The fenbs blogAll articles
Page 21 of 48
- Turning Feature Requests Into Tasks With an AI AssistantForward the emails, paste the chat threads and the call notes into ChatGPT or Claude, and let it split, de-duplicate, classify and draft each request as a task, with who asked and why. Then you approve what gets filed. · 7 min read
- MCP Tools, Resources and Prompts: The Three Building BlocksAn MCP server can offer three kinds of thing: tools the model calls, resources the application reads in, and prompts the user picks. What each one is, who decides when it is used, and how hosts put them in front of you. · 7 min read
- Jira MCP for Data Center and Self-Hosted JiraAtlassian’s own MCP server is built for Atlassian Cloud. If your Jira runs on your own servers, here is what the documentation says today, what the alternatives look like, and the network and token decisions you will have to make. · 7 min read
- Using the Asana MCP Server in CursorAsana’s MCP server needs a registered app before Cursor can sign in. The four steps that get you connected, what to ask first in Agent chat, and how to keep Cursor from changing Asana without asking. · 6 min read
- Notion MCP vs Notion AI: What Is the Difference?Notion AI is the assistant inside Notion. Notion MCP lets an assistant outside Notion, such as Claude, ChatGPT or Cursor, work in your workspace. What each is for, how access and data handling differ, when to use which, and how they fit together. · 7 min read
- Managing AI Projects: A Lightweight MethodWhen the thing you are building is an AI feature, “done” stops being obvious. A lightweight method: write the success criteria first, build an evaluation set before the feature, treat every prompt and model change as a task, and accept work by its scores rather than by a demo. · 7 min read
- Using AI Agents for PR Review Without Losing OversightCopilot and Claude can both read a pull request and leave useful comments within minutes. The trick is to let them find problems without letting them decide what merges — and to write down what the review found where the next person will look. · 7 min read
- A Solo Developer’s Workflow With AI Coding AgentsWorking alone with Claude Code, Copilot or Cursor, you are the product manager, the reviewer and the whole of QA. A day-to-day workflow that keeps the speed and loses the chaos: one backlog, small tasks, a plan for each, a review before every commit, and a note for tomorrow. · 7 min read
- Linear MCP vs CLI vs API for AI AgentsThree ways an AI agent can reach Linear: the hosted MCP server, a command-line tool, or the GraphQL API with a key. How each signs in, what it costs in context, how it is limited, and which to pick for which job. · 7 min read
- Notion MCP vs an Obsidian Vault for AI AgentsA hosted Notion workspace reached through Notion’s MCP server, or a folder of Markdown files an agent reads straight off the disk. How each is reached, synced and shared, what stays private, and what agents do well with each. · 7 min read
- Cursor With the GitHub MCP ServerHow to add GitHub’s official MCP server to Cursor: the hosted server with a token read from your environment, or the local one with a browser sign-in; a read-only entry beside a read-write one; and what to ask first. · 6 min read
- MCP vs LangChain Tools: Which to BuildA LangChain tool is a function inside your agent’s process; an MCP server is a separate program any MCP host can use. How they differ, how LangChain now loads MCP tools itself, and when to write which. · 6 min read
- GitHub Copilot Agent vs Cursor Agent: Handling Multi-Step TasksBoth agents plan, edit, run commands and keep going until a task is done. The differences are in the structure around that loop: how each plans, asks before acting, lets you undo, hands work to the cloud and reads your rules. · 8 min read
- Context Engineering vs Vibe Coding: What Changes in PracticeVibe coding is about how little you read of what the AI writes. Context engineering is about how carefully you choose what it reads. They are not opposites, and moving from one to the other changes a handful of concrete habits. · 6 min read
- AI Context Files: CLAUDE.md, AGENTS.md and Cursor Rules ComparedEvery coding agent reads a standing instructions file, and each looks for a different one. Which file each tool reads, where it goes, what wins when several apply, how imports work, and a layout that keeps one source of truth. · 7 min read
- AI Governance Lifecycle: From First Approval to RetirementAI oversight continues after launch. Learn what to review as a proposed use becomes a pilot, enters daily work, changes, and eventually retires. · 5 min read
- AI Leasing Assistants: Questions, Follow-Ups, and ReviewAn AI leasing assistant can help draft answers and organize follow-ups. Keep availability, screening, accommodations, and commitments under clear human control. · 6 min read
- ChatGPT and RPA: Where Each Fits in a Work ProcessUse ChatGPT to interpret and draft, use repeatable automation for defined actions, and make the handoff between them explicit. · 5 min read
- ChatGPT for Debugging: Reproduce, Isolate, and VerifyGive ChatGPT a small reproducible failure, test competing explanations, and verify the fix. Keep the useful evidence and the lesson for the next debugging session. · 5 min read
- ChatGPT Memory Deletion Failed: What to Check NextCheck whether deletion actually failed, the screen is stale, or another source still contains the fact. Troubleshoot carefully without erasing unrelated conversations. · 5 min read
- Extract Data From a PDF With AI: Fields, Tables, and ChecksTurn PDF content into structured rows without losing the page it came from. Define the output, check difficult layouts, and keep unresolved values visible. · 6 min read
- Knowledge Management and Data Governance: Where They MeetKnowledge management helps people reuse what the organization knows. Data governance sets the responsibilities and rules for the information that knowledge depends on. · 6 min read
