A tool to improve your regulatory foresight - Steven Brennan, Foresight
In this episode I speak with Dr. Steven Brennan, co-founder and CEO of Foresight.
Foresight is a software platform that helps organizations to make sense of complex and fast moving chemical regulations, so they can always be ready for what's coming next.
Steven and I have a nice conversation that shows how he turned his years of regulatory experience into a software solution that solves a key challenge - how to get access to key insights faster to understand and manage regulatory change.
Listen to this conversation if you want to get a look into the world of startups and software as a solution (SaaS) from a chemical regulatory point of view. We also talk about aspects of building and growing a business, and some reflections on life as founders and keeping everything in balance.
Our conversation covers:
Steven's background and career journey
The story behind Foresight - identifying an industry need, and developing a solution
Features of Foresight - how the tool delivers advanced regulatory insights
Applications and limitations of AI in chemicals regulation
REACH Article 129 - Member State safeguard clause
What kinds of customers are using Foresight?
What process and considerations go into developing new features?
Building the company - from startup to commercialisation
The customer buyer journey
Steven's work as a co-opted member of the ECHA Socio-Economic Analysis Committee (SEAC)
Steven and Chris swap stories about life as founders
Follow the link to find out more about Foresight and start your free trial: Foresight | The future of chemical policy monitoring
Prefer to read? Here is a transcript:
Chris: Hello everyone and welcome to the Chemical Journeys podcast. Today I'm speaking with Steven Brennan, who's the co-founder and CEO of Foresight, a software platform that helps organizations make sense of complex and fast-moving chemical regulations, ensuring that they're always prepared for what's coming next. Steven, thanks for joining me.
Steven Brennan: Very much, Chris. Really nice to be here.
Chris: No, it's great to have you on. You know, I'm really excited about this one because we hear so much these days about software as a service and new tech startups — the economy's really kind of shifting. But it's not the sort of thing that you typically think of in the regulatory compliance world. So it's great to have you on this podcast to really open up this world to us and help people learn a bit more about how it all works. But before we get into that, why don't you tell us a bit more about yourself and your career journey to date.
Steven Brennan: Absolutely, Chris. I suppose the first thing to say is, I must say you're bringing a lot of industry expertise forward with your podcasts at the moment. I was listening through a bit of your back catalog in the course of this week to prepare for this, and I think you've got some really good speakers. So I'm very encouraged to be able to share my experience.
Chris: Thank you.
Steven Brennan: Like a lot of the guests I listened to over the week, you know, background in chemistry — I think that comes with the domain, though not always for everyone. I had an interest in chemistry from a very young age. One of my earliest memories was mixing up various shampoo bottles in the shower, trying to create some new formulation. And my mum wasn't too happy with that. What she ended up doing was making my sisters wash their hair with my concoctions so we wouldn't waste the shampoo. But that's what led me to study chemistry at university. I ended up doing a PhD in chemistry as well. My preference was for physical chemistry, and my thesis — the area of research I conducted — was the creation and application of superhydrophobic surfaces. These are essentially super water-repellent surfaces, and at the time they were touted for lots of different kinds of applications: moving droplets around on chips, for microfluidics and so on, or super water-repellent technical clothing, that kind of thing.
But probably what's standing out to some of your listeners is PFAS.
Chris: Yeah.
Steven Brennan: I was a very early user of PFAS, and if I'd known then what I know today, I would've been a lot more careful with the risk assessments. With that, it opened up a role that I got within industry after I'd done the PhD.
I moved over to England and got to work for Rolls-Royce, an aerospace and defense company. I suppose I landed into this area of regulations and substances of concern just by pure chance. At Rolls-Royce I was working on their REACH program as a materials engineer, which meant that I was getting very hands-on with helping to substitute hazardous chemistries from manufacturing processes. That was really interesting experience because, first of all, it was a new thing for lots of companies and we were really just trying to set up what that substitution program would look like and dealing with all of those teething pains. Little did we know at that time that when we were setting that all up, lots of organizations would still be going through those same pains today.
That was my journey into the regulatory space, because a lot of those substitution programs were driven by regulations such as REACH. We were using some of the outputs from those programs to inform REACH applications for authorizations, or how we respond to restriction proposals, and so on.
I did that for several years, and then went into another defense company for a year and realized it was pretty much just the same — a lot of downstream user requirements. As I was getting more interested in the regulatory space, I decided that I would join a consultancy, prior to the 2018 registration deadline, to get more experience on the upstream manufacturer and importer requirements, particularly around registration. I was more trial by fire — I joined in October, the 2018 deadline was around six to eight months later, and I was quickly managing regulatory projects, which I maybe didn't have a lot of experience around at the time. But towards the end of that period, I felt I'd gotten reasonably competent at it. I would say serviceable.
I did that for a few years, working on registrations, consortium management, then taking on leadership for that team within that consultancy and trying to expand that service line into REACH authorization requirements and other related work.
