- AI Business Insights
- Posts
- 12 Features, most use two
12 Features, most use two
Google with better manners
In partnership with
Open Claude. Type a question. Get a nice answer. Close the tab.
That was the entire shape of my relationship with it for longer than I want to admit, and it left me quietly wondering why everyone online calls this a superpower when it mostly felt like a search engine with better manners.
Then I read a LinkedIn post that named the problem so cleanly I have not stopped thinking about it.
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
For its first CTV campaign, Jennifer Aniston’s DTC haircare brand LolaVie had a few non-negotiables. The campaign had to be simple. It had to demonstrate measurable impact. And it had to be full-funnel.
LolaVie used Roku Ads Manager to test and optimize creatives — reaching millions of potential customers at all stages of their purchase journeys. Roku Ads Manager helped the brand convey LolaVie’s playful voice while helping drive omnichannel sales across both ecommerce and retail touchpoints.
The campaign included an Action Ad overlay that let viewers shop directly from their TVs by clicking OK on their Roku remote. This guided them to the website to buy LolaVie products.
Discover how Roku Ads Manager helped LolaVie drive big sales and customer growth with self-serve TV ads.
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
*Ad
The tool was never the issue
The author spent months working out which tools actually moved the needle and which ones just manufactured busywork. Claude landed near the top for them, but only after they understood what it could really do.
Their framing is that Claude is not a chatbot. It is infrastructure for how you think, build, and work.
Most people use one or two features. The founders and operators who feel genuinely ahead are using six or more, and they argue that is not a coincidence.
So here are the twelve, each with the reason it matters rather than just the label.
The first four are the ones you have probably met
Long document analysis reads up to one million tokens. Drop in a contract, a full codebase, or a dense research paper and pull out what actually matters rather than skimming and hoping.
Writing and content creation covers emails, proposals, posts, and scripts. The instruction the author attaches is to define audience, tone, and length up front, which is the difference between usable output and mush.
Claude Code connects straight to your codebase. It finds bugs, refactors files, writes tests, and reviews pull requests. Add a
CLAUDE.mdfile with your conventions so it follows your rules instead of inventing its own.Building artifacts turns outputs into live web pages, React components, charts, and calculators inside the chat window. Nothing gets copied out into a separate editor to become real.
The middle four are where it turns into a workspace
Projects store instructions, files, and memory across sessions. The suggestion is to treat each one as a dedicated workspace rather than starting cold every single time.
Extended thinking raises the reasoning effort on hard problems. Legal analysis, financial models, and genuinely nasty debugging are where that extra depth pays for itself.
MCP connectors plug Claude into Gmail, Google Drive, Slack, GitHub, Notion, and Jira. As the author puts it, this removes the copy and paste tax entirely.
The Cowork desktop agent gives Claude access to your local folders, where it reads, edits, and organizes files on its own.
The GTM Playbook Behind Warmly's Acquisition
Warmly ran pipeline, outreach, and lead scoring on autopilot for hundreds of startups — before a single sales hire.
HubSpot acquired them for it. Now the cofounders are walking you through the exact system, live, before they disappear into product. Join the Builder Session on August 12.
Leave with an agentic GTM stack you can replicate this week. Plus HubSpot Credits when you join HubSpot for Startups.
*Ad
The last four are the ones almost nobody opens
Web search and research searches, summarizes, and cites sources in real time. The tip attached here is to hand it a research goal rather than a search query.
Image and vision analysis takes screenshots, charts, diagrams, and photos and gives you real analysis back instead of a description.
Multi-agent workflows run several specialist agents in parallel. Best used on recurring, structured pipelines rather than one off questions.
Voice mode works on iOS, Android, and web, and it is genuinely good for brainstorming when you are nowhere near a desk.
Nobody is hiding any of this from you
That is the part I found clarifying. These features are not gated behind some secret tier. They are spread across the product, and if you only ever open a chat window, you never bump into them.
Extended thinking, Projects, MCP connectors, the desktop agent. Those four are the ones that convert a question and answer tool into an actual work system, and they are the four I had never touched.
Look at the pattern hiding in the list
Almost every item on the author's list arrives with a small operating instruction attached.
Define audience and length before writing. Add a CLAUDE.md for your code conventions. Give research a goal instead of a keyword. Treat each Project as a workspace.
That is the whole gap between poking at a chatbot and running infrastructure. Not the feature. The rule you bring to it.
Two Minutes to Know What Slow Billing Is Costing You
Most SaaS finance teams know their billing process is slow.Most SaaS finance teams know their billing process is slow. Few know what it's costing them.
The Tabs Billing Lag Calculator puts a dollar figure on it in two minutes — benchmarked against top SaaS companies.
*Ad
You do not need all twelve by Friday
Start with your biggest time sink. If you are drowning in long documents, begin with long document analysis. If your tools refuse to talk to each other, begin with MCP connectors.
Add one operating rule per feature and learn that feature properly before chasing the next one. Scope beats novelty, which is the author's entire argument compressed into three words.
Then move from asking to delegating. Voice mode and web search are asking tools. Projects, Claude Code, and multi-agent workflows are for handing off real work that repeats.
Finally, count your number. Track how many of the twelve you genuinely use, because the claim is that the people who feel ahead are past six. That is a concrete benchmark, and it is uncomfortable in a useful way.
I counted three and winced
I ran the exercise on myself before writing this, and I got to three. Long documents, writing, and a bit of code review. Everything from Projects onward was theoretical for me.
What changed is not that I discovered a hidden feature. It is that I stopped treating Claude as a bar I type into and started treating it as a system I build around. Long documents, code, connected apps, local files, parallel agents. Different jobs, one place.
So do the count tonight. Read back through the twelve, mark the ones you have actually used in the last month, and then open the single highest one on the list you have never touched. If your number is one or two, there is an enormous amount of room sitting there unused. And if someone on your team is still using it purely for chat, forward them the list.