Our opinions

Great software has opinions. Here are ours.

Five strong views on how enterprise climate risk management should work — and how software should help companies do it well.

Five strong opinions

01
Climate models do not predict the future.
02
Risk assessments are useless until they are used.
03
Reporting is not the point. But it is the most important thing.
04
AI is amazing, but insufficient.
05
Climate risk isn't special.

Why opinions matter

Good software helps a customer do a task. Great software has opinions - it shows the customer how the task should get done. Like any opinion, that might alienate people who disagree, or think they know better.

The worst thing software can be is opinion-less. At best, that makes it bland. At worst, it becomes a Frankenstein's monster: 1,000 features, built to please everyone, used by no one.

So, here are a few of Beehive's opinions on the category of enterprise climate risk management.

Climate models do not predict the future.

It's a common experience that if you ask five model providers what your flood risk is at one address, you'll get five different answers. And if you compare their assessments for 30 years from now against what actually happens, you're likely to be disappointed in the results.

No model can predict the future. That doesn't make this process unhelpful. It's just important for you to understand how these models work and what they mean.

At Beehive, we acknowledge the limitations of climate models while valuing their importance in understanding the world around us. So instead of pretending that our models are right and the other guys' are wrong, we lean into two core philosophies: transparency and auditability.

Transparency means that we show you everything about how we build our models. Anyone can read our documentation. Our approach is to take existing best practices - government data sets, published papers on how to build models - and bring them into usable software, with as much transparency as possible.

Auditability means that every data source comes from a credible database you can reference, like FEMA or the EU Joint Research Centre. It means that when we share risk to an asset, we provide the inputs that led us to that estimate, give the customer tools to adjust some of those assumptions, and give customers and auditors tools to follow the logic all the way through. We know that your lived experience in a place or at a building may be different from what a global model says. So you should be able to control that.

So it is our strong opinion that climate models are not good at predicting events. But they are still useful. What is important is that you understand how your model provider reached their conclusions, and that you have the tools to incorporate your own knowledge about the business, so that you can use the information wisely.

Risk assessments are useless until they are used.

The proliferation of climate modeling means that anyone can get a risk score for an asset around the world, sometimes for free, and often quite credibly. But knowing that the risk exists is, on its own, not enough to be useful to a company. What matters is the action you take on that assessment.

At the very least, you need to inform the right people. The sustainability director knowing that a new real estate acquisition is in a wildfire zone doesn't help the business prepare for a wildfire. You need to get that information to the manager of that property, so they can consider it in their decision and their business continuity plans.

Sometimes, the risk assessment is valuable because it feeds a compliance report. This common workflow informs how Beehive builds our product. If the majority of our customers need risk assessments to comply with a law, and the law requires not just an assessment but a report (TCFD, IFRS S2), then it doesn't really solve the customer's problem to just assess risk. The risk assessment, on its own, is not useful until it goes into a compliance report. Beehive needs to embed reporting directly into the product experience, and it should be seamless to move from risk assessment to report.

The best thing to do with a risk assessment is to act on the results. Draft business continuity plans, build flood walls, buy insurance, offer employee training. The point of a risk assessment isn't to be proud you did it. It's to be proud that you did something with it.

So it is our strong opinion that a risk assessment on its own - without a way to credibly communicate its results to the right stakeholders, to report on it, or to act on it - is useless. But when you connect it to those workflows, it can be the start of something very useful.

Reporting is not the point. But it is the most important thing.

In a recent survey of 100,000 sustainability professionals, zero said that they got into this work for the love of reporting.

The survey is made up, but the point stands. Beehive - like many of our customers - does not see reporting and compliance as the point of all of this work. We want to help build companies that are more in sync with the world around them: decarbonizing to limit the impact of climate change, and managing risks to build resilience. Reporting does not play a direct role in that, and we - like most of you - wish it did not distract us from our most impactful work.

But we - like most of you - know that reporting to align with regulatory and investor requirements is the reason most corporate sustainability work exists right now. Compliance budgets pay for the majority of sustainability work, and reporting to align with raters, investor requests, and regulatory frameworks is what most of us spend all of our time doing. So as much as it is not the point of our work, it is the most important thing.

Therefore, we do not shy away from the problem of reporting. We lean in. Everything in Beehive integrates with the reports our customers need to write. AI can draft TCFD and IFRS S2 reports based on the risk work you conduct in Beehive, and we build technology to make this necessary step as quickly and to as high a standard as possible.