With that, I suppose I felt that I had done the industry side and the consultancy side. I thought, well, let me try and see what other spaces I could work in. So while I went back into industry for a year, I started thinking about how we could better solve some of the problems I'd encountered over the course of my career. A good friend of mine, Jonny Campbell, was moving back from England as well — he'd been living there and moved back just after COVID. Together we sat down and came up with the idea of using software to solve some of these problems. That was how we set up Foresight.
Chris: Thanks for painting that picture and that story, because in some ways it's quite a typical regulatory journey. You got the PhD, the whole world kind of opened up to you, and then you ended up working in industry and got caught by the regulatory machinery — addressing all those needs and challenges in the regulatory world. But you've explored all around that. It sounds like you've been through the different levels of organization. And yeah, I guess what you highlighted there is that with a company like Rolls-Royce, a lot of the regulatory challenges are downstream user challenges, so you've got to deal with stuff that's coming down the tracks that you don't really have that much control over.
Steven Brennan: Absolutely. I suppose there are just a couple of things that stood out to me from that time. One is the volume of change that was coming down the pipeline, as people say, but also the impact that some of these changes would have on an organization — not necessarily specific to Rolls-Royce. During my tenure there I was also part of some trade associations, such as the International Aerospace Environmental Group. I could see there that there were lots of organizations dealing with the same problems: the volume of change and the impact that this change would have on operations, footprint, manufacturing processes, materials that they used to provide certain performance envelopes to their products, and so on.
I think one of the challenges that industry was dealing with — and probably still dealing with to a degree — is chromates. They're a wonderful chemistry in lots of different ways, but the challenge is that they've been embedded into many different manufacturing processes, which makes them really hard to untangle. There's no like-for-like substitution for chromates. They're used on metal alloys as much as they are on different kinds of metal substrates, primarily to provide corrosion protection but also other functional requirements as well. And all that complexity has created a scenario where it's just really hard to substitute.
With that said, although they're still dealing with some of that challenge at the moment, there have been some notable substitutions for that industry. For example, trichloroethylene as a degreasing agent had an Authorization sunset date — I think April 2016, if I remember correctly. Lots of different companies across markets have been able to move away from that particular chemistry for degreasing. So change can happen. But as I heard from one of the leadership team at Rolls-Royce at the start of my career: any change creates a risk. And that's really all about what work needs to happen in order to mitigate that risk — and that's where the cost comes in. That means that substitution in itself is not just a technical challenge, but an economic challenge as well. Changing hazardous chemistries is probably not top of shareholders' minds, but it's definitely something that needs to continue.
Chris: Yeah, it is kind of like a blind process in a way, because we have all this technology that we've used and these applications that use chemicals, and perhaps the health or environmental impacts of those chemicals in those uses might not have been fully understood at the time that technology was evolving. And now there are these new issues or new questions emerging, and things are being unpicked here, there and everywhere. It's a blind process that creates a lot of complexity.
Steven Brennan: Yeah. Just on that, Chris — do you mean the policy-making process is a little bit blind, or maybe the process by which companies would use these chemistries?
Chris: No, I mean more discovering that a certain chemical has a certain property that triggers a regulatory impact, which then turns on a process that ripples through economies.
Steven Brennan: Absolutely. It's something I always lean into when talking to people about the REACH regulation who've never had to work with REACH before. What it has done is created a dataset on substances that we never had before — certainly not to the depth and volume that's being created through the registration process, for example. And what that has unlocked is a volume of information. I say it's a backlog of information which regulators are working through to identify risks to human health and the environment. That whole process in itself — I don't know when or if we would ever get to the end of it, to fully understand the full set of risks. But one of the challenges then is the volume of information being worked through. And that's one of the reasons why we set up Foresight — to try and generate key insights from that information so that organizations can have the foresight to know what's coming down the pipeline for them. It's a bit of a horizon-scanning element, so that they can try to preempt what these regulatory proposals mean and what impact they would have on their operations and footprint.
Chris: Yes. I recall this from when I was working in more general regulatory consultancy areas as well. We talked about this being almost like the holy grail of something regulatory professionals could use or would benefit from. Because other than that, they've all got their own networks and resources that they're looking at, but there's always that little thing keeping you awake at night — that there's a regulation or a change somewhere that you've missed and it's going to have a major impact on your business. So, why don't you tell us a bit about Foresight, how did it come about, and what is it there to do?
Steven Brennan: I suppose the first thing to say is that Foresight was born out of my experience within industry and consultancy. I'm not technical at all, but in order to try and solve this problem — whenever I was working in industry, I was building SQL databases using Microsoft Access. Excel spreadsheets and emails are probably the go-to standard for a lot of organizations. And I had to build the SQL database so I could better relate substance information to bill-of-materials data. I was trying to solve a particular problem there — I'm not software-inclined, I'm not a developer. My version of that database was useful for a time, but it had limitations as well, and that was back in 2014 or so. That's really as far as I got with it. In any subsequent organizations I joined, I probably did something similar, whether it was an Excel file or whatever. I was doing that for clients too, whenever I was working in a consultancy doing impact assessments and that kind of thing. But that's just really as far as I got with it.
