Rolled drawings, a black binder and a pencil on a plywood table in front of open stud walls, with a bundle of blue cable run through the studs to a wall box.

Technology often enters a residential project after the rooms, millwork, and mechanical systems have taken shape. By then the decisions that matter have been made by default.

Many teams do better. When it slips, I would rather have the boring conversation early: seven decisions worth settling while the drawings can still absorb them.

1. Define the experience before selecting equipment

Two documents govern it, and they differ. Design intent is the goal. A control narrative is the mechanism: which conditions produce which outputs, including when a device, the network, or the power is missing. Intent without a narrative produces a system nobody can test. A narrative without intent produces a system that works exactly as specified and feels wrong.

2. Coordinate the network as building infrastructure

Lighting, entry, cameras, and shades may all depend on the network, so access points, cable pathways, and service clearance belong on a drawing, not a purchase order.

Then decide what survives an internet outage, a different question from a power outage. Through the first, control platforms that process locally keep their core functions running; their cloud features stop. Functions that route through a vendor's cloud, common in consumer locks and thermostats, fail with the connection.

Give the equipment a home with ventilation and access, not a closet that later fills with coats.

3. Coordinate lighting and controls together

How a fixture dims changes the electrical scope. How many control points a room gets drives back box locations and wall elevations. Keypads are visible hardware with finishes and lead times, not a device count.

4. Coordinate shades, windows, and millwork as one detail

Fabric and hembar are the designer's call. Ours is power. Wired motors need cabling in the pocket while the walls are open, low voltage or line voltage depending on the motor. Battery motors move the cost into maintenance: who reaches them, and how.

The openness factor of the fabric drives glare. Window dimensions constrain shade width, and the pocket, roll, brackets, and clearances are the millworker's problem before they are ours. Either power choice is defensible. Discovering it after the ceiling detail is drawn is not.

5. Treat doors and access as a complete assembly

One door can involve the architect, hardware consultant, millworker, electrician, security contractor, and three systems that have never spoken: access control, intercom, and automation. For each opening that matters, agree on operation, guest entry, cabling, behavior when the network or power is out, and which system owns each function.

Anything electrified at an opening is also a hardware and code conversation, one for the architect and hardware consultant, not a technology vendor working alone.

Door position sensing gets missed because it belongs in the hardware schedule. Specified there, the contact is concealed. Added later, it is usually surface mounted on a door somebody chose carefully.

6. Design sensing around the space, not around gadgets

The question is not where sensors can go. It is what the house has to know, and where that must come from to be true.

A bedroom needs temperature from where people sleep, not the hallway thermostat. Leak detection goes on the floor under the equipment, not on the wall the installer can reach. Every answer is a mounting location, power or a battery someone has to reach, and, where detection geometry matters, a sight line. All cheap on a drawing, expensive in a finished ceiling.

7. Decide what the client receives at handoff

Expect an equipment schedule, a network diagram, configuration documentation, client-held credentials, backups, and an Operations Guide written for the household, not a technician. Ask about it before anyone quotes the work, along with five boring questions.

Commissioning proves it: verify the house against the control narrative from decision one and the agreed requirements, document the deficiencies, and require the corrections before acceptance. A system is not done because everything turns on.

The drawing set is the first control surface

In year four, the next person can still tell what was designed, and why. That is coordination, not procurement.

RyzaLab designs and integrates residential technology in New York and Austin. The Blueprint is where these decisions get written down: a paid assessment and a written plan.