AI for ops
Ops is a generalist role — process design, tooling, internal support, the work that holds an organization together. AI is genuinely transformative here because so much of ops is structured writing: SOPs, status updates, vendor briefs, FAQs. The trap is automating processes that didn't need to exist.
What AI handles well
SOPs from how things actually get done
The problem. You ask a team how they do X. They explain it. By the time you write it down, the doc misses three steps and reads like generic process. Adoption: zero.
What AI does. Record the team explaining the process (Otter / Granola). Paste transcript into AI with the SOP-from-tutorial prompt. Output: an actual checklist that matches their workflow.
Vendor evaluation / RFP responses
The problem. Three vendors pitched the same product. Each demo was great. You need to make a defensible recommendation in 48 hours.
What AI does. Use AI to structure the comparison: features matched against your requirements, hidden costs, switching cost, security posture, integration. Output: a one-page recommendation with reasoning.
Internal newsletters worth opening
The problem. Monthly ops newsletter. Open rate is 30% and falling. People skim "exciting updates from across the team" with the same enthusiasm as a spam folder.
What AI does. Use AI to extract the 3-5 things that ACTUALLY matter to readers from your raw input. Lead with what they need to act on. Skip the corporate excitement framing.
Project status updates for stakeholders
The problem. Weekly status update is due. Slack threads, meeting notes, Linear tickets — extracting "what changed" takes longer than doing the actual work.
What AI does. Paste raw inputs into AI. Get a structured update: progress, decisions made, decisions needed, risks, next week. Stakeholders can scan in 30 seconds.
Meeting prep + agenda design
The problem. You have 6 meetings tomorrow. None has a real agenda. Half will overrun. The other half will end with "we should follow up about that."
What AI does. For each meeting: paste the goal + relevant context. AI generates a tight agenda with time boxes, the decision the meeting needs to make, and what to NOT discuss.
Templating recurring docs (briefs, kickoffs, retros)
The problem. Every ops project has the same 5 documents. You copy-paste from the last project and edit. Half the time you forget a section.
What AI does. Build templates with AI as scaffolding. The template includes ALL sections, with [TODO] markers for project-specific fill-in. Future projects inherit completeness.
Cross-functional brief writing
The problem. You need to brief leadership on a problem touching sales + CS + finance. The data lives in 4 systems. The audiences need different framings.
What AI does. Use AI to draft three audience-specific versions from one master brief. Same facts, different emphasis depending on who's reading. You edit each for tone.
Decision frameworks for "should we automate this?"
The problem. A team asks for an automation. It would save them 2 hours a week. Building it would take 2 weeks of eng time. Is it worth it?
What AI does. Use the decision framework prompt. Surface the actual question (often it's "is this process even right?" not "how do we automate it?"), reversibility, and the smallest version that earns information.
Your AI stack
Start with the foundation. Add specialized tools as the work calls for them.
Foundation LLM
Specialized add-ons
Prompts ready to use
Get started in 35 minutes
Pick an LLM and load your company context
10 minClaude Project with: org structure, tools, common processes, voice/tone guide. Now any prompt is pre-loaded with your specifics.
Run the SOP-from-tutorial prompt on the most-asked-about process
15 minPick the process people keep asking how to do. Record someone explaining it, transcribe, run through prompt. Ship the resulting checklist.
Set up a weekly status template you'll actually use
10 minBuild it once with AI. Save as a Notion template or Custom GPT. Now each week is a 5-minute fill-in instead of a 30-minute write.
Common mistakes
Generic SOPs that don't reflect actual workflow. The team you wrote it for can spot it instantly. They go back to tribal knowledge.
AI-generated newsletters with corporate excitement framing. "We're thrilled to announce" is not an upgrade over the boring update — it's an insult to readers' time.
Over-automating. Just because AI can generate a template doesn't mean every workflow needs one. Some processes work better as living conversations than as ossified docs.
Pasting confidential ops data (vendor contracts, internal financials, employee info) into public AI. Use enterprise tiers or anonymize.
Using AI for political/judgment-required decisions. "Should we restructure the team?" — AI can list pros and cons. The actual call requires reading the room, not analyzing the data.