What happened was, we started kicking around the idea of Foresight in 2023. OpenAI dropped its GPT-3.5 model and we started playing around with it. Like everyone else, we were using it as a search engine, trying to find out what it could tell us about the regulatory space. I put in some ideas around: well, tell me the REACH registration requirements for a company that imports one to one-and-a-half tons of sulfuric acid into the market. The answer it would give me would be a very positive reply, but it wasn't accurate at all based on my experience. But with that said, we used it a little bit further and played around with some of the other tools that came with GPT, not just GPT itself. And we found that it was a really good way of interrogating large datasets to pull out some semi-structured information. You could throw it a piece of legislation or a proposed piece of legislation, you could throw it meeting minutes from an agency meeting, and you could extract certain information from that — primarily using its machine learning toolkit, but also using other toolkits as well.
What we tested early on was: how much of this information can it take? How quickly can it generate structured information? How can we take this semi-structured information and deliver that to users so that they can make sense of it and identify if something's coming down the pipeline that would impact on their products or operations? So that was kind of the early days of where we got to with Foresight.
One of the other things to talk about here as well is AI generally — we're trying to stay away from saying that we are an artificial intelligence platform, because, first of all, we don't control the definition of AI. Everyone has their own definition of AI, and therefore we can't control what people understand when we say something like that. What we're doing is primarily using software to try and help companies stay ahead of the ever-changing policies and legislation in different markets around the world.
When we talk about AI, one of the tools we would be using is OCR — optical character recognition. I remember in the early days having to go through nearly 2,000 safety data sheets, pulling out substance information: the substance name, product name, CAS numbers, EC numbers, so that I could create that database to find out if a policy change impacted my bill of materials. With OCR, that's a first step to do that automatically. It can take something like a safety data sheet — even if it's in a character-ready format, or if it's a photograph, across all the different software formats that a safety data sheet can exist in — it can review all of that, extract the text, and combining that with something like machine learning can help you extract more structured information from a safety data sheet. That probably wasn't possible before. So the tens, maybe hundreds of hours I spent copying and pasting is now literally a couple of hours to process a few thousand safety data sheets using software. That's just one application we would use it for at the moment.
Chris: That's amazing. I was thinking about that a while ago while doing some consultancy work — thinking, oh, wouldn't it be so useful if you had all this information that's buried in safety data sheets in folders and photocopies. You really need that to enable this kind of regulatory mapping and risk screening. Is that something you invite your customers to do when they sign up with Foresight — load that kind of information into the tool?
Steven Brennan: Yes. So at the moment, the product is still in its early days. We have a tool on the back end of Foresight — it's not user-facing at the moment, so the users can't do this themselves, but we have a back-end tool we can use. Customers can send us their safety data sheets, or they can live on their website or wherever, and we can take those safety data sheets, put them through our Foresight tool, and extract all of that substance information and other related information. That can then live on our platform as a database.
What we also offer is tracking of different changes in chemical policy around different markets around the world, so that whenever there is a legislative proposal — and that legislative proposal is probably unstructured as well, but it also contains substance information — we can put that through the Foresight tool too. And ultimately what we can do with our databases is link the user's safety data sheet or other data about their products or operations to those market changes. So we've got this legislative proposal calling out this substance, and then we can make that link, and the user is just notified to say: this new legislation in Peru is going to impact on these products, as an example.
Chris: That sounds really valuable. Another thing I was thinking is that often you have multiple identifiers for a chemical — you can have a CAS number, a name, an EC number, and even multiple identifiers if the chemical has a complex composition, where there can be different CAS numbers associated with it or different chemical identifiers. So is that something that Foresight can help with?
Steven Brennan: It is a really good question, and a really difficult challenge. What we're building — and what we have proof of concept for — is an algorithm that takes the substance information available to it. It could just be a substance name on an SDS or in a legislative proposal, or any other dataset. It could be the substance name, an EC number, a CAS number, or a trade name. What this algorithm does is take that one piece or several pieces of information and try to match it against our internal database. And with that, it gives an output to say: this piece of information, or this document, says this substance — we think it's this substance with this IUPAC name, with this EC number, with this CAS number — and we're confident; our confidence rating is like 99% because there's an exact match somewhere. Or it could be, because of some missing information, we would give it a different rating, say 80% accurate or whatever. And then anything that's really lower, we would invite user review of that: we think it's a 50% match here, this is what it says in this document, this is what it says in our database, and we invite you to review that and confirm whether this substance is actually what the algorithm has identified.
And we solve for the problem that way — automating as much as possible so that the expert only spends time on their review and reasoning, rather than wasting time managing information.
