The old version
Monday: Google “AI agent platforms”, open twenty tabs, start reading, take notes, get distracted, come back, take more notes. Three to four hours. Tuesday: keep reading. Try to organise the findings. Realise you’re missing pricing data. Go searching. Three to four hours. Wednesday: build the deck. Write slides. Find images. Format. Polish. Five to six hours. Total: twelve to fourteen hours over three days. Most of which is finding things, not thinking about them.With Fluso
Monday, 9:00 AM. You ask:“Research the top five enterprise AI agent platforms. Compare pricing, features, target market, funding, and competitive positioning. Create a 12-slide board presentation with citations.”Five to ten minutes later, two things land in chat. A research report: market overview and sizing, profiles of the five platforms with comparison tables, pricing analysis, funding and growth data, a section on competitive advantages and vulnerabilities, and 25-plus cited sources. A presentation: twelve slides, professionally laid out, charts comparing pricing and features, key takeaways framed for board discussion, speaker notes that contain the supporting context. Monday, 9:15 AM. You read the deck. Tweak three slides where you have a stronger opinion than the model. Add a couple of insights only you would know. Monday, 10:00 AM. Done. The deck is ready. Three days early. About an hour total, including the time you actually spent thinking.
Variations
For something lighter:“Quick research on Rippling’s product, pricing, and recent funding. Five-bullet summary, no deck.”For when you want sources scoped:
“Research enterprise AI adoption trends. Sources from 2025 or later. Prefer industry analyst reports over blogs.”For combining external research with your own context:
“Research our top three competitors. Combine public information with what we’ve discussed about them in emails and Slack.”The third one is genuinely different from what you’d get from search alone, because it weaves in the things only your knowledge graph knows.
Tips that move quality
Specificity drives output quality the most. “Top five platforms compared on pricing, features, target market, and funding” beats “research AI platforms” by a lot. The model can do almost anything; what you tell it to do is what you get. Ask for the artifact, not the research. “Create a 12-slide board presentation with citations” gives you the deliverable. “Research it” gives you a report you’d then have to convert. Verify the load-bearing claims before you’re standing in front of the board. The citations are linked. Click through.Related
Skills
Research depths, video summaries, decks, documents, and the rest of the catalog.
Going deeper
Power-user patterns: projects, custom skills, prompt habits.