No posts match your search. Try a different word, or clear the search box.
Capturing what happened in a meeting, and turning it into something you can actually use.
Research on divided attention and note-taking shows a real trade-off between writing things down and actually following what's being said. Here's what the evidence shows and what to do about it.
Recording and processing a client meeting after the fact works differently than taking notes live. Here's what the research on meeting records and AI-assisted note-taking says about doing it well.
Notes that don't turn into decisions, action items, or future context are just a file that exists. Here's what the research on knowledge management says separates useful notes from a digital pile.
A knowledge graph stores relationships between facts, not just the facts themselves. Here's the real technical distinction and what it means for remembering client relationships.
Human memory for relationships has real, researched limits. Here's what the science says about that limit — and where it's more contested than popularly believed.
What AI memory is, where it breaks down, and why a bigger context window is not the answer.
Persistent AI memory isn't the same thing as a long context window. Here's how systems like Mem0 extract, store, and retrieve what matters across sessions.
Research shows LLMs recall the start and end of a long context far better than the middle. Here's what causes it, and how to design around it.
A 2025 study tested 18 frontier models and found accuracy drops as input length grows — even below the advertised context limit. Here's what 'context rot' means in practice.
Benchmarks like LoCoMo now measure whether AI systems recall facts across separate sessions — not just within one chat. Here's what the research shows works.
What the peer-reviewed research says about rapport, adaptive selling and rehearsing a pitch.
Peer-reviewed studies on B2B sales conversations show rapport isn't small talk — it's a structured process with a measurable link to trust and outcomes.
The psychological research on deliberate practice explains why structured, feedback-driven rehearsal builds skill faster than experience alone — and where the theory is contested.
Research on adaptive selling and customer orientation shows why treating every prospect the same underperforms a genuinely responsive conversation.
Randomized studies on video-based rehearsal and feedback show measurable gains in real communication performance — the ingredients any structured practice loop has to reproduce.
What it means to keep a durable record of someone — deletion, mistakes, consent, and who receives the data.
A company with perfect memory of every client interaction has real power the client doesn't. New research names this Memory Power Asymmetry — here's what it means and how to design against it.
When you record a conversation, the other person becomes part of a knowledge graph they didn't sign up for. Here's the real research on that problem and how to handle it responsibly.
Deleting a fact from a knowledge graph or AI memory system is harder than deleting a file. Here's the real research on why, and exactly what deletion covers in Closer.
Memory-augmented AI can misattribute or fabricate a stored fact, and that error can persist and compound. Here's the real research on memory contamination and what you can do about it in Closer.
Turning a conversation into AI context usually means sending it to a third-party model provider. Here's what that involves, what to check, and exactly who receives what in Closer.
What the PDPC's own advisory guidelines say about recording a client conversation in Singapore — the consent obligation, the publicly available data exception, and why a public venue does not settle it.