How Meetings Are Matched To Opportunities
Last updated: September 17, 2026
Behavior Overview
1. Overview
Every meeting that comes through Spotlight.ai goes through a multi-step process to determine two things: which account (company) the meeting belongs to, and - if that account has more than one active deal - which specific deal it relates to. The process moves through several checks, escalating from simple, direct logic to more sophisticated AI-based reasoning only as needed.
2. How the Account Is Identified
Spotlight.ai starts with what's already known about the meeting - the meeting title and the companies represented by the attendees - and uses that to determine which account in your CRM the meeting belongs to. This happens in stages:
Gathering context: attendee email domains are compared against your own company's domain to separate internal participants from the customer side, and the customer's company name is inferred from the title and those domains.
Searching your CRM: that company name is used to find candidate accounts already in your system so the process only ever considers accounts you actually have on file.
Narrowing to relevant accounts: candidates that don't currently have any active opportunities are omitted, since a meeting can only sensibly be matched to a live opportunity.
3. Confidence-Based Escalation for Account Matching
Rather than applying AI to every meeting uniformly, Spotlight.ai uses a multi-step approach:
Level | When it applies | How the account is chosen |
Direct match | The candidate account is unambiguous - for example, only one plausible account was found, or one candidate is clearly and distinctly the best fit. | Resolved automatically with straightforward logic; no AI interpretation needed. |
AI-assisted match | More than one candidate account is plausible and no single one clearly stands out. | Spotlight.ai applies AI-based reasoning over the meeting title, attendees, and candidate accounts to judge the most probable one, cross-checking it against real deal activity before finalizing. |
Attendee-based fallback | Even AI-based reasoning can't confidently settle on one account. | As a last resort, Spotlight.ai checks which candidate account the actual meeting attendees have a real, existing relationship with, and only proceeds if this points to exactly one account. |
This escalation means the majority of clear-cut meetings are matched instantly through simple logic, and AI reasoning is reserved for the genuinely ambiguous middle ground.
4. How the Right Deal Is Selected
Situation | What happens |
The matched account has one active deal | That deal is used directly - there's nothing to choose between, so no AI interpretation is needed. |
The matched account has more than one active deal at the same time | Spotlight.ai applies AI-based reasoning, informed by the categories of information in Section 5, to determine which deal the meeting most likely relates to. |
5. Categories of Information Considered for Deal Selection
When a customer has multiple active deals, picking the right one takes more than just knowing the company - several deals can share the same account. To make that judgment, Spotlight.ai brings together information about the conversation itself and information about each candidate deal:
Category | What it captures |
Meeting content | What was actually discussed - drawn from the meeting title and a summary of the conversation, so the match reflects the substance of the call, not just who attended. |
Deal identity | How each candidate deal is named and which account it sits under, to distinguish between multiple opportunities on the same account. |
Deal stage & status | Where each candidate deal currently sits in the pipeline, and whether it's active. |
Deal ownership | Who owns each candidate deal, on both the deal and account side - useful context when a meeting's attendees line up with a specific owner. |
Deal commercial profile | The scale and forecast category of each candidate deal. |
Deal timing | Key dates associated with each candidate deal, such as when it was created and when it's expected to close. |
Deal type | The nature of the opportunity (e.g. new business vs. expansion), where relevant to distinguishing candidates. |
Spotlight.ai weighs these categories together to judge which deal the conversation most plausibly belongs to, rather than relying on any single category in isolation. It does not use information outside of what's already stored on the deal and account records in your CRM.
6. Illustrative Example
Suppose a call takes place with contacts from "Acme Corp," and Acme Corp has two active deals - an initial deployment deal in an early stage, and a separate expansion deal further along in the pipeline. Spotlight reads the meeting title and a summary of what was covered. If the conversation centered on rolling out the initial deployment, that content - combined with the relative stage and timing of each candidate deal - points Spotlight toward the deployment deal rather than the expansion deal, even though both belong to the same account.
7. What Happens When No Confident Match Exists
Spotlight.ai does not force or guess a match at any stage of this process. If the available information isn't sufficient to confidently identify an account or a deal, the meeting is simply left unmatched rather than being linked incorrectly.
Why this matters: An unmatched meeting is expected, intentional behavior in ambiguous situations - it reflects a deliberate choice to prioritize accuracy over forcing a match, not a system failure.
8. Frequently Asked Questions
Can the same logic pick the wrong deal? Yes, in principle - this is why the process is designed to leave a meeting unmatched rather than force a low-confidence pick, and why account/deal data quality (clear naming, up-to-date stages) directly improves match accuracy.
Does this rely on any data outside our CRM and calendar? No - everything considered comes from your own connected systems.
Is the process consistent across every meeting? The categories of information considered are consistent, but which level of matching applies (direct, AI-assisted, or fallback) depends on how clear-cut each individual meeting is.
9. Data Sources and Scope
All matching draws only on information already present in your connected CRM and calendar/meeting systems - no external data sources are used, and no new data categories are introduced beyond what your teams already work with day to day.