phase 03 · first quarter
currently booking Q3 2026 onwards · discovery & sparring open sooner · last update 25.04
Midnight Labs

midnight labs · learning in the work

People learn from the work around them, not from the course catalogue.

one sceneA regional line in a 40,000-person bank went from a crowded LMS to peer-reviewed AI-assisted credit memos in eight weeks: same stack, new habits, visible in the work.

We help teams and leaders get sharper at real work. The environment (what surrounds the job), the social (who gets heard and who models what), and the technical (tools, data, and flow) are one system: we make them line up, help groups use AI together with clear norms, and help individuals keep judgement when models are always on offer. Fewer programs. More durable habits.

principles

We change the shape of the work, not the slide deck.

Courses sit outside the job. We work where people already spend their time: the room, the tools, the handoffs, and the written record. If the work does not teach something better after we leave, we have not earned our fee.

01 / environment

The three layers beat another module.

Physical and digital surroundings, who gets heard, and the tools in between: we redesign the parts that are teaching the wrong lesson. Growth shows up in the job, not in completion rates.

02 / signal

If we cannot see it in the work, it did not happen.

We pick a few concrete behaviours (decisions, recoveries, handoffs) and read them in how the work actually shows up: what people ship, how they recover, how handoffs land. Within six weeks of a change going live, something observable should move. If not, we change the plan.

03 / cadence

Small habits that survive busy quarters.

One weekly ritual beats a flagship offsite. We tie practice to artefacts the business already needs, with a named owner. If it dies when the sponsor goes on leave, we did not build it right.

read the long form →

services

Three areas of work.

The same three streams you will see on services: ecosystem design and learning strategy, AI capability in teams, and judgement for how individuals use AI. Most projects start with a short discovery so you can see signal before a larger build.

Consilium: a private space for your point of view.

Consilium is a long, quiet writing practice for senior leaders who want their judgement to stay sharp. Not a course, not a dashboard. The writing is private; the clarity shows up in how you lead, decide, and show up in the work, aligned with the same signal we look for in ecosystem projects. Who pays: seats are either bought by the organisation (often bundled in a long partnership) or by the individual through the instrument’s published tiers. Not included in a typical discovery on its own.

We only open a few seats each phase on purpose.

enter consilium →
contact

Phase 04 opens Q3 2026 (conversation now).

We run two deep builds and a handful of lighter annual partnerships at a time. Full builds are scheduled forward; discovery, sparring, and fit calls still happen earlier (say so in your note). If you are considering work with us, the earliest conversation is usually the most useful. A short note about the problem is enough, or book a strategy sparring session (30 minutes, no slides) if you want a fast read on fit. Ask in the same email for a scheduling link if you prefer a calendar to threads.

Principals: Tom (APAC) · Janna (Europe) on LinkedIn

tom@midnightlabs.com.au · janna@midnightlabs.com.au for anything else.