Skip to content

Why

Is your knowledge AI-ready? The amplification trap One source of truth

Platform

The knowledge foundation Governed & validated Security & compliance Integrations

For whom

Government & public sector Finance & insurance Healthcare Beyond the contact center

Customers

Customer stories

Company

About Polly Leadership Newsroom Careers Partners Contact Book a demo

Company · Leadership

A dedicated, efficient team behind every customer.

There is no account management layer here, and no tier one. The person who answers your question is usually the person who can change the thing you are asking about. That only works with a compact team, and it is the main reason we keep it that way.

Stefan Mackaaij

Stefan Mackaaij

Chief Executive Officer

Took over as CEO in July after years of shaping the product and the positioning from the inside. If a customer wants to know where Polly is going, this is the conversation.

Leadership

Who leads Polly.

The chief executive, the chief financial officer and the founder who started the company in 2015. The rest of the team is on the About page.

Stefan Mackaaij

Stefan Mackaaij

Chief Executive Officer

Runs the company and sits in the room when a customer decides what their knowledge is worth.

Peter Vleugels

Peter Vleugels

Chief Financial Officer

Finance, contracts and the paperwork a procurement team sends over.

Cees Zwezerijn

Cees Zwezerijn

Founder & Chief Strategic Officer

Started Polly in 2015, after watching the same question get three different answers.

In his own words

Without reliable knowledge, no reliable AI.

In June 2026, shortly before he took over as CEO, Stefan Mackaaij spoke to Erik Bouwer of Ziptone about what generative AI changes for knowledge management, and what it does not.

“AI does not make knowledge management redundant. It makes it business-critical.”

A language model is not a knowledge base.

“A lot of people think that as soon as you use LLMs such as ChatGPT or Claude within knowledge management, you are automatically working with generic information. That is not true. A language model is not a knowledge base. It is a technology you can use for specific tasks, such as rewriting, summarizing, checking or formulating answers based on validated knowledge.”

Three things have to come together.

“You need technology to manage, find, publish and distribute knowledge. You need good content that is current, validated and suitable for different audiences and channels. And you need an organization in which ownership, review processes, feedback and governance are properly set up. If one of the three is missing, it becomes hard to make knowledge truly reliable.”

Somebody has to own it.

“Everyone has an interest in good knowledge, but if nobody owns the process, it often gets left. That does not mean you have to build a large knowledge management department straight away. You can also start with roles: who owns which subject, who reviews changes, who guards the quality, who looks at feedback and who makes sure knowledge is reviewed periodically?”

Start with the knowledge, not the tool.

“The first question is not which AI tool you need. The first question is which knowledge you trust enough to let employees, customers and AI use. Start there. Map your sources, assess the quality, set down ownership and put a process in place to keep knowledge current. The technology comes after that.”

Source: Erik Bouwer, “Zonder betrouwbare kennis geen betrouwbare AI”, Ziptone, 23 June 2026. Read the full interview in English

In production

Trusted where a wrong answer costs the most.

More than a hundred organizations run their knowledge on Polly. Government, healthcare, financial services, utilities. Places where roughly right is not good enough, and where one outdated answer becomes a complaint, a claim or a headline.

100+
organizationsin production, not in pilot
15+
sectorsfrom government to logistics
10+
countriesacross Europe, North America and Asia
2015
building thisbefore knowledge became urgent

Where they sit

Government & public sector Financial services & insurance Healthcare Retail & e-commerce Charities, lotteries & NGOs Utilities & energy Leisure & tourism and eight more sectors

How we build

Four things we will not trade away for speed.

We work in short feedback loops with real users, design for scale and simplicity, and use automation and AI where it adds value, never where it costs us quality.

01 Clarity We simplify the complex, in the product and in the way we talk about it. If an answer needs a manual, the answer is not finished.
02 Reliability Customers put their most critical information in here. That is a different promise than uptime, and it is the one we are judged on.
03 Innovation The platform improves on what people actually run into, not on what the category is talking about this quarter.
04 Partnership We started as the people implementing this for others. That is still how we work: alongside the team, not across a service desk.

The same foundation, in your own tenant. Read-only to start, nothing to migrate.

Book a demo

Make your knowledge AI-ready.

Before you trust AI, trust your knowledge. See how Polly builds the foundation your people, your customers and your AI can rely on.