You can export everything in Beehive as a spreadsheet for further analysis. You can export some modules as PowerPoint presentations, so you can report to your board or management committees.

So it is our strong opinion that while reporting is not the point, it remains the most important thing. And if both of those things are true, we think it is our responsibility to help customers report very, very well and very, very quickly. Then, you can get back to impactful work, like decarbonization and resilience.

AI is amazing, but insufficient.

Like everyone in the technology industry, we're enamored with AI. It's kind of awesome.

In Beehive, we use AI to run transition risk assessments, to draft climate risk reports, to understand climate models, and more. The results are amazing, and getting better every day. We can say with some confidence that an IFRS S2 report drafted with Beehive's AI in three minutes - compared to a consultant who can take four months to get you a draft - is pretty good. In fact, one Fortune 500 customer told us Beehive's answers were about 65% perfect.

Not bad. But 65% perfect is far from finished. While we want it to get even better, we have a strong opinion that AI will not be doing the last mile, ever.

So our goal is to consistently and incrementally drive up that 65% number, but we also acknowledge that very capable humans will always take the AI's output over the finish line. We do not build the AI to get to 100% perfection, because we don't think most customers will ever accept the AI as 100% perfect or complete. We build it to get a good enough output very quickly, then we build technology and human services around that to help our customers get to a final report.

That technology can be even more AI - like asking the agent to make an answer longer, or remove some of the financial statements, or hedge the response. But it is often good old-fashioned software - like editing tools, benchmarking modules, and audit trails.

And human expertise must underpin all of it. That expertise can come from our customer success team, or consulting partners, or SMEs at the customer. But it needs to be a human responsibility, in the end.

So it is our strong opinion that AI is amazing, but insufficient. Our favorite analogy is that AI is like your intern who recently graduated with a JD/MBA from Harvard. Super impressive! Unfortunately, they got those degrees straight out of college and still live in their parents' basement. Smart, but needs coaching.

Climate risk isn't special.

Climate risk does not deserve any more or less attention than any other risk that may be material to a business.

Companies have advanced systems to manage cybersecurity risk, geopolitical risk, succession risk, regulatory risk, and plenty more. Climate does not deserve any more attention than any of those just because it's climate and we care about the planet. But it does not deserve any less attention when it is material.

Not all risks are created equal. To a tech company, cybersecurity risk is more important than climate risk. To an energy company, that might be flipped. To a hospital, they might be equal.

We do not think that all three of those industries should invest equal amounts in climate risk management. But we do think all companies should be able to quickly and credibly evaluate the materiality of climate risk, then act on it accordingly. If you determine that climate is not a material risk for your business, great. You shouldn't have to spend much time or money to reach that conclusion. If you determine that it is a material risk, you should have the technology and expertise to act as you would in the case of any other risk.

So it is our strong opinion that climate risk is not special. And because it is not special, it should be easy, fast, and cost-efficient for everyone to have a baseline program. If you find climate risk to be material and deserving of more attention, you should have the resources to help. And if it is not relevant to your company, you should have confidence in that statement.

The synthesis

What these add up to

You may have noticed that each opinion holds two ideas at once. Models can't predict the future, but they're still worth using. A risk assessment is worthless on its own, and the beginning of everything once it's used. Reporting isn't the point, but it's the most important thing. AI is amazing, but insufficient. Climate risk isn't special, which is exactly why it should be easy.

You may also have noticed that none of them flatter us. Every one is a limit on what we'll claim: our models can't predict the future, our risk scores stop short of useful, reporting drives this industry, our AI stops well short of finished, and our entire category may not even be material to your business. We could probably sell more software by saying the opposite of all five.

Like we said at the top, opinions may alienate people. If you want a model that hands you one confident number, or an AI that claims to finish the job, or a technology that goes no further than assessment, we're probably not for you. That's the cost of having opinions, and we'll pay it. We'd rather be clearly wrong to some people than vaguely agreeable to everyone.

We also think most of this category ends up here eventually. We'd rather say it now, out loud, on a page you can hold us to. If you think we've got one of these wrong, tell us. We hold these strongly enough to build a company on them, and loosely enough to want the debate.

See the opinions in practice

Climate risk software should help you understand, report, and act.

See how Beehive turns transparent, auditable assessments into reports and risk-management workflows your company can use.