I'm Molly Kovite, General Counsel at Coefficient Giving. I joined in 2021 as the organization's first lawyer and have since grown the team to (soon to be) seven lawyers and three legal admins.

Over the past couple of months, a number of AI safety organizations have asked me about whether they should hire in-house counsel and if so, what they should look for. This post is my attempt to write down what I've told them: how to tell when you've outgrown outside counsel, what in-house counsel does and doesn't get you, how to scope the role, and how to run a process that actually tells you whether someone will be good at it. This is one lawyer's experience at one organization, so treat it as calibrated opinion rather than settled wisdom. I'd be glad to hear where people disagree.

How do I know when I should hire an in-house attorney?

Some signs to look for:

• You’re unsure whether you’re bringing the right things to outside counsel
• Important plans or projects have been derailed or significantly delayed two or more times in ~6 months due to unexpected legal issues cropping up, despite running things by outside counsel
• You feel uncertain about the quality and calibration of the advice outside counsel is giving you, and this is true of more than just one or two external lawyers
• People at your org are having to spend an annoying amount of their time talking to outside counsel

One of the real values of in-house counsel is issue spotting, which is something that outside counsel can't really do because they're not broadly aware of what you're doing. They're only aware of what you bring to them, so the first indication that you might want to consider in-house counsel is that you feel pretty unsure whether you’re bringing the right things to outside counsel.

Another big value add of in-house counsel is that they can learn your culture and calibrate their legal advice to your organization's priorities, risk tolerance, and communication style. Even if you work with excellent outside counsel, they're just not gonna be able to do this as well. While in-house counsel won't have the breadth of knowledge of an entire law firm's worth of attorneys, when working with outside counsel, they'll be able to present the issue to outside counsel much more succinctly than a non-lawyer in-house. They'll be able to keep the outside counsel focused on the specific issue, which can be more of a challenge than you'd expect.

They'll be able to push back when outside counsel is being overly conservative, which they tend to be, and they'll be able to drill into what the actual legal and enforcement issues are more confidently than a non-lawyer in the same conversation. They'll be able to communicate that advice back to the rest of the team because they'll understand the level of knowledge and context that the rest of their team members have.

So if you find yourself frustrated with working with outside counsel because you feel they're too conservative, you don't know how seriously to take their advice, and their communication style is not working for you, that's probably a good reason to consider at least fractional in-house counsel. Though it's also probably a good reason to consider changing outside counsel.

What are some of the downsides of hiring in-house counsel?

Attorneys do tend to be among the higher-paid employees at a given organization, so cost is definitely something to keep in mind. If your annual legal bills are in the $100,000 to $200,000 range, you're probably not going to consume a full attorney's worth of legal services.

You could consider hiring a fractional general counsel - an experienced lawyer who integrates into your organization but works on a part-time basis and may do so as a contractor across multiple organizations. You might be a bit less likely to get a really ideal culture fit or mission-motivated hire if you go this route.

Because whoever you hire will have limited expertise as compared to a whole law firm, you won't be able to just hire one lawyer and get those outside counsel costs down to zero unless what you're doing is incredibly specific (such that you can just hire one person with that skill set). Even if that's the case, lawyers don't like to operate alone, especially on high-stakes issues, so an in-house counsel would likely want to at least occasionally get a sanity check from outside counsel on important matters. And it’s arguable whether this is a feature or a bug, but having in-house counsel won’t even necessarily reduce your spend on outside counsel at all. They might come in and notice a bunch of compliance issues that need to be fixed, or notice things that the organization wasn’t previously getting advice on. So possibly more money and annoyance upfront buys you more resilience and less chaos down the line.

You could also consider hiring an attorney that also covers other operations portfolios; attorneys are often good operators who slot into other realms with relative ease. This is especially true of compliance-heavy realms like running a grants team, running a contracting team, or employment and HR functions.

This does create some issues on the margins about when legal privilege applies, although I think people are often too worried about maintaining privilege. Privilege only matters if you end up going to litigation, and most things are just not at risk of going to litigation. The privilege issue of having an attorney working on non-attorney matters is that if an attorney is acting in e.g. their capacity as an HR advisor, then a given conversation would not be privileged, because the attorney is not giving legal advice, but HR advice. Whereas if they're acting in their capacity as an attorney, then a conversation would be privileged, and sometimes there's going to be ambiguity about that.

How do I scope the role?

You probably won't successfully predict exactly what the role will look like if you've never had in-house counsel before, and that's okay. When I was hired at Coefficient Giving (Open Philanthropy at the time) the basic idea was that I could help the org save money on some of the more rote stuff, like updating our grant agreement templates and reviewing contracts and that kind of thing.

It became apparent pretty quickly that I could add more value in other ways, especially since I was way more attuned to our organization’s goals and risk tolerance than outside counsel were. So my role changed a lot over time. I don't know whether that's more or less likely to happen with legal roles than other roles, but I think in general, as long as the needs of the organization are being met, having flexibility to take advantage of your employees' biggest value adds is a good thing.

That said, you do, of course, need some sense of what this person is going to do all day every day. If there is a person who is currently working with outside counsel, you should expect that that person will only rarely need to interact with outside counsel after in-house counsel has been hired. Not because in-house counsel will do all the work that outside counsel is doing, but because in-house counsel can be the one to interact with outside counsel. The work stream of interacting with outside counsel should be taken off everyone's plate.

That means in-house counsel will have to have some basic understanding of the work of anyone who interacts with outside counsel, so getting spun up on that knowledge will be a major component of their first several weeks. It's also likely that when in-house counsel first starts, they should do a basic audit of whatever compliance processes and systems are currently in place, as well as any portfolios that might carry some legal risk.

