How to Turn Felo Search Results Into a Research Deck

November 12, 2026

Felo Search exports a search result to slides in one click. What you get is a section outline with headlines and bullet summaries, which is about 30% of a deck. The remaining 70% is sourcing, one real number per slide, and cutting the sections nobody asked about.

This is the workflow I use for competitive and market decks: a 90-minute cycle that ends with something presentable, not a draft nobody trusts.

What does the Felo slide export actually produce?

One slide per answer section, each with a headline from the section title and two to four bullets paraphrased from the answer. Citations are preserved as links but usually not attached per slide.

The output has three predictable weaknesses:

  • No numbers. Bullets describe categories, not figures. Every deck needs figures.
  • Even weight. Twelve slides of equal importance, when a real deck has three points and nine supports.
  • No argument. It summarises the topic; it does not advance a claim.

Knowing this saves you from the classic mistake of exporting, restyling, and presenting a summary nobody needed.

How do you restructure the export into an argument?

Write the conclusion first, in one sentence, before touching the export. “Our pricing is 40% above the two closest competitors and customers still choose us for onboarding speed.” Now the deck has a job.

Then:

  1. Keep two or three exported sections that support the conclusion.
  2. Delete the rest. Exported sections that neither support nor complicate the claim are filler.
  3. Add one slide with the counter-argument. Decks without one get challenged in the room.
  4. Reorder so the strongest evidence lands second, not last.

That last ordering rule is not stylistic. Attention peaks at slides two and three and drops after seven.

How do you get real numbers into each slide?

Back to the citations. This is where most of the 90 minutes goes, and it is the part that makes the deck credible.

For every slide, pick one figure from a cited source and put it in the headline, not the bullet. “$12 per seat vs $19” is a headline. “Pricing comparison” is not.

Rules that keep you honest:

  • Every figure needs a source line with domain and date.
  • If two sources disagree, show both and note the gap. That is a finding.
  • Round consistently. Mixing “about 40%” and “38.7%” in one deck signals sloppiness.
  • Date every slide with the data period, not the presentation date.

Three to five verified numbers carry an entire deck. You do not need twenty.

Can you iterate a Felo search to fill deck gaps?

Yes, and it is faster than starting over. After restructuring, you will find two or three claims with no support.

Run a targeted Felo query for each gap: name the claim, name the source type you want, add the year. “Enterprise AI adoption rate survey 2026 percentage of firms” returns survey pages rather than vendor blogs.

Then re-export only when the new result adds sections you actually need. Re-exporting for one bullet wastes time and renumbers everything.

What are the limits of this workflow?

It does not work for original research. If no page contains the number, Felo cannot find it, and neither can your deck. Interviews, internal data and surveys are outside the tool’s reach.

It also struggles with fast-moving topics. A retrieval engine prefers pages with authority, which often lag breaking news by days. For anything from the last week, search directly.

The export formatting is basic. You will rebuild the visual styling in Google Slides or PowerPoint. Budget ten minutes for that, not thirty.

Who this suits

Analysts producing recurring market or competitive decks, where the topic is well documented and the bottleneck is assembly. It does not suit board-level strategy decks where the argument is yours and the sources are internal.

Quick Summary

  • Felo’s slide export gives one slide per answer section, with no numbers and no argument.
  • Write a one-sentence conclusion first, then keep only sections that support or complicate it.
  • Put one verified figure per slide in the headline, with domain and date attached.
  • Fill gaps with targeted follow-up queries instead of re-exporting the whole result.
  • Works for documented market topics; it cannot produce original research or breaking-news data.