- Knowledge Management Roadmap Example for AI-Assisted TeamsUse this phased roadmap to test one knowledge problem, improve the handoff, and expand only when the record is useful. Copy the example and adapt its owners, checks, and review points to your team. · 5 min read
- Microsoft Knowledge Management: Documents and AI ContextChoose a clear home for Microsoft documents, team conversations, and approved answers. Keep a separate, portable work record when your team also uses assistants outside that environment. · 6 min read
- Notion AI for Research: Keep Sources and Findings ConnectedUse Notion AI to organize evidence around a specific question, inspect its citations, and separate a useful research finding from an unsupported conclusion. · 5 min read
- How to Share a Claude Design Project With OthersShare the design through its access controls, or export a deliverable for review. Keep the approved version, design-system rules, and follow-up work clear. · 6 min read
- Use AI to Analyze Data: A Practical Workflow With ChecksStart with a clear question and a clean data copy, then inspect the calculations and assumptions before turning an AI-assisted analysis into a decision. · 6 min read
- Zettelkasten Agent Memory: Linked Notes an AI Can UseBuild a small network of source-backed notes an assistant can consult. Keep evidence, interpretations, instructions, and accepted lessons distinct. · 5 min read
- AI for Market Research: Choose Tools and Check the EvidenceUse AI to find sources, organize customer feedback, and compare competitors. Choose tools by the question they help answer, then preserve the evidence and the limits behind your decision. · 7 min read
- AI for Students: Tools for Studying and Staying OrganizedChoose student AI tools by the help you need: explanations, practice, source checking, or keeping a course project organized between sessions. · 8 min read
- ChatGPT Memory Not Available: What to Check FirstMissing memory controls and missing recall need different checks. Identify the account, setting, and conversation involved before changing anything. · 5 min read
- Claude Prompts for Students: Study, Practice and FeedbackCopy Claude prompts for working through course sources, testing your understanding, and reviewing mistakes. Build a repeatable study session around your own attempts. · 5 min read
- Improve Writing With AI: Feedback You Can Act OnUse AI to identify specific problems in your draft, test possible edits, and learn from the changes you accept. Keep your meaning and voice under your control. · 5 min read
- AI Capabilities: What Assistants Can and Cannot Do at WorkAI assistants can draft, organize, compare, and use connected tools. Learn how to separate a useful suggestion from verified work and an authorized action. · 5 min read
- AI for Academic Papers: Research, Draft, and Verify Your WorkUse AI to organize an academic paper while keeping the evidence, argument, and authorship yours. Start with permission, verify sources, and document the help you used. · 5 min read
- AI in Risk and Compliance: Evidence, Reviews, and OwnershipRisk and compliance teams can use AI to prepare policy comparisons, organize evidence requests, and draft review summaries. Keep source verification, accountable judgment, and final approval with people. · 7 min read
- Claude Code Prompts for Planning, Debugging, and ReviewUse these reusable Claude Code prompts to define scope, investigate bugs, review changes, and report evidence. Each template includes the context and boundaries a coding assistant needs. · 5 min read
- Claude for Data Analysis: Work With Files and Verify ResultsGive Claude a clear dataset brief, inspect the transformations it proposes, and check the outputs before sharing them. Save the analysis rules and next checks so another session can continue. · 5 min read
- Codex CLI Memory: Instructions, Recall, and Saved SessionsCodex CLI memory includes several different ways to carry context forward. Separate repository instructions, local memories, saved sessions, and shared project progress so each has a clear job. · 5 min read
- Context Window Benchmarks: What They Actually MeasureA context window benchmark measures how well an AI uses long inputs, not just whether it accepts them. Learn to separate recall, relevance, reasoning, and results that matter for your work. · 5 min read
- Figma AI Design: From Brief to Reviewed PrototypeTurn a clear brief into useful design explorations, then review the flow, content, and accepted changes before treating a prototype as ready. · 5 min read
- How to Access ChatGPT History and Find Past WorkFind an earlier conversation through Search, check archived and locked chats, and separate missing history from memory. Then save the outcome you need next. · 5 min read
- NotebookLM for Research: Organize Sources and Check FindingsBuild a focused source collection, ask questions you can verify, and turn checked findings into useful notes and follow-up work. · 5 min read
- NotebookLM vs Notion: Research Sources or Team Knowledge?NotebookLM helps you question research sources. Notion gives you a flexible workspace for shared pages and organized knowledge. Choose by the work you need to keep. · 8 min read
- Open WebUI Context Window: Configure and Check What FitsYour usable context depends on the connected model, provider settings and everything added to the request. Check each layer before increasing a number. · 5 min read
- AI for SOPs: Turn a Real Process Into Checked InstructionsUse AI to organize observed work into a standard operating procedure, then test it with someone who performs the task. Keep the process owner responsible for missing steps, exceptions, and approval. · 6 min read
- AI Governance Best Practices for Everyday AI WorkMake responsible AI use part of the work: name an owner, check sources, review consequential outputs, and turn mistakes into better rules. · 5 min read
- AI Governance Workshop: An Agenda That Produces DecisionsUse this practical workshop agenda to agree where AI may help, where a person must approve, and who will carry each decision into daily work. · 6 min read