platforms.
Complete stacks on Mautic with multiple MTAs (KumoMTA, Sparkpost, Amazon SES) and the VDMS suite for volume management, reputation, and deliverability. Some clients we operate ourselves, others we deliver turnkey and hand off to the internal team.
Specialized agents.
Agents that live inside the client’s processes. Campaign analysis, predictive scoring, first-line support, editorial assistance. Designed with a bounded domain, explicit memory, and a clear protocol for when to escalate to a human.
Data architectures.
Pipelines, data lakes, attribution models, dashboards that show what matters and nothing else. Built on AWS, Hetzner or OVH depending on where the client wants to keep their data.
Integrations.
Fast, durable connections between the systems the client already uses. SFMC, Braze, HubSpot, ActiveCampaign, Kit, Brevo. Custom ETLs when no standard connector solves it.
Products.
Sendability (agentic email and CRM platform) and BrandExpand (human and AI content engine) are the two products that have come out of this practice so far. When an internal project proves value across multiple contexts, we extract it as a product.
How a build project starts.
A typical build project starts with two discovery sessions, where we look at the existing architecture, available data, and what the internal team is trying to solve. A short document comes out of that: what we are going to build, in what order, with what measurable milestones. The first month is usually a small functional version that already delivers value. Subsequent months expand and refine. Most build relationships evolve into mid-term operation.