Adam Jones|HomeBlog

Why product managers are uniquely suited for tech policy roles

Headshot of Adam Jones

Adam Jones

If you have product management skills, you're likely able to have a huge impact in technology policy. I think you'll be at an advantage getting a policy job, and your skills will enable you to excel in the role. This article sets out why.

This isn't just theory. I worked as a software engineer alongside great product managers at Palantir, Starling Bank and BlueDot Impact. I then worked on AI policy in the UK Government, getting first hand experience with the policy process.

Why consider policymaking? Good policymaking is essential to the functioning of the government, and the country as a whole. Policymaking is also one of the most powerful levers for preventing AI or biosecurity catastrophes.

Unfortunately, many people look over policy careers because they don't understand what's involved. Maybe you think it requires a politics background, or is only about putting out nice speeches.1 I'm going to show you this isn't true!

What do product managers and policymakers do?

Most product managers spend their time between four key activities:

  • identifying problems, for example through market research and user interviews
  • finding and selecting solutions, for example through hypothesis testing and stakeholder negotiation
  • delivering, often by directing software engineering teams
  • monitoring, for example through defining and tracking KPIs

Guess what policy advisors do all day? The same!

Not only do they do the same kinds of work, many of the detailed actions have equivalents too. I'll explain this for each policymaking activity.

Identifying problems

Policymaker areas tend to already be set by ministers, the specifics are rarely fleshed out (especially in tech policy). For example, in AI a minister might direct staff to focus on reducing AI harms, boosting AI innovation in industry, or increasing government effectiveness with AI - but they'd rarely specify all the details of how this would happen. Product managers similarly have ownership of a certain area, such as the onboarding flow of a product, even if the specifics are for them to figure out.

Both product managers and policymakers usually start by meeting people who they're trying to help. Product managers call these user interviews, while policymakers call this public engagement. Because user interviews and public engagement are so similar, principles from books like The Mom Test apply to both:

  • ask people about their lives, not your ideas
  • ask about specific instances in the past, not generics or opinions about the future
  • understand how problems are being solved today, and why that solution is not working
  • challenge vague or misleading responses

Let's see how user interview skills help in practice.

Imagine you're tackling AI discrimination:

  • A mistake would be to jump to a solution: ban discriminatory systems used for employment decisions, and fine companies large amounts should they violate this. You suggest this idea to people, who agree that this should be banned and companies should be fined. However, in practice you haven't learnt much and your solutions will be worse for it.
  • Doing this well, you'd focus on people's lives. You'd discover that a key problem with AI discrimination is that people are not able to get clear evidence of it. You'd learn how they tried to solve this in specific cases: for example, through the courts' disclosure processes. You'd then identify why those solutions aren't sufficient: for example, disclosure is refused because courts consider these systems to be protected trade secrets. Through this you've learnt a lot about what the problems actually are, so you can begin building better policies.

Finding and selecting solutions

Next, you'd identify potential solutions to these problems! While they might have less experience as to what government policy levers are available,2 people with product experience often generate more technically-grounded ideas, and can more accurately predict how companies will react.

Once you've generated solution ideas, you'll select which ones you want to proceed with. This might involve gathering more data, designing experiements, or consulting stakeholders. Again, these are all similar to how product managers prioritise solutions.

Delivering

Next comes delivery (sometimes called implementation). This is quite varied: sometimes it's actually carrying out the policy, and other times it might be drafting new legislation to allow it to be carried out.

Delivery often requires specific skills, such as legal expertise for legislative drafting. You'll therefore often be leading delivery teams rather than delivering directly yourself, in the same way a product manager might lead software engineers.

Policy can have a huge impact on people's lives, so many people will often have a stake in the outcome. This can mean there is also a need to negotiate with stakeholders (frequently other government departments). Any product manager in a large organisation will be used to coordinating with different teams (e.g. such as marketing, finance, legal, and security departments).

Monitoring and evaluation

After you deliver your change, you'll want to check it worked!

Product managers and policymakers usually measure progress with KPIs that tell you how you're doing at your goal. Just as there's a risk of using vanity metrics in product development, picking the right metric for policy evaluation is crucial. The NAO has an interesting report on how the UK government does evaluation.

Once you have learnt from your changes, you can then start the cycle again: iterating and continuing to improve on the policy. The EU publishes a lot of its own learnings and reviews publicly - you can read an summary of GDPR reviews or see the REFIT dashboard for a wider array of legislation being reviewed.

Getting into policy

The world needs more top talent in tech policymaking: especially as new tech might enable an AI catastrophe or cause a catastrophic pandemic.

If you're interested, learn more about developing policy skills at 80,000 Hours. They also offer free career advising and an excellent jobs board.

I also run the AI Safety Fundamentals courses. These are free online courses that can give you the knowledge, skills and connections to succeed with a career transition into AI safety. A colleague at BlueDot also runs the Biosecurity Fundamentals course, which is similar but with a focus on preventing catastrophic pandemics.

Footnotes

  1. Speechwriting can be useful in particular areas of AI governance work.

    However, it certainly isn't a necessary skill for domestic AI policy in the UK government. In fact, most people working on AI policy in the UK government probably don't write any speeches.

    Areas where these might be more important:

    • International policymaking, particularly on the foreign relations side of things
    • Policy research at a think-tank, research institute or university
    • Politician's offices, which are closer to the politics (more common in US policy)
  2. We run a free online course on AI governance, which covers what these policy levers are if you want to pick them up!