How 75% of Parallel Web Systems’ team self-serves analytics with Hex

How 75% of Parallel Web Systems’ team self-serves analytics with Hex

Challenge: Mounting data requests, zero self-serve

In the beginning, Parallel Web Systems’ analytics was a cobbled-together combination of Google Sheets and AI-generated SQL queries that pulled data directly from the warehouse.

For the team, it was the quickest way to get the answers, insights, and dashboards they needed, but as the company scaled, it soon became disorganized and unreliable.

With their rapid growth trajectory — which included a $2B valuation and a Google Cloud Partnership — data requests went through the roof.

Leadership, engineers, and go-to-market teams all needed answers but key metric definitions, business logic, and institutional knowledge was scattered across reports and teams. This led to team members creating their own way of getting answers and having no clear source of truth.

Worse still, the fragmented cycle of downloading data from the warehouse and uploading it to Sheets via scripts, made it nearly impossible to track business and product trends over time.

The Parallel Team needed one analytics platform where everyone could quickly find, explore, and share trusted data, with the right permissions and context baked in.

Enter Hex.

Solution: A true self-serve data solution with Hex from one data analyst to 75% of the company

Parallel chose Hex because it accelerated the data team's most complex work, and gave the broader org a governed, reliable path to self-serve answers.

Before going live, Parallel’s first Deployed Engineer, Khushi Shelat, and Finance lead, Darien Lee, consolidated all of the company’s institutional knowledge to build context for Hex's agent. They mapped column names to definitions and wrote instructions to join tables based on how they envisioned data would be queried.

This created a single source of data that everyone could trust.

"Trusted and governed data now lives inside Hex, and is accessible through Slack and our various agents via the MCP. This has dramatically increased our ability to self-serve data questions across the organization.” — Khushi Shelat, Deployed Engineer, Parallel Web Systems

Hex’s Slack integration and MCP connector proactively drives insights

Before Hex, self-serve at Parallel meant using Claude and Codex to build SQL queries that could pull data directly from the warehouse. However, getting a consistent answer was near impossible when everyone had their own definitions and processes.

To increase self-serve analytics from a single, trusted source, the team turned to Hex. With Hex's Slack integration and MCP connector, users could access trusted insights from Slack, Claude, or another chatbot without needing to access Hex’s UI. Simultaneously, admins had the ability to see what questions users asked via the Context Studio, giving Khushi and Darien the ability to build additional context or workflows to solve repeatable questions.

Beyond increasing access, Parallel built agent tasks that sent slack notifications to their customer-facing teams when one of their customers experienced a usage spike, error, payment failure, or was featured in the news. This gave customer and account teams the context they needed to act faster, maximize product use, and spot new customer expansion opportunities. Parallel aptly named this tool Customer Watch, an always-on agent that alerts the right account manager in Slack when one of their customers

[CALLOUT: Learn more about the Customer Watch program Parallel built with Hex.]

Being able to ask questions and get trusted insights from where the team already worked was a game changer. By connecting Slack and other AI chatbots to Hex, the entire organization was able to maintain and create new workflows with trusted data. Now, over 75% of the organization self-serves analytics with Hex daily.

"Today, 75% of the organization incorporates data into their day-to-day work with Hex in a way that wasn't possible before. The ability to access our data across multiple surfaces and tools has transformed our data culture." — Khushi Shelat, Deployed Engineer, Parallel Web Systems

Business metrics leadership could finally trust

Before Hex, Parallel had no agreed-upon way to calculate its most important business health metrics. Definitions and business logic lived across multiple sources, and there was no single version the whole org could align on.

Using the Hex Agent, the team built a canonical dashboard of business health metrics, allowing them to visualize insights and replace multiple spreadsheets in a matter of hours. Now leadership has one clear, trusted report in Hex where they can quickly see revenue, retention, and growth trends and explore them. When they want to dig into a spike or dip, they can leave comments to relevant team members or chat directly with the Hex Agent to dive in further.

"With Hex, our leadership team can dig deeper and get answers mid-meeting instead of waiting on an analyst, which means faster decisions and a lot less back-and-forth between teams." — Khushi Shelat, Deployed Engineer, Parallel Web Systems

Catching API issues before customers do

For Parallel's engineering team, API performance is everything. They needed to be able to look at product telemetry and ongoing evals to identify issues and investigate: Are the agents accurate? Are they drifting? Are they getting faster or slower, and when did that change?

Before Hex, those answers were fuzzy and static. Existing reports in Google Sheets referenced stale data and its rudimentary visualizations couldn't show important nuances on API performance. By the time a latency spike was noticeable enough to flag, it had usually been compounding for a while.

Hex gave the engineering team something they'd never had before: real time access to trusted data. For the first time, engineers could see performance data across different API endpoints simultaneously, spot anomalies as they emerged, and compare performance across product lines over time.

But clarity was only half the story. The other half was speed. Eval metrics that were sporadically updated throughout the week now refreshed continuously, reducing time-to-insights from every few days to every 24 hours on a schedule . When something shifts like a drift in accuracy, a spike in latency, an unexpected improvement, the team sees it immediately (instead of waiting for a stale overnight snapshot), investigates it in Hex, and acts.

“When you're building agentic systems, a performance issue that goes unnoticed for 24 hours is too long. Hex closed that gap.” — Khushi Shelat, Deployed Engineer at Parallel

The impact: From scattered spreadsheets to a self-serve data culture

Parallel graduated from "having data" to running on it. With Hex, data is available instantly, reliable, and easy to build workflows with. Engineers catch issues before customers do. Leadership makes decisions in the meeting, not after it. And anyone on the team can get a trusted answer in the time it takes to send a Slack message.

With Hex, Parallel achieved:

• 75% of the org using trusted answers on their own — starting with one analyst as a bottle neck, to everyone self-serving analytics
• 44% of Parallel’s Hex users engage primarily through the Slack integration.
• 24-hour scheduled product telemetry updates — from sporadic manual updates to scheduled nightly runs.
• One canonical dashboard of key business metrics replacing a patchwork of scattered, inconsistent spreadsheets.
• Proactive GTM customer alerts that equip account executives with details that make every call tailored and efficient.

As for the future, the team plans to integrate data across more workflows and expand their Customer Watch program across every stage of the customer lifecycle.

"Before Hex, one person built the majority of the company’s dashboards and insights manually, now the majority of the company does it themselves." — Khushi Shelat, Deployed Engineer, Parallel Web Systems