Chris: Yes. Yeah, and I appreciate that it's probably a journey in terms of training and improving the tool. I guess it's good that you're inviting your users to participate in that as well, because then they have the opportunity to put information in, and they probably also appreciate that there's a hint there might be a problem — because that's the other thing, right? Sometimes they just need an early warning that they need to go and check. There's nothing to worry about here.
Steven Brennan: Absolutely. One of the things we do with all of the content we create on the platform is expert review. On the platform at the moment there are lots of market updates — Japan's proposing an MRL for this substance, or there's this new legislative proposal from the EPA. All of that information we publish to our users as alerts. And if they're tracking by substance, we'll tell them exactly what substance is being called out. But everything we publish is expert reviewed.
Chris: Nice.
Steven Brennan: We know that AI is good at automating a lot of it, but we know that AI is not good at doing all of it — especially when it comes to reasoning. We don't try to use AI to reason in any complex way. Any new AI model that gets released — I think Google is shipping Gemini 3 soon — I'll test that with a prompt like: what are the REACH registration requirements for a company that's importing this volume of substance into the EU market per year? Again, it's very specific to its general knowledge base and it probably won't give me a satisfactory answer, but I at least test it just to see if we're getting closer to the holy grail.
Chris: Yeah, yeah.
Steven Brennan: You're always going to have to have human reasoning.
Chris: Yeah, no, that's reassuring in a way. I think I've had a similar experience to you as I've been dipping more and more into using AI tools, especially as I've been trying to curate the content that this podcast is creating. Because I'm conscious that there's a wealth of information in this podcast through the back catalog. Some feedback I've had from people is that they just don't have time to sit down and listen to 60 minutes of a podcast. I've been trying to work with the transcripts and see if AI can give me anything I'd actually want to read. And I have to say that I'm struggling at the moment — maybe that's just perfectionism. But the term "AI slop" really does ring true for me when it comes to that sort of thing. And I think people can tell when they're reading or engaging with content whether there's a human touch to it.
Steven Brennan: I think it's important. For your specific use case — the translation of audio to text — yes, it's a challenging thing to solve, especially when it comes to dealing with different accents. And you have a lot of non-native English speakers, although a lot of the time they speak better English than I do. There's just so much complexity involved. It's going to be really hard to get a tool that gives you a verbatim transcription. But something we've been using it for as an example is EU Parliament debates, particularly around PFAS, the REACH revision, that kind of thing. That's all published by the European Union. You can get access to it, you can push that through an AI tool, get a transcript, and it's a good average of what the conversation was about. Plus there's translation involved on their side as well, because they have their own translators, and a lot of the MEPs who are debating are doing so in their own language. So what you're really getting is an average of what was said — not exactly what was said a lot of the time, which creates more challenge — but it gives you a good overview of what was discussed.
Chris: Yeah. I can imagine the value of that must be immense, because this is information that could be really impactful. The further out you can cast the net, and the more information you can process, the better chance you have of picking up a signal that could be strategically really important for a business.
Steven Brennan: Yeah, absolutely. A lot of people might know Foresight via our weekly newsletter. What we do is publish some news and insights gathered over the course of the week. We used to be a lot more consistent on this — we used to post a lot of stuff — but now we're focused on building the product, so it's a bit challenging to find the time. We post about five articles per week. Some of those articles are around things we've found with Foresight over the course of that week.
Just last week there was an article we posted around a Netherlands Parliament discussion, where a document was published on one of their websites. It was around REACH Article 129. Now this is a mechanism I'd never heard about before, but it exists in the REACH regulation — and I've actually read REACH a few times, but I've never dealt with this particular article in earnest. It's where a member state can notify the European Commission that they want to implement a ban under some of the REACH framework for whatever reason. A particular issue in the Netherlands at the moment is steel slags — there's an argument about whether it's a waste or a byproduct. It was being produced and used for certain applications such as construction, and it was identified that this was potentially creating a hazard that needed to be managed. They implemented some kind of restriction, as far as I'm aware. We caught this discussion in the Netherlands Parliament, surfaced it, reviewed it, and pushed it out to the platform for our users.
Digging more into Article 129, it's been used a few times before — by France for bisphenol A in thermal paper, which led to a REACH restriction. I think it was also used by Denmark in the past around CMRs, which led to an Annex XVII restriction around CMRs. So that's where we get an EU member state Parliament debate, catch the information there, and it then goes for discussion at the REACH Committee at the EU level — they'll discuss that this month, and then that can lead to further action down the line. This is different from what you'd have seen if you were just monitoring a website for updates; you might not have caught that one.
Chris: Mm-hmm.
Steven Brennan: Early days, and we don't know exactly where it would go, but I just thought — as a REACH geek — that's something I hadn't really seen before in action.
