About Von: What It Is, Who It's For, and How It Works

Von is the AI headcount for revenue teams. It connects to your revenue data, learns your business the way a senior hire would, and then takes on the analysis and busywork that used to sit on someone's plate for days. Built by the team behind Rattle, Von is now generally available.

If you've ever waited a week for a report, dug through Salesforce hoping to spot a pattern, or wished you had a data analyst on call around the clock, Von was built for you.

What is Von?

Von is an AI headcount purpose-built for revenue teams. It connects to your revenue data sources, including your CRM, call recordings, emails, calendars, and data warehouses like Snowflake. You hand it a task in plain English, and you get an answer or a finished piece of work back in a minute or two.

Von is not a chatbot bolted onto your CRM, and it is not another dashboard. Underneath the conversational surface is a real data science layer. Von runs SQL and SOQL queries, builds scoring models, runs predictive analytics, and surfaces patterns that would otherwise take a dedicated analyst weeks to find.

The simplest way to think about Von: a brilliant analyst who knows your entire sales operation, has read every call transcript, and can do the work in minutes instead of days. The difference is that this one onboards in a few days and then works for your whole team at once.

Is Von a chatbot or a copilot?

No. Von is a super agent that reasons across both your structured data (CRM, data warehouse) and your unstructured data (calls, emails, playbooks) at scale. A copilot helps one person write a little faster. Von takes on the work you would otherwise hire for, and it does work that was never possible before because no one had the bandwidth to attempt it.

The framing that resonates with revenue leaders is headcount, not copilot. Von is an additional team member that already knows your business, connects to your real data, and shows its reasoning so you can audit the work.

Who is Von for?

Von is built for anyone on the revenue side of a business who needs answers from sales data.

Sales leaders use it to prep for forecast calls and 1:1s. RevOps teams use it to audit pipeline health, build scoring models, and improve CRM accuracy. AEs use it to prep for meetings and understand their deals. Marketers use it to analyze campaign performance and find high-intent accounts. Finance and executive teams use it to understand win/loss patterns and forecast accuracy.

If your team asks questions about sales performance, pipeline, deals, or customer conversations, Von can help.

How does Von get set up, and how long does it take?

Von onboards like a new hire. After a brief setup call, it connects to your stack and spends roughly three to five days autonomously building a deep context graph of your business. It learns your fields, your stages, your fiscal year, your definition of churn, and the language your team actually uses

From there it is ready for your whole GTM team. Most customers see a meaningful productivity boost within the first two weeks. Corrections from any one user apply across the org, so the whole team works from one source of truth.

Which AI model does Von use?

All the major ones. Von routes each task to the model best suited for it, automatically. Claude for reasoning, and other leading models where they perform better, with the mixing and matching handled for you behind the scenes.

Most people don't care which model is under the hood. They care whether the answer is right. Von's job is to make that decision so you never have to.

What can you ask Von?

The best way to understand Von is to see the kinds of work people hand it. These are real examples from customers. They are a starting point, not a limit.

Pipeline and forecasting

• "How confident should we be in the current commit number? Flag deals with low engagement, slipped dates, or quiet champions."
• "What moved in or out of commit since last week, and why?"
• "What is my projected revenue for this quarter, overall, by month and by owner?"
• "Across all open pipeline this quarter, where is our biggest exposure by segment and by rep, and what are the top 10 deals driving that risk?"

Coaching and 1:1 prep

• "Prep me for my next 1:1. Summarize meaningful activity from the last 7 days, plus wins, losses, and coaching topics."
• "What patterns in this rep's calls or emails suggest coaching opportunities?"
• "Who are my top sellers for new business in the last 2 quarters? Show new logo ARR, win rate, and average sales cycle by rep."

Deal prep and account intelligence

• "I have a meeting with Jake from ACME. What have I already told him about pricing, and what expectations has he shared?"
• "For my top 20 opportunities this quarter, summarize the last 3 calls and key emails, and highlight any new risks or buying signals."
• "Which open opportunities are at high risk of slipping to next quarter, and what should the owner do this week to reduce that risk?"

CRM accuracy and data hygiene

• "Check all past-due meetings and confirm whether they actually happened using call recordings, calendars, and email history."
• "Find opportunities that have been in the same stage longer than our historical median for that stage. Group by owner."
• "Using call transcripts and emails, find opportunities past S1 with no clearly identified decision-maker or timeline."

Competitive and product intelligence

• "How many times have we competed against [competitor] this quarter, and why did we win or lose?"
• "For all opportunities where a competitor is mentioned, summarize win rates, average deal size, and key reasons we win or lose."
• "What feature requests come up most often in calls, and how many deals have we lost waiting on them?"

Predictive analytics and scoring

• "Based on the last 12 months of closed-won opportunities, build an account scoring model using signals like number of touches, marketing engagement, website visits, and title/role mix."
• "Using all closed-won and closed-lost deals in the last year, build a model that predicts win likelihood."
• "Compare engagement patterns between closed-won and closed-lost opportunities: number of touches, average time between touches, and number of unique contacts involved."

Marketing and demand gen

• "Identify which intent signals most strongly correlate with opportunity creation within 30 days, ranked by predictive power."
• "For all accounts without an open opportunity, score them 0-100 using historical intent and engagement. Output the top 100 accounts we should prioritize."

Churn and renewal intelligence

• "What are the top renewal risks for Q1 accounts based on the last year of conversations, with specific call references?"
• "Stack-rank churn risks by frequency, and compare what's documented in Salesforce against what customers are actually saying on calls."

