Three inputs. One embedding space. Rooms ranked by projected value.
The matching engine that runs the Concierge, Sponsor matching, and Private Room guest lists is the same engine, and it is not a set of rules. It borrows the same class of technique that powers the world's most effective content and product recommendation systems: represent every person, company, and association as a vector in the same high‑dimensional space, then let proximity do the work of the traditional sales team.
What the engine actually consumes.
Structured signal
Association rosters, conference speaker lists, industry directories, capital‑stack disclosures, CRM data you connect, and open sources. All indexed and refreshed on a schedule.
Fresh context, fetched on demand
Where the structured index is thin, automated research fills in the gap in real time. Search, page fetches, and normalization into the same vector schema as everything else.
Twenty years of domain networking
The engine is seeded from twenty years of Business Development LLC's own relationship graph and refined by the vetted host marketplace's on‑the‑ground judgment. It is not a cold‑start algorithm.
Everyone lands in the same space.
You (goals, industry, seniority, geography, capital stack, tech stack, prior engagement, open interests). Companies (product, ICP, buying committee, deal history). Associations (charter, membership profile, event roster, adjacent industries). Events (topic, speakers, historical audience, room dynamics).
All four collapse into vectors of the same shape in a shared high‑dimensional space. When you take the assessment, your vector gets projected into that space. Everything else in the space is scored by distance and by the specific dimensions that matter for your goal. The result comes back with the reasoning shown: not just "top match", but "top match because your goal and this association's charter overlap on capital raising in the mid‑market energy transition, and three named principals in the association have historical fit with your buyer stage."
The engine borrows the same class of technique used by the world's best recommendation systems, applied to executive networking rather than short‑form video. What is unusual is not the math. It is the domain graph feeding it.
You see the reasoning behind every match.
Every ranked recommendation shows the top three dimensions that drove the score, the specific overlap points, and any gates you would have to clear to get into the room. If the recommendation is wrong, you can see why it was wrong and correct the input.
XODIAK attribution — the reason the engine keeps getting better.
Every intro made through the platform is logged. Every follow‑up is tracked. Every closed opportunity is written back to XODIAK with a reference to the intro that started it. The engine reads that outcome data on the way back around, and matching quality improves cohort over cohort.
Discipline is the feature.
No purchased lists
The engine is not seeded from purchased contact lists. Every person and company in the graph is there because a real signal put them there.
No black‑box scoring
Every score has a reason. If the engine cannot explain why a room was recommended, the room does not get recommended.
No public embeddings
Your embedding, your CRM data, and your engagement history stay on your workspace. Nothing leaves without an explicit export you control.
Six minutes to see the engine work on your goals.
Take the assessment. See your top association matches with the reasoning shown. Get a personalised proposal back within 24 hours.
Take the assessment