Why we built Helix for personal development, not AI surveillance

When we started building Helix, we spent a lot of time thinking about a question that had very little to do with AI itself: would people actually feel comfortable using Helix at work?

AI is increasingly capable of analysing workplace conversations and turning them into useful insights. That's exciting from a product perspective, but it also comes with a responsibility. If employees think an AI tool is quietly monitoring them, scoring them and reporting back to HR, it can quickly become something people feel they need to work around rather than something that helps them.

That's not the product we wanted to build. That’s why we designed Helix, our AI skills intelligence platform, around measurable skills development and performance improvement, not workplace surveillance.

The problem with AI surveillance at work

There's a growing conversation around employee monitoring and AI surveillance at work, and some of the concern is entirely understandable.

When you introduce AI into meetings and workplace interactions, people naturally want to know what is being recorded, who can see it, or what consequences using this tool could have further down the line.

Those are reasonable questions. We’ve all heard horror stories about AI notetakers recording confidential conversations, or the wrong people being able to read AI meeting notes. With any new technology, it’s natural that people would want to know exactly how their data is being used.

From a product perspective, I believe the answer shouldn't be to tell people not to worry. We should proactively reassure users before they ever use the product so they have an explicit reason to trust it.

That's shaped a lot of the decisions we've made with Helix.

Helix is designed to support personal skills development

The fundamental purpose of Helix is to help people identify their workplace skills and explore opportunities to develop them.

Rather than relying solely on yearly assessments or asking employees to self-report how good they are at something (unsurprisingly, this approach isn’t very accurate or meaningful), Helix can use the conversations that are already happening at work to generate useful skills insights for the people demonstrating those skills.

That might help someone recognise a strength, spot an area they'd like to improve or understand how they're showing up in meetings. Users have told us that Helix has helped them identify blind spots in their own behaviour or revealed strengths they never realised they had.

The important thing is what happens with that information. To be clear, Helix isn't designed to make disciplinary decisions or determine whether someone is a high or low performer, or a 'good' or 'bad' employee. It doesn't make automated employment decisions, like who should be promoted or who should be laid off. It’s not there to spy on you or report back to your manager or HR about where you could improve.

It's designed solely to support learning and development. HR can benefit from aggregated data, insights and trends from across the workforce, but your skills insights are to inform how you personally show up at work.

Privacy isn't an afterthought

We didn't want privacy to be something organisations had to figure out around the edges of Helix. It needed to be part of the product to avoid people viewing it like an AI surveillance tool.

That's why Helix includes controls around how its insights are used, including options relating to anonymity, scores, written feedback and email summaries.

Organisations can configure Helix’s assessment and reporting model, including whether scores, written feedback and certain organisational views are available, so that the product matches the agreed development use case.

Customer data is also not used to train AI models. Your workplace conversations and skills data aren't being fed back into a general-purpose AI model to improve it for other users. This is different from some mainstream consumer AI tools, like ChatGPT or Claude, where conversations and uploads may be used to improve models depending on the user's settings.

There's another important element of control, too: Helix doesn't have to be invited to every meeting. It's there when it's useful. Teams can choose which meetings Helix joins rather than treating it as an invisible participant that sits in the background of every conversation. If you want to cover confidential information mid-meeting, you can remove Helix as a participant. If you simply want a break from skills analysis, don’t invite Helix to your meetings and you can carry on as normal.

For me, that's an important distinction between an AI development platform and AI surveillance at work.

Your conversations aren't an HR record

One of the biggest misconceptions about workplace AI is that if an AI system is present in a meeting, someone must be able to access everything that was said afterwards.

That's not how Helix works.

5app can't see meeting transcripts, and neither can HR.

The transcript isn't there to create a searchable record of individual employees' conversations. The data is used to generate the skills insights that Helix is designed to provide. It’s not like using Google Gemini to record a meeting transcript or summarise the conversation – there won’t be a document saved to everyone’s calendars and emailed to all the participants. Your Helix skills insights and feedback are for you, in your skills profile and your emails. That’s it.

That's an important part of the trust model we're trying to create: the technology should be useful without creating a new mechanism for personal scrutiny.

Skills heatmaps are about development trends, not employee ranking

Helix can also give organisations a broader view of their skills landscape through a heatmap.

This is where I think the distinction between skills intelligence and performance surveillance is particularly important.

The heatmap is designed to help HR teams understand trends. Where are skills developing strongly? Where might there be opportunities to invest more in learning? What capabilities could the business strengthen?

It's information to help organisations make better decisions about development. Does the sales team need more support around a specific skill set? Are leaders equipped to manage their teams effectively? Are your coaches and trainers adhering to the company's preferred coaching framework?

It isn't intended to become a leaderboard of employees or a mechanism for making individual performance or disciplinary decisions. You won’t earn points or be secretly ranked behind the scenes.

In other words, we're interested in what the data can tell us about how an organisation can improve, rather than using it to spy on individual users.

AI should earn people's trust

As a product leader, I don't think trust can be bolted onto an AI product at the end.

If you're asking people to bring AI into something as human as a conversation about their work, you need to think about privacy, transparency and control from the beginning, not as an afterthought.

That means being clear about what the technology does. It means giving organisations meaningful controls. And it means being deliberate about what the data is, and isn't, used for.

The wider conversation about AI surveillance and employee privacy is going to become increasingly important as AI becomes more embedded in the workplace.

Employees need to understand why an AI tool has been introduced, what it’s analysing, who will see the results and what those results can be used for. Transparency around AI solutions, including Helix, ensures users are fully informed about the tools they’re using. If an AI tool is introduced by stealth or its purpose isn’t fully explained, there will always be concerns that it’s being used for surveillance.

We believe there's a better approach.

The point isn't to watch people. It's to help them grow.

I think the most useful way to look at Helix is this:

Workplace surveillance asks: "What are people doing?"

Personal development asks: "How can we help people become better at what they do?"

We built Helix around the second question.

The technology behind it is powerful. But the purpose behind it is deliberately simple: give people useful insight into their skills, help them develop and give organisations a clearer picture of where they can support their people.

That’s why Helix isn’t a tool for AI surveillance at work. It's for personal development, genuinely interesting insights and continuous development support, right when you need it.