Can Von update Salesforce?

Yes. Von doesn't only answer questions. It can log calls, update opportunity fields, create records, build flows and validation rules, and make bulk changes across multiple objects. Von always confirms before making changes and respects your existing Salesforce permissions and validation rules.

How is Von different from just connecting Claude or ChatGPT to my CRM?

This is the most common question, and it matters. Wiring a general-purpose AI to your tools with connectors gets you a demo, then you hit three walls.

1. Search quality. Connecting tools directly gives the model keyword search without vectorization. A question like "why are we losing to competitor X?" returns shallow, keyword-matched answers. Von runs true semantic search across all your data, so it understands concepts, not just matching words.

2. Context limits. A general model has a large context window, but a single call transcript can run 10,000 to 20,000 tokens. Two hundred deals at three calls each is several million tokens, far more than the model can hold at once. It analyzes the calls that fit, quietly drops the rest, and never tells you which it skipped. Von pre-processes and vectorizes all of your unstructured data ahead of time, so it can analyze 10,000+ calls in a single query and give a real answer, not a sampled guess.

3. Setup and enablement. With a general model, each user wires up their own connections, and making it work for every persona (CRO, manager, AE, CSM, SDR) falls entirely on RevOps to figure out and maintain. Von's admin connects once. Von already knows what works across dozens of revenue teams, so there is no skill file to write and no per-user lift.

I have great engineers. Can't they build this in a weekend?

They can build a prototype in a weekend that impresses the room. The weekend isn't the problem. What takes three to five engineers and six-plus months is everything after the demo: a pipeline that pre-processes every deal nightly, a semantic layer that searches concepts across 100,000+ calls, a model of your business that learns your fields and terminology, org-wide memory so the whole team shares one source of truth, and production connectors that don't break. Then you maintain all of it forever, with engineers you'd rather have shipping product. Von is that system, already running on day one.

What are teams getting out of Von?

Von processes over 10,000 revenue tasks every week across its customer base, and the people using it most are often the ones who came in most skeptical, because they can see the reasoning and verify the work.

What customers consistently report is that Von feels like additional headcount, not another login. It handles analysis they previously couldn't justify dedicating a full-time hire to, and it turns projects that lived permanently on the "someday" list into work that gets done every day.

Trusted by revenue teams at Tapcart, QuickNode, Coalesce, 15Five, DemandScience, Qualified, and more.

"Von isn't just a tool, it's genuinely an additional teammate on my RevOps team. It answers complex questions in seconds that used to take me hours, from pipeline analysis to deal risk assessment to Salesforce configurations." — Taylor Kelly, Head of Revenue Operations, Tapcart

"We had 'Sales GPT' on our internal roadmap because everyone from CS to the CRO was asking for it. Von solved this gap before we could even start building it ourselves. Von is now the intelligence layer across our entire revenue stack." — Rob Janke, Director of Revenue Operations, QuickNode

"Unlike ChatGPT or other AI tools, Von is directly connected to our data sources, which makes it actually applicable to my day-to-day work instead of just theoretical." — Evan Briere, VP of Partnerships, DemandScience

How does Von relate to Rattle?

Von is a new product from the same team that built Rattle. It does not replace Rattle. If you're a Rattle customer, you can keep using Rattle as you always have, and we encourage you to try Von to see what it adds for your team. Existing customers get extended free trials to explore Von's capabilities.

What do I need to use Von?

You need Salesforce or HubSpot as your CRM. To get the most value, you'll also want to be recording sales calls and connecting your other revenue data sources. Setup is fast, and you can start handing Von tasks as soon as it has connected and learned your business.

Is Von secure?

Yes. Von is SOC 2 compliant, with additional security assessments including ISO 27001, Google CASA Tier 2, and regular third-party penetration testing. You can learn more about our security practices at trust.vonlabs.ai.

Quick answers

What is Von? An AI headcount for revenue teams. It connects to your CRM, call recordings, emails, calendars, and data warehouse, learns your business, and does work that used to take days in minutes.

Who makes Von? The team behind Rattle. Von's CEO is Sahil Aggarwal.

Does Von replace Rattle? No. Von is a separate product that complements Rattle.

Which CRMs does Von support? Salesforce and HubSpot. More CRMs are planned.

What data sources can Von connect to? CRM, call recordings, emails, calendars, and data warehouses like Snowflake. Native one-click integrations cover the most common tools, with many more available through MCP.

Do I need call recordings to use Von? You can start with just your CRM, but connecting call recordings and other revenue data unlocks the most value.

How fast does Von return answers? Usually a minute or two. Because Von runs real queries and analysis rather than guessing, it takes a moment, and that is still far faster than the hours, days, or weeks the same work used to take.

Can Von do predictive analytics? Yes. Von builds scoring models, identifies patterns, runs SQL and SOQL, and performs predictive analytics.

How long does onboarding take? Von connects to your stack and spends three to five days building a context graph of your business. Most teams see a productivity boost within two weeks.

Can Von update my Salesforce data? Yes. It can create, update, and modify records, and build flows and validation rules. It always confirms before making changes and respects your permissions and validation rules.

How much does Von cost? Contact our team for pricing. Free trials are available so you can experience Von before committing.

Get started with Von

Stop waiting for reports. Stop digging through your CRM. Start handing off the work and getting answers back.

‍

Book a demo

‍