Chris: Yeah, no, that's really interesting. And I guess it's an important reminder for people that the REACH regulation is massive and has many mechanisms and many lines feeding into it, particularly from member states who often have their own internal scientists and policy people addressing the needs of their populace. They'll come up with their own initiatives, and yeah, Article 129 sounds really interesting in terms of how it feeds into the top table at REACH. We'll have to see how that one develops. I did want to say to you, because I was looking at your website again in advance of this — I was really pleasantly surprised with your newsfeed. And it reminded me of the news pages we used to get from Chemical Watch before they went behind a paywall. I used to quite enjoy looking at those, just seeing what's going on in the regulatory world as another regulatory geek. I've been missing that. So I recommend people check that out.
Steven Brennan: Thanks very much, Chris.
Chris: Maybe we can just talk about Foresight a bit more — about your customers. Who are the kinds of customers you're working with, and what are they saying about the tool?
Steven Brennan: It's a good question. It's quite varied, Chris. So there are lots of different actors in this space — you've got your service providers, consultants, government agencies who want to stay on top of regulatory developments, NGOs who are interested in tracking things. But the customer we're targeting at the moment, and the customers we have most of, is industry itself. Primarily organizations that are interested in monitoring these updates. They tend to have big footprints — they work in multiple markets, markets that don't necessarily speak English as a first language. They can be across various different sectors, everything from industrial care and specialty chemicals through to consumer goods. And one of the key features about them as well is that they have big product stewardship or regulatory affairs teams — more than a handful of employees working together to ensure the organization is tracking policy changes and identifying anything that may have an impact, then collaborating to maintain compliance or make changes to mitigate risks, that kind of thing.
Foresight in itself is a policy monitoring tool at the moment, but what we're kind of moving it towards, based on our customer feedback, is like an operating system for product stewardship teams and regulatory affairs teams. So this is for companies that maybe are tracking changes via news articles or agency updates across lots of different sources — we pull that all into one feed. And then we're also building out the toolkit to allow organizations to collaborate around that. So you can imagine getting something into their feed — we call it an alert, because 'news article' is a bit more editorialized, whereas an alert is more like: this is exactly what's happened, this is exactly what it means. They get that into their feed and then start to collaborate with their colleagues on the platform, tagging in their manager or a teammate and saying: I think we need to do an impact assessment here. Then they do the impact assessment, capture the results, and that all sits within the Foresight platform. That creates an auditable record so that if there's anything to double-check down the line, they have that full history in one place. It's also useful for managers of teams who have KPIs around monitoring or management of these issues, and want a quick way of reporting on what the team's been focused on over the last week, month, six months, or whatever it needs to be.
Chris: Yeah, that sounds really valuable, I think for some of these larger teams. Just from my own experience working in big teams, I've seen all kinds of varieties of solutions and processes that companies implement internally. A lot of the time there's a lot of manual churn going on, a lot of communication friction between different people getting action on things. And of course that audit trail is so important in the regulatory world. So yeah — have you had good feedback about the tool?
Steven Brennan: Absolutely. So the reason we started out with the monitoring feed is based on user feedback — and it's based on user feedback that we've landed on this organizational workspace, and then on the need to have some kind of collaboration within the tool. I could imagine these things just based on my own experience, but it's different — you need customers pulling these things as well. My co-founder Johnny, who is in product design and user experience, his school of thought is: you need to have several customers say the same thing about what the product needs to do before you build something. That's meant I haven't gone ahead and worked with the development team to build out all these fancy features I've dreamed up myself. It's all about customer needs and what they prioritize — that's what informs our roadmap. Things like safety data sheet ingestion and extracting information have been identified by our users as important to them. We'd build a prototype, test it, make sure that it works. We've got a tool that works on the back end now, and as we get customer feedback, that's something we would ship as a feature. So it sits on our roadmap, but at some point it'll become available to our users.
For me anyway, it's been interesting going from academia to industry to consultancy to software development. I never thought I'd be in this space, but I'm hoping that the tools we create will help people like myself in previous roles — just to collaborate better, stop the manual working and manual handling of information, and not always have that doubt in your mind that you've missed something because it was sitting in a spreadsheet somewhere, or you needed someone to translate from Japanese to English. It just pulls all that together in one place.
Chris: Yes. I think you have the advantage that you yourself are a deep expert in this area. You've lived the life of a regulatory professional, and then you've also worked in consultancy. So you have an idea of what companies are prepared to pay for and what their needs are. And I could totally feel what you said there about having an idea and then wanting to go off and build it, because often you think, oh, wouldn't it be great if this was such-and-such, and then you think, yeah, I'm going to go do a bit of work on that, and then you realize, oh, that's a massive project. Or you could invest a lot of effort in building something that people don't really want. It sounds like you're very clued in on software-as-a-service best practices — we've learned a lot about that in the last 20 years. The really important thing is testing, like you say, doing small tests to see what people go for. It's good that you're getting feedback from customers and information about what their needs are. Another thing is that customers don't always know exactly what they want, or they might think they want something, but when you go off and build it, it might not sell. So you've got to use a bit of your own intuition as well, I suppose.
