AI Discovery
Playbook

AI answer engines rarely research a person live. They lean on what's already been written and structured — Wikipedia, IMDb, licensed press, structured data — and reassemble it into a confident-sounding answer. If those sources are thin or inconsistent, the AI's portrait will be thin or inconsistent too, no matter how strong the actual filmography is.

This is a distinct discipline now — Answer Engine Optimization, or AEO — sitting alongside, not replacing, the usual site and press work.

The pattern: nearly everything describing Aleksandar is self-authored — his own site, his own IMDb text, his own LinkedIn. That's ordinary for an independent filmmaker. But AI systems weight independent, third-party confirmation heavily when deciding how confidently to state a fact — and volume of self-written bios doesn't substitute for that. A handful of outside citations does more work than dozens of near-identical self-authored ones.

AI models get confused by drift — the same person described slightly differently on five platforms reads as five weaker signals instead of one strong one. The fix is a single canonical fact set, repeated verbatim everywhere:

• Roles — director, cinematographer, writer, producer (add “art director” explicitly if that's a title he wants surfaced — it isn't in his existing bios yet)
• Base — New York City; originally from Serbia
• Studio — Kostic Films LLC
• Signature themes — perception, memory, consciousness, technology's effect on human connection
• Key works — The World of Vija Vetra (2008/2010), Persistence of Memory (2012), Parallel Dreams (2015), Allure (2014, DP/producer), Artsanity, Capio (upcoming feature)
• Collaborators — Joan La Barbara, Morton Subotnick, Marina Abramović, Javier Pérez de Cuéllar, Hilly Kristal
• Press — New York Times, LA Times, Hollywood Reporter (cinematography on Allure)

Then write it once, at three lengths, and reuse word-for-word across every platform bio field:

Happy to draft all three, word for word, whenever useful.

1. Schema.org markup — Person schema (name, jobTitle, worksFor, sameAs) plus Movie/CreativeWork schema per film. The single highest-leverage technical fix.
2. An llms.txt file at the site root — an emerging convention giving AI crawlers a clean summary of who he is and where the key pages live.
3. One consolidated About page holding the long canonical bio, rather than fragments scattered project by project.
4. Basic upkeep — current copyright year, fast load, mobile rendering; a dated-looking site is a soft signal that a profile isn't actively maintained.
5. A real Press page — the NYT/LA Times/THR mentions, linked, not just asserted.

• Wikipedia notability review — with NYT, LA Times, THR coverage and selections at Tallinn Black Nights and MaerzMusik, a page may already be within reach. It's one of the strongest single levers in AI visibility.
• A Wikidata entity — machine-readable facts that help disambiguate him from an unrelated "Aleksandar Kostic" credited on Gucha: Distant Trumpet, and from the well-known Serbian director Aleksandar Petrović.
• The actual NYT / LA Times / THR pieces, tracked down and linked — a dated citation outweighs an unsourced "praised by."
• Festival databases kept current and consistently spelled — PÖFF Tallinn Black Nights, MaerzMusik, Queens World Film Festival.
• One independent interview — No Film School, Filmmaker Magazine, a festival blog — creates a quotable, third-party source AI systems favor over owned platforms.

AI retrieval increasingly favors recent, specific, dated content over static bios. A light, sustainable rhythm:

• Short, dated updates on Capio's progress — pre-production notes, casting, festival submissions.
• Consistent cross-linking between site, IMDb, LinkedIn, and Vimeo, so each confirms the others.
• A current "Known for" line on IMDb as Capio advances — this field carries real weight in AI summaries.

Quarterly, or after any major release or press hit, ask ChatGPT, Claude, Perplexity, and Google's AI Overview: “Who is Aleksandar Kostic?” and “What films has Aleksandar Kostic directed?” Watch for name confusion, stale credits, missing recent work, or dropped press mentions — and for Capio beginning to appear, which will confirm the fresh-content push in Reel 06 is working.

• Take 1Shoot firstLock the canonical bio and sync it across site, IMDb, LinkedIn, Vimeo.
• Take 2Shoot firstAdd Person and CreativeWork schema markup to the site.
• Take 3Shoot firstBuild a Press page with real, working links to the NYT / LA Times / THR coverage.
• Take 4PickupsAssess and pursue Wikipedia notability; file a Wikidata entity either way.
• Take 5PickupsAdd an llms.txt file to the site root.
• Take 6PickupsLand one or two independent interviews or features tied to Capio.
• Take 7OngoingRun the quarterly AI-answer check from Reel 07.
• Take 8OngoingKeep "known for" and recent-work fields current as Capio develops.