Pulzio

The challenge

Managers rarely know how their team actually feels until someone resigns. The standard instrument is an annual engagement survey, and it fails in three predictable ways.

• It is stale before it is read. A survey that closes in March and is presented in May describes a team that no longer exists.
• People are not honest in it. If a response can be traced back, even in theory, it gets sanded down into something safe.
• Nobody acts on it. The output is a slide deck of percentages. There is no step between "engagement dropped 6 points" and "here is what to do on Monday".

The product had to solve all three at once: continuous instead of annual, anonymous by architecture rather than by promise, and ending in an action rather than a chart.

What we built

Pulzio is a browser-based feedback platform. Employees answer one question a day; HR and team leads get a live read on mood plus an AI assistant that turns raw sentiment into a prioritized plan.

• A 30-second daily check-in. Five emotions on a single scale, one click, plus an optional comment tagged with a hashtag such as #workload or #culture. Short enough that people actually do it every day, which is what makes the trend line meaningful.
• Anonymity in the data model, not in the copy. Votes and comments are stored without a link to the employee record. Nobody can de-anonymize a response, including an administrator with full access to the database. That constraint shaped the schema first and the UI second.
• A real-time dashboard. Average mood, weekly vote volume, participation rate, and the mood trend, broken down by department and by hashtag, so a dip is attributable to a team and a topic rather than to the company as a whole.
• A moderated feed where anonymous comments are visible to the company and HR can reply in public. This is the part that stops the tool from feeling like surveillance: people can see that a comment produced an answer.
• Departments, roles, and invitations by email or link, with QR-code onboarding so a whole floor can join from a poster.
• PDF reports for the people who still need something to bring to a board meeting.
• No app to install. Everything runs in the browser, on a phone or a laptop.

The role of AI

AI is the layer that makes the feedback useful rather than merely collected.

• Sentiment analysis over anonymous comments surfaces recurring themes and detects when the tone of a department shifts, including cases where the numeric mood score has not moved yet.
• AI action plans turn those signals into specific, prioritized checklists: what to change, who owns it, what to watch afterwards. The assistant then tracks whether the metric it flagged actually recovered.
• An AI chat over aggregated data lets a manager ask questions in plain language ("what is driving the drop in engineering this month?") against anonymized data only, so the answer can never resolve to an individual.

Engagement and stack

DForce took the product end to end: discovery and product definition, interface design, the front end, the data model, and the AI layer. The platform is a web application with an analytics dashboard and an AI sentiment pipeline, built with GDPR-compliant storage and export from the start rather than added later, because anonymity guarantees cannot be retrofitted onto a schema that already stores the link.

Pulzio is currently in closed beta with early companies onboarded on a Growth plan, which is deliberate: the roadmap is being set by what those teams actually use rather than by what tested well in a demo.

Services: Product, Design, AI engineering. · pulzio.net

Service: AI product development