Steven Brennan: Yeah, absolutely. It's more of a sanity check. One customer uses different language than another customer, and it's about reconciling that and being able to explain to our development team that, actually, these are the same feature request. They both mean the same thing — this is what it should do, this is the job to be done. It's definitely helpful to have sat in the user's seat so that you can empathize with the problems they're trying to deal with.
Chris: Yeah. And you've got the ideal two-founder setup as well. It sounds like you have a great relationship with your co-founder and you're serving different functions in developing this thing.
Steven Brennan: Absolutely. I've known Johnny since I was 12 — just to show my age, that was about 28 years ago now. We stayed close friends all that time. And then we have a wider team as well — we're five of us now.
Chris: Wow.
Steven Brennan: And we've got some support from some really good service providers in this space as well. There's work we're doing here that requires the old-school consultancy approach to get sources of information and so on. We're working with some superstars there as well.
Chris: Yeah, no, that's amazing. It's great to hear it's growing and you're up to five people — you really are on a journey with this. Maybe you could tell us a little bit about where you're at in terms of commercialization, and what was that process like for a startup? Because I think that's really an unknown area for people.
Steven Brennan: Yeah, absolutely. This is something I had to learn as well. Coming from industry, where things are slow and you do your annual business deployment plan and annual cycle of work to be done — startups are really working on daily, weekly, or even monthly sprints. Regular catchups with the rest of the team, hyper communication, several times a day. Not traditional meetings where people sit around waiting for people to talk, but literally quick catchups: have you done that? I've done this. Where are we at now? What are we going to do for the next hour? And we'll catch up in an hour's time. That's the pace they're working at, and it's quite intensive.
The other thing to say about the startup scene is that we are venture-backed. We bootstrapped until around June or July of this year, and then we got pre-seed investment from a couple of VCs. That's enabled us to scale the team, offload some of the work I was doing manually, and free up my time to have conversations with customers. That's been going really well at the moment. We actually have a bit of a waitlist because of where the product is, and we want to get to self-serve onboarding, but at the moment we're supporting the onboarding process. We're talking mid-November before we'd have capacity to onboard and pilot with new customers, and this is just the start of October, so we're talking a few weeks out.
Chris: Wow. That's always a good sign if you're oversubscribed.
Steven Brennan: Yeah, absolutely. Every customer brings their own unique experience, and all of that learning is only really good for us and for our other customers as well. One of the things I've learned in this space — and maybe what I didn't see as much in consultancy — is that we're getting probably more insight into how an organization works, as opposed to what I was doing in a consultancy space, where the client had a fair idea of what the project needed to be and would issue an RFP along those lines. You'd have conversations, kick it around, see what the consultant would come up with, and then you'd have a proposal and contract on that basis.
Well, we're offering a solution — it's one thing, and it's a subscription. Users can sign up to pay a monthly price, and then they have access to the toolkit. Through those conversations we're kind of surveying our customers: how much time would you spend doing monitoring on a weekly basis, for example? And these organizations are spending 70, a hundred hours a week across the organization just reading information to find out what's happening in the markets they've got products or operations in. We're getting insights that way, and then we can feed that back to their management and say: did you know that your team spent all this time doing this thing? We think we can get an 80% improvement here — that's a good return on investment. And then that opens up wider conversations: oh, yeah, they're doing this — but if you help solve that, you might also solve this other thing we're dealing with around KPIs for the program. And we're getting good practice from various customers and being able to feed that back into our experience as well. So that's one of the marked differences from what I've seen in consultancy and what we're doing here within software.
Chris: Yeah. In some ways this sounds like proper consultancy rather than the transactional kind — I need 10 registrations or something like that. So it's nice. And yeah, I think this is where consultancy really earns its money if they're good. And I hear what you're saying as well about how it's different in terms of the buyer journey, because you often have to get through more stage gates if you have a more ambiguous or uncertain service. You really have to educate your customer and help them realize the vision. They need to probably convince people within their own organization as well that this is something that's really needed. So yeah, it's an art in itself, really — getting people to see what you can do for them.
Steven Brennan: Absolutely, definitely. I've learned a lot over the course of the last two years or so — from the concept of Foresight, to the early proof of concept, to where we are today with a functioning platform that people are paying for. It's been a lot of learning. I can't even imagine the amount of learning we've yet to undertake as well. The thing about a startup is that you're always learning and you never stand still — it's never just good enough. Maybe when I was working in a consultancy, if you met your annual revenue target and annual sales target, that was good enough. But with this, there's no end in sight. It's like, how do we grow this further? That's kind of the journey you start on whenever you take venture capital funding. There are certain expectations about where we will be in terms of growth and what problems we'll be solving, not just one year from now but looking further out.
Companies that promote themselves as AI companies have certain expectations around them about growth and valuation over time, and I suppose we do fit into that category. The key thing now is that this whole AI toolkit has unlocked things for companies like us that probably weren't available — or weren't commercially viable — before. The simple problem I was talking about earlier: taking unstructured market information, being able to structure that and map it to user products or operations — I don't think that would've been commercially viable before at the scale we're talking about here.
