How Oyster and Cresta Put an AI Revenue Headcount to Work

RevOps and analytics leaders share a quiet frustration that rarely makes it into vendor demos or conference keynotes. The ad hoc analysis requests pile up, forecast calls eat hours that could go toward actual strategy, and the headcount request to fix all of it keeps getting denied.

That was the starting point for our recent webinar, where two revenue team leaders shared how they're using Von as AI headcount. Not a dashboard or a chatbot, but actual capacity that does the analytical work and tasks their teams haven't had bandwidth for.

Sonia Kozlowski, Senior GTM Operations Manager at Oyster, and Ted Ranney, Head of GTM Analytics at Cresta, walked through specific use cases, the trust-building process they each went through, and what changed once they made the shift.

The Real Source of Truth Isn't Your CRM

Sonia opened with a mindset shift that will sound familiar to anyone who has spent years chasing sales reps to update Salesforce fields.

"If it's not in Salesforce, it doesn't exist" is a phrase she'd repeated at her last three companies. She now considers it outdated. Reps talk to customers and prospects every day. They update Salesforce once a week if you're lucky. The actual source of truth lives in the calls, emails, and customer interactions happening between those updates.

Sonia had spent years optimizing picklists, validation rules, and field structures trying to solve data hygiene. None of it worked without creating friction for the sales team. When she connected Von to her revenue data sources, both structured and unstructured, the problem resolved itself. Von pulls context directly from where the interactions happen, not from fields that may or may not be current.

As Sonia put it, "I finally fixed data hygiene by doing nothing. I just plugged in a system that's connected to my actual source of truth."

Forecast Calls That Focus on What Matters

Forecasting came up repeatedly throughout the webinar, and for good reason. Sonia described it as a process built on opinions. A rep has a bad day, they pull back on a commit. But the business can't run on whether someone is feeling optimistic on a Tuesday.

With Von, Sonia now compares what her sales team has committed against what Von sees in the actual interaction data. The forecast call shifts from reviewing every deal to focusing only on the mismatches, the places where the rep's call and the data tell different stories.

That means fewer forecast changes, more confidence in the numbers, and significantly less time spent in review meetings.

Closing the Product Feedback Loop

Sonia also walked through a use case that product and ops teams will recognize. Collecting product feedback from the sales floor is one of those things every company says they do, but few do well at scale. Feedback lives in emails, call recordings, Slack threads, and meetings across SDRs, AEs, and CSMs. Pulling it all together manually is a massive lift, and it usually happens once a quarter at best.

With Von, Sonia can collect and synthesize product feedback from across all of those sources in minutes. That alone is valuable, but she took it a step further. When a requested feature actually ships, Von can go back and surface every prospect who didn't convert and every customer who churned because of that gap. That gives the sales team a ready-made re-engagement list tied to a real product update, turning lost deals into pipeline again.

Scaling a Two-Person Analytics Team Without Adding Headcount

Ted Ranney runs GTM analytics at Cresta. His team is two people. The volume of requests coming in from sales, CS, and leadership is not scaled for a two-person team.

He described three problems Von helped him solve.

End-of-quarter chaos. Ted is responsible for reconciliations, audits, internal reporting, board decks, and a two-hour QBR presentation at the start of each quarter. He used Von to build his entire quarterly business review, including a full-year forecast that landed within 5% of the FP&A model. Without Von, he said the QBR simply would not have been possible alongside everything else on his plate.

Renewal and competitor analysis at depth. Ted used Von to go beyond the picklist reasons logged for churn and downgrades. He had Von analyze six months to a year of actual conversations to surface the real drivers, then take it a step further and recommend strategies to address them. He took the same approach for competitive analysis. Instead of relying on fields AEs may or may not populate, Von reviewed transcripts from won and lost deals to identify which competitors showed up and what made the difference.

Slack-channel triage. Like most RevOps teams, Ted's team has a Slack channel where requests pour in. Reports, data pulls, one-off analyses. Each one might take 30 minutes to a few hours. Von now acts as a first responder. In one example, a colleague asked for an NRR analysis sliced a specific way. Von ran it in the background and flagged that the sample size was too small to be meaningful, saving Ted the time of building the whole thing out before delivering that same conclusion himself.

Trust Is Earned, Not Assumed

Both speakers were candid about skepticism. Neither trusted Von immediately, and both said that skepticism was healthy.

Sonia described a progression where she started by asking Von questions she already knew the answers to. When those checked out, she gradually moved to using Von as her first stop for any question. The turning point was when she could validate analyst work that had taken three to four days in a matter of minutes.

Ted took a similar approach but leaned heavily on Von's Organizational Memory feature. He described it as "dumping his brain" into the system, teaching it how Cresta calculates pipeline, which filters to apply, which test accounts to exclude, and how win rate is defined. He spent dedicated one-on-one time training and verifying before he shared any Von output with the broader team.

His advice to anyone evaluating the product: "Every minute you spend training Organizational Memory is invaluable. Take all the brains on your team and make sure they're dumping all their knowledge into the system before you start asking it questions."

The Shift from Reporting to Strategy

When asked what changes for revenue teams in the next year, both speakers landed in the same place. The role shifts from operational work to strategic work.

Sonia put it simply. She no longer needs to spend hours pulling data and assembling reports. That time now goes to deciding what to do with the data. Ted described running Von prompts in a side window during live meetings, getting answers in minutes that would have previously required days of follow-up.

Von didn't replace their judgment. It compressed the time between question and answer so they could spend more time on the decisions that actually move revenue.

Your CRM Data Isn't the Full Picture

Sonia closed her presentation with a question that stuck with the audience: what would break if you stopped relying on your CRM data as your single source of truth?

For most revenue teams, the answer is a long pause while they think about what would actually change and what might get better. Your reps are having dozens of conversations every week that contain more signal than any Salesforce field. The real insights about deal risk, churn drivers, competitive positioning, and forecast accuracy live in those interactions. Most teams just don't have the capacity to access them.

Von changes that. It connects to your CRM, call recordings, emails, and data warehouse, learns how your business actually works, and delivers the analytical output and revenue tasks your team has been backlogged on for months. That's what headcount does. It shows up and does the work.

‍

Want to explore whether Von is a good fit for your revenue team? Book a strategy session with Sahil, Von's CEO, to talk through your AI strategy and specific use cases.