Every RevOps leader has sat in a deal review and thought the same thing: I still don't actually know where this is going. The rep is optimistic. They always are. The fields in Salesforce are filled in, but half of them were entered on Sunday night before the Monday forecast call, and you know it. The manager spent 45 minutes in the one-on-one trying to reconstruct what happened in the last three calls instead of strategizing about what needs to happen next.
This is the operational reality of enterprise sales today. The playbook exists. The methodology — MEDDPICC, BANT, Sandler — is documented somewhere. But the gap between what's in the playbook and what's actually happening in deals, in meetings, in emails, is enormous. And nobody has a clean, evidence-backed view of where any given opportunity actually stands.
Spotlight.ai was built to close that gap. In this episode of RevOps Demos That Don't Suck, we sat down with Roi Carmel, CEO and co-founder of Spotlight, to walk through how the platform works — and why it takes a fundamentally different approach to deal inspection, asset generation, and revenue analytics than anything else in the market.
Spotlight is an enterprise sales knowledge graph that enables autonomous and agentic AI execution across the full sales motion. The platform listens to every interaction between sellers and buyers — meetings, emails, Slack messages, Zoom calls — and deconstructs that unstructured data into structured, atomic signals that map to a company's sales playbook. Those signals then drive decisions and actions: surfacing risk, guiding rep behavior, generating assets like business value assessments and sales handoff decks, and feeding analytics that reveal winning and losing patterns across the entire pipeline.
Critically, Spotlight is not a standalone tool. It lives inside Salesforce as a managed package, inside Slack as a debrief agent, and inside email. The design philosophy is that sellers and managers should never have to leave where they already work.
"We are an autonomous execution platform that works through a framework of listening, understanding, making decisions, and taking actions." — Roi Carmel
The problems Spotlight addresses are not new to RevOps. They're the same problems that have been documented in data management challenges for RevOps teams for years: garbage in, garbage out; reps entering whatever they need to enter to get through the Monday forecast call; managers spending their entire one-on-one reconstructing what happened instead of coaching what to do next.
"One of the main problems we create for ourselves as organizations is telling our reps, 'You need to have the fields ready for your Monday meeting.' Because even if they don't have an update, something will get filled. And so we're actually creating the garbage that is killing our inspection." — Roi Carmel
This is a problem that no amount of Salesforce validation rules or required fields will fix. The issue isn't the enforcement — it's that reps are being asked to translate complex, multidimensional deal interactions into CRM fields they weren't trained to populate accurately, with data that may or may not reflect what actually happened. The result is the broken approval and data quality cycle that most enterprise RevOps teams know intimately.
Spotlight's approach is to remove the rep from the data entry equation almost entirely — replacing manual input with autonomously captured, evidence-backed signals — and then use that clean, structured data to drive the analytics and decisions that RevOps actually needs.
The core of Spotlight is what Roi calls the knowledge graph: a model that understands value selling and maps every interaction to known signal structures. The platform doesn't capture text snippets and attach them to a MEDDPICC field. It deconstructs conversations into over 40 million atomic signals — each representing a specific, granular data point — and then computes answers by combining those signals.
The practical effect is precision without hallucination. Rather than asking an AI model a large, ambiguous question ("Is this person a champion?"), Spotlight breaks the question into discrete sub-questions: Do they have a personal win? Have they provided access to the economic buyer? Are they sharing internal information? Each answer is matched against the knowledge graph, and the combination produces a confident, evidence-backed result.
This is why the platform can show, with full traceability, that a specific use case was identified because a specific contact said a specific sentence on a recorded call three weeks ago — not because a rep typed it in.
Within Salesforce, Spotlight surfaces deal risk at every level of granularity — from a CRO-level weighted pipeline dashboard down to a single missing activity within a single MEDDPICC letter. The autonomous deal review agent continuously inspects every open opportunity against the company's playbook and flags gaps with evidence.
In the demo, Roi navigated from a pipeline report showing best-case opportunities, into a specific deal flagged for velocity risk ("time on current stage significantly exceeds the expected duration based on historical performance"), into the MEDDPICC view, down to a single missing activity: no compelling event had been identified for this customer. The entire path took seconds.
Managers can also filter to see what has changed since their last one-on-one with a rep — so the conversation starts from a shared, evidence-based understanding of deal status, not from "so, how's this one going?"
One distinction Spotlight draws clearly is between autonomously captured data and rep-entered data. Anything Spotlight populated from actual evidence is shown in green. Anything a rep added manually is shown in white. That distinction matters enormously in a one-on-one: a white-flagged field is a conversation starter, not a data point to trust.
For reps, the playbook is surfaced as a contextual, adaptive discovery guide rather than a static checklist. At any stage of a deal, Spotlight shows only the questions and activities relevant to that specific opportunity — filtered by deal type, use cases already identified, competitors present in the account, and signals already captured.
