So staying current costs hours a week, and gets done unsystematically — or quietly not at all.
The subject areas you’re accountable for, and what you actually do with them. This is what everything downstream is measured against.
Channels, shows and publications worth following in your field — across YouTube, podcasts and Substack, with Reddit and the open web for one-off searches.
Not the description or the first minute. The whole source, including transcribing audio and video so it can be searched properly.
Each item scored against your topics, with a short summary, the exact timestamp of the relevant passage, and a line stating why it matters to your work specifically.
A real one. Someone tracking agentic architecture for insurance work — the summary, the reasons and the timestamps below are the actual output.
Researcher
AI in insurance — architecture & compliance
Google I/O 2025 marks the moment the agentic protocol stack — MCP, A2A and AG-UI — crystallises into a de-facto standard for how AI agents discover tools, delegate work and stay under human control.
Why this matters to you
Directly analyses MCP and five other agentic protocols as foundational layers for enterprise AI systems, with explicit focus on security, tool integration and agent coordination — core to understanding agentic architecture in regulated enterprises like insurance.
The passages worth watching
“The substrates for agents actually shape the customer experience.”
The core claim: protocol decisions aren’t purely technical — they shape the user experience directly, and with it the process quality of insurance workflows.
“Tool access is not a feature toggle even though it’s treated that way in a lot of user interfaces. Now it is a security boundary that you’re crossing.”
Directly relevant to insurance IT: deploying MCP without governance is a compliance risk under the AI Act, DORA and GDPR.
“A procurement agent will need a supplier agent. A travel agent needs a hotel agent. A finance agent may need a tax agent. Work is distributed across owners and permissions and domains and expertise.”
Matches the reality of insurance processes exactly: claims, underwriting, legal, fraud and compliance are distributed domains — A2A is the architectural answer.
People who are accountable for being current in a specialist field.
Marketing & growth
“What’s actually working in paid acquisition this quarter”
The tactics move faster than any published source. The people who know publish on YouTube and in newsletters, at length, between everything else they talk about.
Insurance & financial services consulting
“AI in claims and underwriting, and what regulators will accept”
Architecture decisions and compliance constraints get worked out in conference talks and long technical interviews months before they reach guidance documents.
User experience & product design
“How research practice is changing as AI enters the workflow”
The useful material is practitioners talking through real cases for an hour — not the write-up that appears later with the specifics removed.
Start free, no card. Prices include VAT — what you see is what you pay.
Free forever
incl. VAT · cancel anytime
incl. VAT · cancel anytime
Paid plans aren’t live yet — Free is what you can start with today.
A research tool for people who are accountable for being current in a specialist field. You give it your topics and your role; it follows the sources your field publishes in, works through them in full, and returns a ranked briefing with summaries, timestamps and a reason each item matters to you.
You can add channels, shows and publications directly, or describe what you’re trying to learn and let the setup assistant suggest sources for that field. Researchers cover YouTube, podcasts, Substack-style newsletters and news sites. Reddit, the open web, Hacker News and GitHub are covered by Last 30 Days searches.
Every item is scored against the topics and professional intent you defined — not against keywords. That’s also what makes the “why this matters to you” line possible: it’s written for your stated question, so the same video can rank differently for two people.
A researcher’s first run takes longer than later ones, because it works through the existing back catalogue rather than just what’s new. After that it only processes what has appeared since. Audio and video have to be transcribed before they can be read, which is the slow part.
Sometimes, and that’s the point. Most of the time the briefing is enough. When it isn’t, each passage carries its timestamp — so you watch the six minutes that apply rather than the full two hours on spec.
Yes. Briefings can be delivered on a schedule instead of you opening the app, so a researcher can run in the background and simply arrive.
Yes. Free gives you 3 researchers with up to 3 sources each, 4 saved Last 30 Days searches, weekly monitoring and 20 explanations a month, covering youtube and substack newsletters. No card required.
Capacity, not access. Every source type is already in Pro (€9,90/month) — no media type is held back for Studio. Pro raises you to 10 researchers with 10 sources each and monitoring every 2 days. Studio (€29,99/month) adds daily monitoring, 25 researchers and 500 explanations. Paid plans aren’t switched on yet.
Your researchers, topics and briefings are private by default — a share link only exists if you deliberately create one. Sources are public material; CopyKit summarises and always points back to the original rather than republishing it.
No. Prices are shown including VAT — the number on the card is exactly what you’d pay, with nothing added at checkout.