Glossary

ABCDEFGHIJKLMNOPQRSTUVWXYZ

A

Atomic Answer

A 40–80-word direct-answer paragraph placed immediately under each H2 on a docs page. The format is optimised for AI-overview citation engines that extract one paragraph per heading; reading-experience benefits follow as a side effect of the structural discipline.

B

Breakout Group

A small subset (typically 4–8 people) of an event's attendees pulled out of the main room for a focused conversation. EventIntro forms breakout groups by balancing complementary seek/offer keywords across each group rather than by topical similarity alone.

C

Cohort

A persistent group of people in EventIntro that hosts ongoing or recurring events together. A cohort is the unit of subscription and the home for cross-event member profiles — distinct from a one-off event, where attendees disperse afterwards. EventIntro is cohort-first by design.

Cohort Facilitator

The person who creates and maintains a cohort in EventIntro. Responsible for inviting members, configuring the survey, scheduling events, and reviewing the matches the platform produces. Facilitators are the primary buyer persona for the product.

Complementary Match

A pairing where person A's offerings answer person B's challenges, and vice versa. EventIntro's matching pipeline favours complementary matches over similarity-based matches because the former is more likely to produce a callable weak tie than a same-affinity meet-cute.

D

Dunbar's Number

Anthropologist Robin Dunbar's 1992 paper proposed humans can sustain stable social relationships with around 150 people. The number is descriptive, not prescriptive; for one-evening events, the cognitive ceiling for evaluating strangers is much lower.

E

Embedding

A numerical vector representation of text — in EventIntro's case, of the seek/offer keyword pairs extracted from each attendee's survey. Embeddings let the platform query for nearest neighbours in vector space, which is how complementary matches surface without a hand-curated taxonomy.

Event Intro (introduction)

The personalised, LLM-generated introduction text EventIntro writes for each match — "you should meet because you're building X and they've done Y". The introduction names what each person has to offer the other, so the conversation starts past small talk.

F

Facilitator

In general usage, anyone running a structured group conversation. In EventIntro specifically, the cohort or event organiser. See Cohort Facilitator.

G

Goal/Challenge/Offering Survey

EventIntro's fixed five-question survey, structured around what each attendee is working on (goals), where they need help (challenges), and what they can offer others. The shape is what lets the matching pipeline find complementary matches rather than just lookalikes. The questions are the same across every cohort by design — signal consistency drives match quality.

Group Formation

The process of assembling small groups (typically 4–8 people) from a larger pool of attendees. EventIntro's group-formation step takes pairwise match scores as input and produces balanced groups; group quality matters more than individual match quality for breakout sessions.

I

Icebreaker

A structured opener at the start of an event meant to help attendees relax and start talking. EventIntro's personalised introductions serve a similar role — they give matched pairs a non-trivial starting point so the icebreaker work is implicit rather than explicit.

M

Matchmaking

The end-to-end process of pairing attendees who should meet. In EventIntro: profile enrichment → keyword extraction → vector embedding → nearest-neighbour search → group formation → introduction generation. "Matchmaking" is sometimes used interchangeably with "matching"; both refer to the same pipeline.

Mastermind

A small, ongoing peer group — typically 6–12 professionals — that meets regularly to support each other's goals and accountability. Masterminds are a high-fit use case for EventIntro because the cohort-first model mirrors how masterminds already work.

N

NPS for Events

Net Promoter Score adapted to events: a single question along the lines of "how likely are you to attend the next event?" or "how likely are you to recommend this event to a peer?". EventIntro tracks this as a follow-up signal but does not yet have aggregate cross-customer data to share.

P

Pre-Event Engagement

Activity attendees do before an event starts — completing the survey, reading their match list, exchanging messages with their matches. Higher pre-event engagement correlates with deeper conversations on the day; we believe this is a structural, not coincidental, relationship.

Profile Enrichment

The first pipeline step after a member completes the survey: an LLM expands their answers into a richer profile and extracts seek/offer keyword pairs. Enrichment is what lets a five-question survey produce match-quality signal.

Psychological Safety

Amy Edmondson's 1999 work defined this as the belief that one can speak up without risk. Small, structured events tend to produce more of it than large unstructured mixers — which is why the depth of conversations differs so noticeably between the two formats.

Q

Quick-Pick

A short, gamified profile-update prompt EventIntro can send to members between events ("here are two skills — which fits you better?"). Each quick-pick refines the profile and surfaces new match candidates without asking the member to fill out the full survey again.

S

Seek/Offer Keywords

The keyword pairs the LLM extracts from each survey response — what an attendee is seeking and what they can offer. Stored as vector embeddings, queried for complementary fits. The pair shape is what distinguishes EventIntro's matching from generic similarity search.

Small Groups

In EventIntro's terminology, events of 10–100 people. The band where Dunbar-style cognitive limits don't yet break down and weak ties have room to form. The platform is built specifically for this band; we deliberately don't compete with conference platforms above 500.

Strong Ties / Weak Ties

Mark Granovetter's 1973 paper distinguished strong ties (close friends, family, daily collaborators) from weak ties (loose acquaintances). Granovetter's finding: professional opportunity flows through weak ties more often than strong ones — but only callable weak ties, not surface handshakes.

T

Top Picks

The personalised list of best matches each attendee sees in EventIntro. Each top pick comes with the LLM-generated introduction explaining why the two people should meet. Top picks are the primary attendee-facing output of the matching pipeline.

V

Vibe-Based Regrouping

A live-event facilitator tool: the host can rebalance breakout groups mid-event based on observed energy or feedback. Useful when a particular grouping has gone quiet or one attendee dominates. EventIntro surfaces the engagement signal; the facilitator's judgement is still the trigger.

Vector Search

The technique of querying a database of embeddings for nearest neighbours. EventIntro uses pgvector — a PostgreSQL extension — to store keyword embeddings and find each attendee's complementary matches. Vector search is how the platform avoids needing a hand-curated tag taxonomy.