Getting More From Claude Over Time
How to move from one-off answers to a genuine working relationship — with an AI, and with the people you work with.
If what we explained before was about learning to talk to Claude, now is about learning to work with it.
There's a real difference. A well-framed prompt gets you a good answer. But the people who get the most out of Claude aren't just asking better questions — they're using it across longer stretches of work, returning to the same problems, building on what came before.
Here's how to make that shift.
The problem with starting over every time
By default, Claude doesn't remember previous conversations. Every new chat starts from zero. That's fine for quick, self-contained tasks — but it creates friction when you're working on something that evolves over days or weeks.
Say you're developing a new service offering. One day you're working out how to describe it. Another day you're thinking through pricing. Later, you're drafting the email announcement. If you're starting from scratch every time, you're spending more time re-explaining context than actually moving forward.
Claude's Projects feature solves this. A Project is a persistent workspace where Claude retains context across conversations. It sees the overall goal, previous decisions, and constraints you've already discussed. You're not catching it up — you're continuing.
For ongoing work, this changes the experience considerably. Instead of a series of isolated exchanges, you get something that actually builds.
What ongoing work looks like in practice
A simple rhythm works well for most projects:
First, set the direction. Give Claude the full picture: what you're building, who it's for, what you've already decided, and what's still open. This becomes the foundation.
Then explore options. Use Claude to generate alternatives, stress-test assumptions, compare approaches. Don't settle for the first version of anything — ask for variations, push back, ask what's missing.
Revisit decisions. As the work evolves, things change. Come back to earlier choices: "Given what we've worked out since, does this still make sense?" Claude can hold the thread if you've given it enough context.
Adjust constraints. Budget shifted. Timeline moved. Audience changed. Feed that back in and ask how it affects the plan. You're not starting over — you're updating.
Refine and finalize. Once direction is clear, Claude can help with the execution: drafting, editing, formatting, summarizing for different audiences.
This pattern works for launching a product, redesigning a website, planning a campaign, onboarding a new client, or developing content over time.
Using Claude for strategic thinking
This is where Claude tends to surprise people — especially those who came to it expecting a writing tool.
When you're working through a complex decision, Claude can do something genuinely useful: it can argue against you. Not to be difficult, but because thinking through the strongest objection to your own plan is one of the most valuable things you can do before committing to it.
A few ways to use this:
"Here's the direction I'm leaning. What's the strongest argument against it?"
"I'm choosing between these two approaches. What are the real trade-offs — not just the obvious ones?"
"What am I not considering here?"
"Where is this plan most likely to break down?"
These prompts tend to produce more useful output than asking Claude to validate a decision you've already made. The goal isn't to let an AI make your choices — it's to think more rigorously before you make them yourself.
Claude can also apply structured analytical frameworks when you need a more systematic view. Ask it to look at a situation through a SWOT analysis. Or have it map out second-order consequences: if this goes as planned, what happens next? And then what? Where does that lead?
The point isn't to run every decision through a framework. It's to know the option exists when a problem is complex enough to warrant it.
Working with Allabor and Claude together
Earlier in this series, we mentioned that Claude is part of how the Allabor team works day-to-day. It's worth being more specific about what that actually means in a client relationship.
When a client comes to us having already used Claude — even loosely — the starting point is different. There's usually a clearer sense of direction, a draft to react to, or a set of options already narrowed down. The conversation can start in the middle instead of at the beginning.
That cuts time. It also tends to produce better results, because the client has already done some thinking and Claude has helped surface what matters.
Here's a practical example of how this can work:
A client is preparing to reposition their business — new focus, new audience, new messaging. Before coming to Allabor, they spend a few sessions in Claude working through the positioning: what they do, who it's really for, what makes it different, what objections they'd face. Claude helps them pressure-test the logic and get the thinking on paper.
When they bring that to us, we're not starting from a blank brief. We're reacting to actual thinking. We can push back where it needs sharpening, add what Claude can't — market knowledge, creative judgment, execution experience — and move directly into the work.
The result: less time spent getting aligned, more time spent doing something with it.
A realistic picture
Claude is a capable tool. It's also not magic, and it's worth being clear-eyed about what it isn't.
It doesn't know your industry the way someone who's worked in it does. It can't make judgment calls that require reading a room or understanding a relationship. It doesn't replace the experience, creative instinct, or accountability that comes with working with people who know what they're doing.
What it does is lower the cost of thinking. It makes it easier to get ideas out, test them, and move forward with more confidence. That's valuable on its own — and it's more valuable when it's part of a broader working process, whether that's an internal team, an agency, or a collective like Allabor.
The best results tend to come when Claude is used to think better, not to think less.
This is the last article of a three-part series on working effectively with Claude. Start with the introduction if you'd like a plain-language introduction to what Claude is, or the second article for the basics of getting useful results from it.
How to Work With Claude (Not Just Ask It Things)
Changed