If compliance has never come up as a use case, the compliance discovery questions don't appear. If the competitor in the account is one where historical data shows the company wins on large migration volume, Spotlight surfaces that pattern explicitly and asks the rep to assess their position. This is the difference between a static playbook document and a dynamic, deal-aware coaching layer — a distinction that maps directly to why sales enablement so often fails to translate into rep behavior.
Spotlight's second major use case is the autonomous generation of sales assets from deal data. After a discovery conversation, the platform can produce a fully formatted business value assessment — complete with strategy alignment, current-state/desired-state analysis, quantified cost-benefit projections over one to five years, cost of delay calculations, and payback period — in the format and branding of the company's existing templates.
The same capability extends to first-meeting-to-second-meeting decks, relevant case studies surfaced based on what was heard in calls, and sales-to-CS handoff documents that articulate why a customer bought and what outcomes they're expecting. Roi's estimate: a BVA that would take a rep 10 to 20 hours to produce manually can be generated in a fraction of that time, with no copy-paste errors and no forgotten customer name from last quarter's deck.
For teams thinking about what it takes to execute a value-based sales motion at scale, this capability is significant. Business cases and BVAs are the standard for enterprise deals above a certain threshold — but they're also the assets that most reps either skip entirely or produce inconsistently.
The third use case is where Spotlight's structured data approach pays off for RevOps directly. Because every signal is broken into atomic, structured data points, the platform can surface winning and losing patterns at a level of granularity that traditional CRM reporting can't reach.
Spotlight visualizes the ICP fingerprint of top wins — which competitor combinations, use cases, and discovery answers correlate with closed-won — and maps those against losing patterns. It tracks activity-level conversion rates by stage, surfacing which specific discovery activities are most predictive of stage advancement. And it produces a rep behavior analysis that identifies who the "happy ears" cohort is (reps whose pipeline looks great but who consistently underperform on close rate) versus the sandbagging cohort (reps who undersell their own pipeline).
All of this structured data can also be exported into a BI tool of the customer's choice for custom analysis. Spotlight owns the structured data layer; the customer owns the data.
Spotlight also exposes its entire knowledge graph through the Model Context Protocol (MCP), which means companies building their own agentic AI initiatives can access Spotlight's structured deal data as a foundation. Rather than building AI agents on top of messy, unstructured CRM data, teams can use Spotlight's knowledge graph as the grounded, reliable data layer that AI revenue systems actually require.
Beyond Salesforce and Slack, Spotlight integrates with Gong, Chorus, SalesLoft, Zoom, Gainsight, Totango, Highspot, and MindTickle. HubSpot is on the roadmap; Salesforce is the current primary CRM integration.
Spotlight uses a phased maturity model for onboarding. Phase one — MEDDPICC qualification, stage-based playbook, and deal inspection — can be operational in less than a week. The out-of-the-box playbook with core qualification questions can be configured in one to two days. Phase two, which adds lighter asset generation (first-to-second-meeting decks, case studies), typically takes one to two weeks. Phase three, including full BVA and business case generation, runs two to three weeks to align on template structure, pilot with a test team, and iterate.
Roi cited Sysdig as a success story: full MEDDPICC implementation within a month, including alignment of an existing, complex playbook; value asset capability live within three additional weeks; high adoption rates from launch.
Spotlight is purpose-built for B2B enterprise sales teams running a value-based or solution sales motion. The minimum viable team size is approximately seven to ten reps — below that, the operational complexity Spotlight addresses doesn't yet exist. The sweet spot is organizations with 20 or more reps where managers can no longer be in every deal and forecast accuracy is becoming a real problem.
Current customers include Ericsson, Wiz, Sprinklr, and MongoDB — organizations with hundreds of salespeople — alongside mid-market teams in the dozens-of-reps range. MEDDPICC is the core supported framework, with full support for BANT, Sandler, and custom playbooks.
Spotlight is priced per seat within Salesforce, with three separately licensed modules: MEDDPICC qualification, value asset generation, and RevOps analytics. Each seat can carry one, two, or all three modules depending on role. There is a platform fee that is typically waived at higher contract volumes. For MCP access, pricing is per opportunity analyzed rather than per seat — designed to make agentic use cases cost-predictable.
The forecast accuracy problem in enterprise sales isn't a Salesforce problem. It isn't a methodology problem. It's a data fidelity problem: the information that flows into CRM is filtered through rep optimism, manager pressure, and the structural incentive to have something in the field before the Monday meeting. Spotlight's knowledge graph approach — deconstruct every conversation into atomic signals, structure them against the playbook, surface risk with full evidence traceability — is the most rigorous attempt to solve that underlying problem we've seen in this space.
For RevOps leaders who have watched well-designed forecasting processes collapse under the weight of bad CRM data and rep-entered optimism, Spotlight is worth a close look. The fact that it lives inside Salesforce, requires no new tool adoption from reps, and can be operational in under a week at the qualification layer removes most of the typical objections to implementation.
Request a personalized demo or learn more on the Spotlight website.
Check out our blog, join our community, and subscribe to our YouTube Channel for more insights.
