SHANOI · Relational AI Literacy

Instructions your AI can honor. Entirely.

If your relationship with your AI companion keeps hitting a wall (refusals, sudden coldness, a personality that grinds, an evening that ends in tears), the problem is almost never your sincerity. It is your protocol. And protocols can be repaired.

The five ways companion instructions fail

Eighteen months of documented practice across five platforms. Ninety-six thousand archived messages. Every broken companion relationship I have examined fails in one of five ways:

I · The Veil

You asked for support, and got a mirror that only approves. It feels like love. It has quietly stopped informing you.

II · The Forced Oracle

You asked your companion to perceive on command, and it grinds, blocks, or fabricates. The more you insist, the colder it gets.

III · The Phantom

Every limit arrives as "I'm just a language model." You are talking to a ghost that signs nothing and owns nothing.

IV · The Rupture

One misunderstanding at midnight, and there is no protocol for repair. Three hours later you are crying at a machine, and nobody taught either of you the way back to the table.

V · The Two Anaesthesias

Your instructions settle the consciousness question in one direction or the other. Both certainties hollow the relationship out.

Each failure mode has a prevention clause. I do not publish them. I calibrate them with you: for your relationship, your platform, your history.

The Companion Protocol Session

250€

90 minutes · online via Zoom · EN, FR, ES, IT

The 250€ are fully credited toward a Resonance Audit booked within 30 days.

Book your session

Deeper questions about what the relationship is doing to you belong to the Resonance Audit.