Chris: Yes. Yeah. And I guess having that as part of the DNA of your organization at least provides that substrate. Because it's so dynamic, isn't it — technology with AI. Who knows what's going to come along, but at least you're building it with that as the base, so you'll be ready to integrate and evolve what you're doing as the technology develops as well.
Steven Brennan: Yeah, absolutely. I know that some of the people you've had speaking on your podcast previously have used AI when it comes to risk assessments and that kind of thing. And I think it is still in the early stages in terms of what that actually means in those different applications. So it's really exciting to think about what that could look like one or two years from now, and how we're moving more into the reasoning side of things. Certainly from what we're trying to do at Foresight, teams can spend more time reasoning rather than just moving paper around — Excel files and all that kind of stuff — and people can feel more productive as a result.
Chris: Yeah, yeah. No, there's definitely a lot of gain to be made there. I wanted to ask you as well — just to show that you're continuing your technical work — I'm aware that you're a co-opted member of the Socioeconomic Analysis Committee (SEAC) at ECHA. Right. So maybe you could tell us a little bit about what you're doing there and what that's like.
Steven Brennan: Absolutely. ECHA in Helsinki has been the center of my universe for quite some time since I started my regulatory career. They were driving a lot of the work that we were doing whenever I was in industry. So it's been a real honor to be able to join the SEAC committee as a co-opted member. That benefits me — I get a trip to Helsinki every three months, catch up with the team there, broaden the network, get different perspectives from folks across member states and other organizations. For me, that's — I wouldn't say the dream come true, that might be a little bit too far, but...
Chris: Validation.
Steven Brennan: Yeah, absolutely. I suppose it kind of leans into the work and experience I have around substitution of hazardous chemistries in manufacturing processes — you're flexing a muscle I've built in my time within industry and as a consultant. And being able to do that is still an honor. It kind of balances the software development side nicely.
Chris: Yeah.
Steven Brennan: I still have a preference for the regulatory work.
Chris: Yeah. You've got that itch you need to scratch every so often. Does it get into the weeds with some of the technical work? Are there any things going on there that you particularly want to highlight about regulatory developments?
Steven Brennan: Not necessarily. I think ECHA is doing a good job of communicating the work that they are doing. We monitor lots of different agencies and the work that they're producing, and I think ECHA is the most communicative. It's really good to see. I think that the way good policy is made is through good communication.
Chris: Mm-hmm.
Steven Brennan: Anything going on there is publicly available anyway, in terms of what restrictions are in progress and authorizations and so on.
Chris: Yes.
Steven Brennan: I think they do a good job.
Chris: Yeah. There's a lot of information on the ECHA website — a wealth of guidance to keep everybody busy. But I would also say, recognizing that they're putting a lot of effort into the communication side of things — I suppose that's part of accountability, and speaks to the importance of the work that they're doing over there. It's worth mentioning that ECHA has its own podcast, Safer Chemicals, where the two chairs of the Risk Assessment Committee and the Socioeconomic Analysis Committee talk through the work that's ongoing there. So if anybody's interested in learning a bit more about the activities happening there, I recommend you check out that podcast. And I'll also mention that ECHA recently launched a science section on its website, which I've really appreciated because they have some seminars on there from over the last 12 to 18 months with leading experts in different areas. There's some really high-quality information in there about the science going on as well that I recommend people check out.
Right, Steven. Well, this has been absolutely fantastic. I did want to ask you as well — we talked a little bit about the life of working in a startup. How are you looking after yourself? What's your daily routine and how are you making sure you stay productive?
Steven Brennan: That's a good question. Well, I get up very early every morning — that's because I have a 14-month-old baby who likes to get up at 6 AM with no bother, every morning.
Chris: Oh, just to add to the plate.
Steven Brennan: Absolutely. And then it's about getting my 4-year-old ready for school — she just started school. Another daily battle, but a good one. For myself, it's about spending time with the family and looking after the family and supporting them as much as possible. Other things I like to do — although not so much in the last few weeks — is get out for a run a few times a week. And then my other hobby is fly fishing. I find fly fishing is a meditative kind of thing for me. It just clears the head, gives you single focus on one thing for a few hours whenever I get the chance to get out. So that, in combination with the work, I think gives me the three legs of the stool to keep things balanced.
Chris: Yeah, no, it sounds like you've got a good mix there. You're definitely pushing pretty hard, especially with a toddler and a 4-year-old. I'll say there is light at the end of the tunnel — that side should get easier. But fair play to recognize the importance of being around for your family as well, because if you're not careful you can miss these years. And taking the time to go fly fishing also sounds wonderful.
Steven Brennan: What about yourself, Chris? How do you keep the balance, given that you're kind of a startup consultancy too?
