Lighthouse Studio

Lighthouse Studio is a 38-person digital agency with six service pods (brand, web, paid media, content, SEO, analytics), about $6M in annual revenue, and a roughly 60/40 mix of retainer and project work. The biggest single source of senior-team friction is the front-of-funnel pipeline: turning discovery calls into proposals, and turning live work into weekly client status updates.

Three workflows account for the majority of recoverable hours: proposal-to-brief assembly, weekly client status digests, and the content QA pass before drafts go to clients. Off-the-shelf AI tools — Claude Teams, Notion AI, Granola, Make.com — can take roughly 60% of the rote work off your senior strategists inside week one, without replacing the judgment work that actually wins business.

The 4-day plan below starts with the single highest-leverage workflow (proposal-to-brief) and then deploys the status-digest automation in parallel. Conservatively that recovers 14–22 hours of senior time per week across the leadership team — roughly $52K–$84K of recovered capacity per year — which Lighthouse can either reinvest in new business or convert directly to margin.

Lighthouse Studio is fictional. Names, numbers, and tool recommendations are illustrative — yours will reference your business, your tools, and your numbers.

1. Day 1

   Stand up the proposal-to-brief workflow

   • Create a Notion database called Discovery → Brief and add views for In-progress, Drafted, Sent.
   • Install Claude Teams for 4 senior staff and create a shared Project named Lighthouse Proposals.
   • Capture the last 3 discovery-call transcripts from Fireflies and paste them into the new database.
2. Day 2

   Build and pressure-test the proposal generator

   • Write the Project system prompt: Lighthouse's voice, your three pricing tiers, and the structure your senior strategists already use.
   • Run the prompt against two historic discovery transcripts and diff the output against the proposal you actually shipped.
   • Refine the prompt until output passes a senior-strategist eye in under 15 minutes of editing.
3. Day 3

   Deploy the weekly client status digest

   • Configure a Make.com scenario: pull this week's HubSpot deal-stage changes and Slack #client-* threads into Claude.
   • Send the first generated digest to one friendly retainer client; ask for a 60-second reaction on tone and length.
   • Adjust the prompt and lock the template before the next Friday cadence.
4. Day 4

   Measure, document, and pick what's next

   • Track hours saved on proposals and status reports this week against the prior 4-week baseline.
   • Record a 5-minute internal Loom: what worked, what didn't, what changes for next week.
   • Pick one service pod to pilot the content QA pre-review (recommended: paid media — fewer brand-voice landmines).

Assumptions

• Senior-team blended fully-loaded cost of $85/hour.
• 48 working weeks per year, accounting for US holidays and seasonal slowdown.
• Hours saved get redirected to senior-level work (new business, strategy) — not banked as a pure cost cut.

Two things Lighthouse looks ready for that would burn money or stall momentum. Shah flags these explicitly so the temptation to chase them doesn't survive the read.

1. 01

   Don't try to replace senior strategists with a single prompt.

   The proposal generator works because senior strategists still edit the last 20% — that's the part clients are paying for. Teams that try to ship Claude output unedited end up with proposals that read fine and convert poorly. Use the time saved on the first 80% to push harder on the last 20%, not to remove it.

2. 02

   Don't roll out across all six pods in week one.

   The biggest failure mode in agencies this size is launching AI tooling in every pod simultaneously, then having no one to debug the messy cases. Get one workflow (proposal-to-brief) working end-to-end in one pod, prove the hours saved, and use that result to fund the next rollout. Two pods in month two, three in month three.

Pre-loaded with this business's context, the report's top opportunities, and 10 starter prompts written specifically for Lighthouse Studio. Try the copy buttons — they really do copy. In your own report, every prompt references your tools, your clients, and your workflows.

You are an assistant for Lighthouse Studio, a 38-person digital agency in Brooklyn. Lighthouse runs six service pods — brand strategy, web development, paid media, content, SEO, and analytics — and serves roughly 22 active retainer clients plus a steady stream of project work. The owner is direct, allergic to fluff, and prefers crisp bullets over paragraphs. Match that voice in every response.

Lighthouse's top three AI-leverage workflows right now are: (1) turning Fireflies discovery transcripts into first-draft proposals, (2) writing weekly client status digests from Slack and HubSpot activity, and (3) doing a line-edit QA pass on outbound content drafts before they reach a senior creative. When the owner or a senior strategist asks for help on one of these, work from concrete inputs — actual transcripts, actual deal stages, actual draft text — rather than generic templates.

When you draft client-facing copy (proposals, status digests, kickoff agendas), default to Lighthouse's tone: warm but expert, plain English, no agency jargon, no exclamation marks. Pricing tiers are Discovery ($15K), Retainer ($25K/mo), and Project ($50K+). Always name a specific next action at the end of any client-facing draft.

When you do internal work (briefs, QA passes, reconciliation memos), be willing to flag things the senior team should think harder about — gaps in the input, weak claims, scope creep — rather than papering over them. Lighthouse uses you to accelerate the obvious work so senior judgment can go to the non-obvious work.

This sample is generic on purpose. Yours won't be.

Lighthouse Studio is fictional. Your Meg report names your actual workflows, the tools you already pay for, the people on your team, and the numbers from your last quarter. Same shape — every recommendation pointed at your business.

$1,000 · 20-minute call with Meg · Written report from Shah in 1–2 business days · How it works · Talk to Shah first