The basic portfolios that I would expect to exist at many/most organizations are:

• Board matters and governance
• HR and employment
• Contract review (which would include grant agreements, where relevant)
• Tax and regulatory compliance
• Possibly some lobbying and political law
• Maybe data privacy and document retention

What basic characteristics should I be looking for in an in-house counsel?

You want:

• A good lawyer
  • Spots important issues
  • Draws correct conclusions from their legal research
  • When areas of law are unsettled, offers a range of interpretations that are well reasoned and defensible
  • Thinks in terms of tradeoffs and makes the available tradeoffs clear
  • Thinks creatively to generate solutions that minimize risk while achieving organizational goals
  • Has enough time practicing under their belt that they can do these things fluidly, without supervision (law school doesn’t make you ready to practice law)
• Who communicates in a way that is native to your organization
• Understands and internalizes the org’s mission well enough to make judgment calls the way the org would want when there's no time to ask
• Knows when to be detail oriented and knows when to move forward quickly with less precision

You probably want other things too.

One thing that is somewhat unique to legal hiring is that mission alignment might matter a little bit less than for other roles. I think this because the legal profession is inherently somewhat mercenary. That’s not to say that there aren't tons of practicing lawyers who are doing what they're doing because they are motivated by a mission; but ultimately, as a lawyer, you have an ethical duty to zealously represent your client no matter who your client is. That is at the heart of the profession. Every lawyer has had experience advocating for something that they weren't 100% heart and soul bought into - and they can do it well because they believe that’s necessary to the functioning of the legal profession.

How should I design my work test(s)?

It’s pretty standard in our eco-system to conduct work tests as part of a hiring process. Historically that’s been pretty non-standard in the legal profession (though is probably normal in tech now), but I’ve received almost no negative feedback about that process. When I went through it myself I found that not only did it give me an opportunity to show I could do the job despite my lack of experience in the relevant fields of law, but also it gave me some insight into whether I would like the type of work I’d be doing. It’s a win across the board.

Designing and grading a good work test can be a bit tricky, especially as AI gets better and better at short and medium time-horizon tasks. Ideally, you just pull the work test from the real world and change a few facts and perhaps amalgamate two or three scenarios into one scenario so that there's a denser set of issues to work with. You’ll likely want to test for:

• Risk calibration
• Communication style
• Ability to get familiar with an area of law and a set of facts relatively quickly
• Willingness to flag uncertainty

One thing to look out for is that someone new to the area of law that your work touches on takes an incorrect turn very early in the chain of reasoning. It can be a totally reasonable misunderstanding, or even a reasonable interpretation of text that just happens to not be consistent with how the law has developed in practice. This can lead to the person getting to a pretty legally wrong but very reasonable outcome. I recommend being very forgiving in those circumstances.

One thing I found to be particularly helpful is to frame a test as a draft of a response to a client to be reviewed by a more senior lawyer on the team. So the test is both to draft the response to the client, but also to draft the email to the more senior lawyer flagging what parts they especially think need review. This framing makes it much more likely that the candidate will actually tag uncertainties, so you'll get a much more calibrated sense of: in real life, would this person have actually sent incorrect advice, or would they have consulted outside counsel or another expert on the right things such that their mistake would have been caught? It also gives you two different frames in which to evaluate the communication style. And lastly, the AIs, when I last did this (June of 2026), had a harder time doing well in the meta frame than they did with just a straightforward email-a-client test.

You'll also want to make sure your work test isn't just a dense series of legal issues, and instead is a broad scenario of "Hey, we want to do this thing. Here are some concerns." That's how issues get presented to lawyers, and often the client's concern is totally orthogonal to what the actual legal concerns should be. I don't like work tests that have red herrings, but I do think it's worth including legal issues aside from just the one that the prompt specifically draws attention to.

One thing that is particularly valuable to evaluate in legal writing is whether the lawyer has appropriately identified and summarized what they understand your goal to be at the beginning. That way, you can understand and make sure that you're on the same page about that. Have they created a document that allows you to skip through the reasoning and find the conclusions, and then go back and double-check the reasoning if you want to get into that level of detail? That kind of formatting ends up being of outsized importance in legal writing.

It's probably useful to have both a lawyer and a non-lawyer grade these work tests, and they don't necessarily have to reconcile their grades. The lawyer is going to respond more viscerally to incorrect legal conclusions and not notice when someone is drifting into slightly difficult-to-understand legalese. The non-lawyer is probably in a better position to evaluate communication style and also how informed they feel about risk calibration.

What should I look for in interviews?

This is probably idiosyncratic to your particular organization, but here are a couple of thoughts.

One theory I've had as I've built my team is that, by default, people don't really want to talk to lawyers. It's a bit like being sent to the principal's office. So to compensate for this, I've tried to build a team that is friendly and fun to interact with, in addition to being mission-oriented and excited about the work that our organization is doing. If your lawyer is someone that people like gossiping with, they're going to know more about what's happening at the organization, and that allows the lawyer to better protect the organization.

I don't think that's absolutely necessary to be successful in-house counsel. What people care about more than anything is getting good, calibrated legal advice and coming away feeling like the lawyer had their back, rather than feeling like the lawyer was trying to justify their icky feelings about what you're trying to do.

So aside from maybe considering whether this person is someone that colleagues would genuinely look forward to interacting with, just look for the things you traditionally look for in a lawyer. Do you feel confused about what they're saying? When you push back on something do they get flustered or double down, or do they pause and consider? Do they react as though risk is a moral rather than legal issue? Do they go on at length without allowing you an opportunity to interject or course correct?

If you’ve had experience with outside counsel, you’ll probably already have a feel for characteristics of lawyers you do and do not like working with. Think through those interactions and jot them down!