Notes from the studio.
- Sales
Claude Sonnet 5 makes account research cheap per deal
Anthropic shipped Claude Sonnet 5 on july 24 — near-opus intelligence at Sonnet cost. For a sales lead: a research brief on every account, not just the top 20%.
- Customer support
AI for your support knowledge base: cited answers, one query
AI for your support knowledge base searches past tickets, product docs, and policies in one query and cites the source. Agents stop digging for answers.
- Operations
Worked example: an ops bot for new-hire onboarding
An illustrative worked example of an AI ops bot that routes new-hire access requests in Slack. The firm and numbers are assumptions, not a client result.
- Customer support
What Claude's self-serve hipaa switch means for support
In july 2026 Anthropic made hipaa setup a one-click toggle with a click-to-accept baa. Here is what it unblocks for a patient-support lead, and what it doesn't.
- Operations
When to replace the airtable running your ops (and when not to)
Your airtable didn't break, it outgrew the tool. Here is how to tell which two workflows are worth moving off it, and which parts to leave where they are.
- Education
Worked example: rubric-driven AI grading for courses
An illustrative worked example of rubric-driven AI grading for an online course company. The company and numbers are assumptions, not a client result.
- Finance
Claude enterprise cost controls: cap AI spend by team
Anthropic just added per-team usage and cost views to Claude enterprise. Here is how a finance lead caps the spend before fable 5 goes metered on july 13.
- Marketing
AI for marketing attribution: tie channels to revenue
An AI layer that stitches GA4, HubSpot, and Stripe so every channel ties to closed revenue. Stop paying three agencies and guessing which one works.
- Compliance
Worked example: contract review with a full audit trail
An illustrative worked example of AI contract review at a regulated finance firm. The team and numbers are assumptions, not a client result or a promise.
- Product
GPT-5.6 pricing: what Luna's cheaper tier does to cost-to-serve
OpenAI's new GPT-5.6 Luna tier is far cheaper than sol. Here is what that spread does to your AI feature's cost-to-serve before the next finance review.
- Sales
AI for lead qualification: score and route inbound
An AI layer that scores, enriches, and routes every inbound lead against your icp and CRM, so reps only open the ones worth a call. No new sdr, no chatbot.
- Finance
Worked example: ai-assisted reconciliation for month-end
An illustrative worked example of ai-assisted bank-to-ledger reconciliation for a generic finance team. Assumed numbers, not a client result.