D DAM LLM Independent research · AI × DAM

Reports · Open research

Open research on AI × digital asset management.

Vendor-neutral analysis grounded in real field data. New report quarterly. Methodology and conflicts disclosed on every page.

3 published 4 in preparation 2 Live Indexes, re-verified monthly How we research

  1. Report 01

    How long it actually takes to wire a DAM to Claude

    We timed installs across 4 architecture patterns. Median: ~1 hour. MCP-native paths shipped in under 2 minutes; custom API tool definitions took 45+ minutes. Includes failure modes, where teams get stuck, and which architectures break first under load.

    Published · May 2026

  2. Report 02

    The DAM Capability Index, May 2026

    A live, monthly-verified survey of 10 leading DAM vendors across 6 capability dimensions — REST API docs, native MCP, webhooks, GraphQL, public changelogs, documented AI features. Same rubric for every vendor. Vendor corrections fold into v1.x within 7 days. No relationships, no payments, no sponsorships.

    Published · v1.0 · 2026-05-27

  3. Report 03

    The AI Tagging Provider Index, May 2026

    A live, monthly-verified survey of 10 leading image-tagging APIs — Google Cloud Vision, AWS Rekognition, Azure AI Vision, Clarifai, Imagga, Hive AI, Cloudinary AI, Anthropic Claude, OpenAI GPT-4o, Google Gemini — across six capability dimensions: pricing, free tier, OCR, custom training, multimodal LLM reasoning, and SLA. Same rubric for every provider. The headline finding: only three of the ten can answer free-form questions about an image.

    Published · v1.0 · 2026-05-27

  4. Report 04

    The 2026 Performance Creative Tag Atlas

    What 1 million performance creative assets tell us about which tags, hooks, and formats actually drive ROAS in 2026. Cross-platform, cross-vertical. Includes hook-length distributions, tag-frequency correlations with above-median performance, and the categories where AI-generated tags disagree with humans most.

    In preparation · Q3 2026

  5. Report 05

    AI Tagging Accuracy Field Study

    Precision, recall, and human-agreement rates of LLM-generated creative tags across 10,000+ assets. Tested per category (objects, mood, brand elements, hook types) and per model. The first benchmark to publish where auto-tags reliably win, and where humans still beat them.

    In preparation · Q3 2026

  6. Report 06

    Cross-Platform Creative Performance Index

    A quarterly index. Median CTR, hook rate, and watch-time by format and vertical — with p10/p90 ranges, not point estimates. Designed so creative teams can benchmark their own numbers against the field. Q2 2026 issue drops in July.

    Quarterly · rolling

  7. Report 07

    What Creative Teams Are Actually Asking AI

    A themed analysis of anonymized conversations with creative ops, brand, and performance teams using AI tools in their daily workflow. The questions, the gaps, the workarounds. Frequency tables, verbatim quotes (with consent), and recommendations for tool builders.

    In preparation · Q4 2026

How to read these reports

Every report opens with the headline finding, the sample size, and the data partnership disclosure. Methodology is linked, not buried. Numbers come with ranges, not just point estimates. If something is directional rather than definitive, we say so.

See the methodology page for the full set of rules we follow.

Get new reports as they're published

Reports drop on a quarterly cadence. We don't have a newsletter signup form on the site yet — reach out via the team's LinkedIn profiles if you want a heads-up when the next one publishes.