Back home
Case study · Prototype

LinkedIn Boost 

An agentic writing assistant that drafts LinkedIn content in the operator's own voice, with a human review step before anything is published.

LinkedIn Boost interface

LinkedIn Boost, the drafting workspace.

Visit LinkedIn Boost

Built with Lovable, Claude, ChatGPT.

The people

Solo founders, operators and consultants who know the platform drives pipeline but cannot sustain a daily posting cadence.

What I heard

Interviewed 12 operators to understand exactly where their content workflow breaks: ideation, drafting or the courage to publish.

The journey

Connect voice samples and goals, generate a week of drafts, review and edit in one pass, then schedule or publish with performance feedback looped back.

How I validated it

Manually ghostwrote for five operators to codify what worked before automating the workflow.

What I prioritised

Prioritised voice matching and human review over scheduling and analytics, so nothing is published without the author approving it.

Where AI does the work

An agentic prompt chain learns voice from prior posts, applies proven hook frameworks and iterates drafts against engagement heuristics.

How success was measured

Posts published per week, time to publish, engagement lift against baseline and inbound conversations attributed to the platform.

The results

Compresses a two hour content session into a ten minute review and publish loop, with referrals from active operators as the growth loop.