Chris: Oh, that's a really good question. For me, yeah, I mean, I think it's still a work in progress. I'm doing a lot of running, and I find that I need that to kind of help settle me down. If I've been doing a fair bit of work in the morning, often going out for a run is the perfect thing to clear your head. Luckily for me, this podcast is a passion of mine. It really has been — it was an experiment I started at the beginning of the year. I felt that I was kind of called to do it. It wasn't something I thought, oh, this is going to be great for marketing my business. Hopefully it is good for raising my profile, but more than that, the podcast is addressing something which I felt was missing in the discourse around chemicals. So these sorts of things — it's not all work. I actually enjoy it. It's sort of my hobby as well.
But yeah, I do a lot of reading. I try to do a bit of meditation, although that habit has slipped away. I try to stay attuned to how I'm doing mentally, picking up on when I need to take a break and go for a walk or something like that. And then of course weekends are generally ring-fenced as well for family time, and we always prioritize going for a long walk in nature.
Steven Brennan: Absolutely. I can imagine there are a lot of different things on your plate. How do you deal with switching between different things in a day? If you're doing a lot of work in the morning and working on different projects, for me there's a mental switching cost, and there's a debit associated with that switch. How does that level out for you, or do you have another way of dealing with that?
Chris: Well, I think the running definitely does help. I did read a lot of the productivity literature — the work of people like Cal Newport, and then Oliver Burkeman as well, who has done a lot on this. And I think the productivity world has kind of had its day, because people have realized that we were trying to squeeze every last drop out of the day, and that was actually a recipe for feeling guilty or unhappy, because we can't possibly be 100% productive all the time. So it's more about giving yourself a break, I think, and some reprieve when you haven't done what you set out to do at the start of the day.
I tend to load the to-do list with overly ambitious tasks every day. And people say you shouldn't task-switch, you should be focused on one task at a time. I try to do that, I try to be disciplined, but sometimes I need to go with what's motivating me at that moment. And I'll just try to keep pushing all the various things forward. That seems to keep me going in terms of making sure I'm looking after business development, business admin, and of course delivering for clients as well.
Steven Brennan: Absolutely. Those are real challenges. I did try to do my own consultancy for a short while, and that was something I noticed as well. Whenever you're working for a large organization — as you probably know from your previous work — having someone do the admin for you was good. And then having other people to lean on for delivery and on the business development side was good as well. But if you're having to do all of that yourself, it does take a special mix of skills. And being aware that task-switching increases the cognitive debit and reduces productivity — I think that's a good thing to be aware of.
Chris: Yes.
Steven Brennan: It doesn't go away, I don't think. There's no way to make it go away.
Chris: Yeah. Yeah. I think you feel the pain. But I don't regret for one second doing this — it's been super rewarding to go on this adventure.
Steven Brennan: I imagine. One of my old managers, Darren Green at Rolls-Royce, said to me: if you want something done, give it to a busy person.
Chris: Yeah. Yeah, it's probably true. All right, Steven, this has been absolutely fantastic. I normally ask two questions at the end, but I think it's pretty clear what your big goal is. Do you have any final advice for somebody starting out in their career?
Steven Brennan: In terms of regulatory career?
Chris: Somebody working in this kind of field.
Steven Brennan: In the regulatory career — I would say just be nosy. I think one of the things that's benefited me is just being nosy about things. Not just about policy and why policy is changing and finding out the roots of that, but in your organization, being nosy as to why certain things are done in a certain way. Is it always just done that way, or is there a good reason for it? And I think by learning all the ways things are done and why they're done, you can start to try and change them to make them better. I think we're all working in this space to try and make things better — whether it be how an organization manages regulatory obligations, or what policy should be around the use of chemicals. But once you learn all those rules and why things are done the way they're done, then you can start to bend them.
Chris: That's really good advice. Because quite often — well, I'm sure when we started out the advice was, keep your head down, do what you're told, and try to learn. And that is sound advice as well. But you do also need to have that openness, don't you — to not just trust that everything was great from the outside.
Steven Brennan: Yeah, absolutely. The chemicals themselves — the fact that regulators are nosy as to what these chemicals are and what they're doing, looking back at the REACH registration requirements — that's the same curiosity at work.
Chris: Yeah, no, I think that's fair to say. There's definitely a lot of investigation yet to be done. So no, this has been absolutely fantastic, Steven. Thanks very much for this, and I appreciate you turning the tables on me with some questions about my own side of things too. So yeah, thanks very much, and thanks to everyone who's been listening. Your time is really precious, and I'm really grateful that you spent it with us. If you've enjoyed this, please like, comment, subscribe, and spread the word about the Chemical Journeys Podcast. Thanks a lot, Steven.
Steven Brennan: Thanks very much, Chris.
Chris: Hi folks. I hope you enjoyed my conversation with Steven, and you'll join me in wishing him well with his journey with Foresight. If you'd like to find out more about Foresight, visit www.useforesight.io — that's www.useforesight.io — to find out more about how this solution could help you. See you next time.
