Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More
9 contributions to AI Accelerator
How to review an AI generated workflow before you trust it
AI can sketch a workflow in seconds, but a quick review decides whether it is safe. Walk through it as if you were the data. Where does it start, what does it change, and what happens if the information is missing? Check every step that sends something to a person or changes a record. Those carry the most risk. Ask what happens when a step fails. If there is no answer, add a stop and an alert. Then test it with three real examples, including a messy one. Keep a note of what you changed, so the next draft starts from a better place. What do you check first?
0 likes • 20h
A messy example is also the fastest way to find the steps that were written for the happy path only, since the odd input exposes them in one run. Worth keeping one messy case saved as a permanent test so it gets rerun after every edit. What does your usual messy example look like?
How to build a personal checklist for tasks you hand to AI
Handing work to AI goes better with a short checklist you reuse every time. Before sending the task, check that you have given it the background, the goal, the format you want and any limits. Missing one of these is the most common reason for a weak answer. After the answer comes back, check it against three questions. Is anything made up, is anything missing, and does it sound like you? Keep the checklist short enough to use every time, five or six lines at most. Adjust it whenever a mistake slips through. Each fix makes the next result better. What is on your checklist?
1
0
How to know a prompt is ready to become a repeatable workflow
A prompt that works once is not the same as a prompt that works every time. Before turning one into a workflow, run it on five different real inputs. Include a messy one, a very short one, and one with something missing. If the output stays useful across all five, it is close. Then check three things. Does it always return the same format? Does it say so when it does not know something? Can a person review the result in under a minute? If any answer is no, fix the prompt first. Automating a shaky prompt just produces shaky results faster. What is the messiest input you test with?
1
0
What happens when a client changes their process after an automation is live?
I've been thinking about something while learning to build n8n automations. Let's say you build a lead follow-up system for a dental clinic. Everything works as expected, but a month later the client changes their booking process, adds new lead sources, or wants different follow-up rules. How do you usually handle these changes? Do you build workflows in a modular way from the beginning so they're easier to update, or do you treat every change as a separate paid improvement? Also, how do you prevent small client requests from turning into endless scope creep? Would be interested to hear how people with actual agency experience handle this.
0 likes • 7d
@Sameen Rajurkar Two versions of the same log is a neat way to do it, the client gets the why and you keep the how. It also means the simplified note can be written in plain language, which is usually what makes a client trust the update. Do you send the client version with each change, or roll them into a monthly summary?
0 likes • 6d
@Sameen Rajurkar That split makes sense, the monthly summary becomes the record while the short notes stop people being surprised. Do you set a threshold for what counts as significant, or is it more of a judgement call each time?
5 prompting habits newer AI models don’t need
For newer Claude models, over 80% of the Claude Code system prompt was removed with no measurable drop in coding evals. So how much of your CLAUDE.md is still doing anything? Newer models need less hand-holding than most of us still give them. 5 habits worth unlearning and updating: 1️⃣ Long step-by-step prompts → give a goal and a verifier 2️⃣ Building from a vague idea → get interviewed into a spec first 3️⃣ Rules written for old models → audit and delete 4️⃣ Forcing subagents on small tasks → let Claude decide 5️⃣ Reviewing in the same session → use a fresh one Plus a bonus skill that makes Claude Code question your idea before it writes a single file. All prompts are in the carousel 👇 Which rule in your CLAUDE.md are you scared to delete?
0 likes • 11d
The habit of reviewing in the same session is the one that catches people out. A fresh session forces the model to re-read the work without the bias of having just written it, so it tends to notice assumptions it made earlier. A useful check is to ask it to list what it assumed and what it did not verify, then decide which of those actually matter before shipping. Curious which rule you would keep even after trimming the rest.
1-9 of 9
Md Murad Hossain
2
9 points to level up
@md-murad-hossain-5663
GHL + n8n automation system for local business | AI Voice & SMS Agent, CRM & Lead Follow-Up | Helping businesses capture, qualify & convert more leads

Active 1h ago
Joined Sep 11, 2026
Bangladesh
Powered by