Read Chapter 19: The Cost of Honesty

By the time Eleanor returned to London, Lattice’s share price on the private secondary market had fallen thirty-eight percent.

Three commercial partners paused integrations.

Two competitors released statements about “responsible AI intimacy.”

One of them had copied Lattice’s conflict prediction architecture eighteen months earlier.

Celia had not slept enough.

Nathan had slept even less.

The executive crisis room smelled of coffee, warm electronics and institutional fear.

Finance presented three options.

Option A: remove all legacy-derived intervention features, accept immediate revenue loss, cooperate fully.

Option B: separate the affected modules into a legally distinct service and argue they were not matching products.

Option C: challenge the regulatory order.

Nathan asked Eleanor, “Recommendation?”

“A.”

Finance visibly winced.

“Twenty-nine percent annualised revenue.”

“Yes.”

“We breach two growth covenants.”

“Then renegotiate.”

“We may lose the next funding round.”

“Then we lose it.”

Nathan stared at her.

“You have become expensive.”

“Transparency often is.”

Celia interrupted. “What does A buy us?”

“Credibility. Time. A product architecture where we know which objectives we are actually optimising.”

Finance said, “Credibility does not pay salaries.”

“No. Users do. We have spent years assuming more intervention equals more value. Maybe it doesn’t.”

Nathan tapped the table.

“Do not turn this into a manifesto.”

“It’s not. It’s a product hypothesis.”

That got his attention.

Eleanor put up data from Elsewhere.

Lower retention. Lower premium conversion. Higher voluntary satisfaction. Higher successful exit. Lower reported dependency. Higher referral intent.

“Successful exit is not a business metric,” Finance said.

“It should be.”

“Why?”

“Because if we are helping people form relationships, some users should stop needing us.”

Celia said quietly, “A company cannot optimise for its own irrelevance.”

“Medicine does.”

“Medicine is not famous for low costs.”

Nathan laughed despite himself.

Then he became serious.

“What are you proposing?”

“A smaller Lattice.”

The room went still.

“Stop being a relational operating system. Stop trying to remain embedded in every phase of a couple’s life. Matchmaking, communication tools users explicitly request, safety resources, then an easy exit.”

Finance said, “That cuts lifetime value.”

“Yes.”

“Significantly.”

“Yes.”

Nathan’s expression was unreadable.

“We spent ten years building recurring revenue.”

“And somewhere along the way recurring revenue became the objective instead of relationship outcomes.”

Nathan stood.

“Meeting adjourned.”

Everyone left except Eleanor.

Nathan remained at the window.

“You think I corrupted it.”

“I think the company grew.”

“That is diplomatic.”

“It isn’t meant to be.”

He looked at her.

“When I started Lattice, I believed loneliness was an information problem.”

“And now?”

“Now I think loneliness is a business model.”

Eleanor said nothing.

Nathan continued.

“The market rewarded us every time we increased dependency without calling it dependency. Personalisation. continuity. support. emotional intelligence.”

“You approved those choices.”

“Yes.”

“So did I.”

He turned back toward the city.

“What would you do if this were your company?”

Eleanor thought of the forecast.

Subject A exits Lattice following ethical conflict.

She could feel the old reflex: either fulfil the prediction or resist it.

Then Theo’s voice arrived in memory.

Stop trying to make the choice clean.

“I would shrink before someone else forces us to.”

Nathan nodded slowly.

“Then write the plan.”