How Supermetrics rolled out an AI sales co-worker to 50 sellers, and hit 96% weekly active usage in 30 days

Supermetrics, the marketing intelligence platform that processes 15% of global online ad spend for 200,000+ businesses, runs a complex global sales org: 50 sellers across EMEA, Americas, and JAPAC, selling into commercial and enterprise accounts where deals involve procurement cycles, multiple stakeholders, and board-level budget approval.

The problem wasn't a shortage of tools or data. HubSpot, Salesforce, Gong, and a growing list of AI products produced plenty of both. The gap was in two places. The signal that should drive a rep's next move: this deal, this stakeholder, now, was buried inside dashboards, transcripts, and notification feeds. And even once a rep found it, acting on it was hours of work: drafting the email, building the business case, prepping the briefing. Every tool was reactive by design: surface information, wait for the rep to do the synthesis and the execution. In a complex sales environment, both costs added up to lost deals.

Ollie closed both gaps by being proactive: not a dashboard to consult, but a co-worker that identifies what needs to happen, does the work, and delivers ready work product (a drafted email, a business case, a meeting briefing) directly in Slack. After starting with a single team, Supermetrics rolled Ollie out globally. The Chief Sales Officer who started as a skeptic personally stood on the all-hands to mandate it for all 50 sellers. Within 30 days of global launch: 96% of the team active weekly, thousands of customer interactions analyzed and acted on, and reps selling on financial impact instead of product features.

Selling into commercial and enterprise accounts means deals are won or lost on execution: how fast you follow up after a key meeting, whether your champion has the right financial justification to bring to the CFO, whether you spotted the procurement delay before it became a 60-day stall. That kind of selling requires real-time deal intelligence, not raw data sitting across systems that no one has time to cross-reference between calls.

The Supermetrics sales stack (HubSpot, Salesforce, Gong, plus a growing list of AI tools) held every signal the team needed. But every tool in that stack was built the same way: surface information, wait for the rep to act. A dashboard showing a stale deal doesn't re-engage it. A transcript doesn't write the follow-up. A health score drop doesn't draft the recovery email. The rep still had to find the insight, interpret it, decide what to do, and then go do it across multiple platforms.

Nowhere was this gap more costly than in deal execution. Take a business case: done properly, it means going through every email thread and call recording to map the prospect's specific pains, then cross-referencing 70+ customer case studies to find the most relevant proof points, then writing it all up in a financially compelling document tailored to that specific account. The information existed. But doing it properly would take hours, which meant that in practice, almost no rep did it. The same applied to mutual action plans and champion one-pagers: valuable assets that moved deals forward, almost never produced because the effort required to make them genuinely good was simply not viable at the pace enterprise sales demands.

The first team identified the blocker fast: if Ollie required a new destination, it would be ignored. The solution was to stop treating Ollie as a platform and start delivering it as a presence inside the tools reps already used: Slack and Salesforce.

But the deeper shift wasn't about delivery channel. It was about what Ollie actually does. Every other tool in the stack, including the AI tools, was reactive: it waited for a rep to ask a question, pull a report, or check a dashboard. Ollie is proactive. It monitors deals continuously, identifies what needs to happen next, does the work, and delivers a ready output: a briefing before the call, a drafted email after it, a business case when the deal reaches CFO stage, a re-engagement message when a deal goes quiet. The rep doesn't consult a dashboard and then decide what to do. The work is already done.

Ed Barrett, Chief Sales Officer, entered the evaluation with a clear thesis: execution gaps are management problems, not tooling problems. He'd seen too many software purchases add process without changing behavior.

What shifted wasn't a demo. It was the Q4 win/loss analysis: Ollie's analysis of over 5,000 individual customer interactions (emails, notes, and call recordings) that revealed how much revenue Supermetrics was leaving on the table due to execution gaps: slow follow-up, missed stakeholder signals, deals lost to competitors while sitting in procurement. The question changed from "do we need this?" to "what does it cost us not to have it?"

Within a few weeks, Ed's framing had transformed from skeptic to sponsor. When the time came for global onboarding (three sessions across EMEA, JAPAC, and Americas), Ed personally led the final session. For 50 sellers watching the most senior sales leader in the company publicly sponsor an AI co-worker rollout, that signal carried farther than any enablement deck could.

That's the part most AI rollouts skip. New tools typically enter sales orgs through enablement: a launch session, a Slack channel, optional adoption. The early adopters lean in, the long tail stays on existing workflows, and six months later the renewal becomes a question. The 96% weekly active number that defines this rollout isn't really a product outcome. It's what happens when the most senior sales leader in the building publicly owns the change, and the org reads that ownership as the new standard.

Ollie gives managers something Gong, Salesforce, and dashboards cannot: not just what’s happening in the pipeline, but why. And a direct line from insight to team action.

Turning insight into team action

Managers can also automate the operational side through Ollie’s task engine: scheduled CRM hygiene nudges, stale deal alerts routed to the right rep, and direct follow-up on specific situations Ollie flags. That moves management time away from chasing data and toward the coaching conversations the data enables.

Optivian's MCP connector makes Ollie's full deal intelligence available directly inside Claude and other AI tools, removing the assembly step that previously made deal-specific AI work impractical at scale. Instead of copying across email threads, transcripts, and deal notes by hand, Claude calls Optivian directly and retrieves a pre-synthesised picture: health score with directional trend, champion qualification with behavioural evidence, blockers categorised by severity, decision process already mapped. The rep moves from "I have data sources, ask the right questions of them" to "I have a deal, work on it."

Emil Berglund, Sales Manager EMEA at Supermetrics, ran a direct comparison of the same task using raw Salesforce, Gong, and Gmail integrations versus the Optivian connector. The raw-integrations run required six discrete tool rounds to assemble the deal picture and still missed email context that would have left a material gap in the analysis. The Optivian run resolved the full context in a single call, surfacing what the other approach missed entirely, and produced a full three-tab cowork sheet (MSP, Time Savings ROI, Business Case Calculator) in one shot without iteration.

Supermetrics has since rolled the connector out to their entire team, with product knowledge and other internal systems connected to Claude alongside Optivian, giving every rep and manager a single AI session that combines deal-specific intelligence with the company's full institutional knowledge.

In those 30 days, the team analyzed and acted on thousands of customer interactions. They reactivated 72 stale deals through 80 Ollie-suggested actions across 33 reps, and generated multiple business cases that unlocked board-level budget approval. One procurement process that projected a 60-day delay resolved in 12.

On the management side, the Monday pipeline-reconstruction routine gave way to a weekly Ollie briefing that arrives already synthesized and ready to act on. Win/loss coaching shifted from CRM reason codes filled in after the fact to specific behavioral patterns extracted from how reps actually worked the deals.

• Forecasting: Ollie has identified a clear correlation between deal health scores and historical win rates. The next step is surfacing that as actionable forecast intelligence, giving revenue leaders a more accurate picture of what will actually close, grounded in deal behavior rather than rep-reported pipeline stages.
• As Supermetrics continues to grow its enterprise presence post-Relay42, Ollie’s deal intelligence becomes increasingly central to the motion: not a supplementary tool, but the operating system for how the team sells.