AI Data Analysis for Sales & Marketing Teams: Real Demos, No SQL Required

2026/05/07

TL;DR: MotherDuck co-founder Ryan Boyd and AE team lead Logan Tosky show how sales reps and marketers can query their own data and build shareable dashboards using AI, no SQL required.

What this covers

Most data tools are built for engineers. This session is for the people who actually have to hit the number: account executives trying to catch churn before it happens, marketers who want to see their lead funnel without filing a ticket, sales managers who need their own view of the pipeline.

Ryan Boyd (MotherDuck co-founder, leads marketing) and Logan Tosky (AE team lead, 10+ years in tech sales) walk through how they each use AI and MotherDuck day to day, then run live demos of what they've actually built.

How Logan manages his book of business

Logan built a custom account health dashboard that shows usage trends, capacity consumption, and growth signals across his entire book. He can drill into any account with one click, filter by growth rate, and spot customers who might be approaching their limits. Before this, that meant pairing with a customer engineer and writing SQL together.

He also demos using Claude as a thought partner alongside the visual dashboard. He'll notice something interesting in a chart, then ask Claude to explain what's actually driving it.

How Ryan tracks marketing performance

Ryan's team built a unified marketing dashboard covering lead sources, email sequence performance, webinar analytics, and web traffic. It pulls from PostHog, Google Search Console, and HubSpot. The whole thing took about two days, and a non-technical marketer on the team built pieces of it using nothing but prompts.

What makes this possible: the MotherDuck remote MCP

The MotherDuck MCP server connects any AI agent to your data warehouse without local setup. Logan walks through connecting Claude to MotherDuck on a fresh account in real time. It took about 20 seconds.

What is a MotherDuck Dive?

A Dive is a shareable, refreshable analytics app you build directly in your AI agent. Instead of a one-off chart that lives in a conversation, it becomes a URL your whole team can bookmark and revisit. Ryan and Logan both demo Dives they've built and show how you can combine Dives from different team members into a single shared dashboard.

You can browse examples in the MotherDuck docs to see what's possible before building your own.

FAQS

Sales reps can connect Claude to their data warehouse through the MotherDuck MCP server and ask questions in plain English. Claude writes and runs the SQL, so you can dig into account trends, usage spikes, or pipeline changes without waiting on a data team. Logan Tosky walks through the setup in the video.

A Dive is an interactive analytics dashboard you create by describing what you want to an AI agent. No React or SQL required. Once it exists, it's a URL with live data — anyone on your team can open it without writing queries. You can embed Dives on internal pages or combine a few into a larger dashboard.

In most cases, under a minute. The MotherDuck MCP server is an official connector in Claude's marketplace — find it, click connect, authenticate. No installation or config files. Ryan Boyd did this live in the webinar and it took about 20 seconds.

Yes. Ryan shows a webinar analytics dashboard built by a marketer with no engineering background, using only natural language prompts. The point is that domain experts—the people who actually know what questions matter—can now build the dashboard they need without going back and forth with an engineering team.

MotherDuck uses "blessed dives." Once a Dive meets the quality bar, it gets added to a shared repository and marked as approved, basically a Git repo with data team sign-off. For marketing, Ryan keeps an official marketing dive that team members contribute to. He decides what gets included. The process is still early, but the idea is: explore freely, then promote the keepers.

Related Videos