👋 Howdy to the 1,251 new legends who joined since our last edition! You are now part of a 569,987-strong tribe outperforming the competition together.

LATEST POSTS 📚

If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions.

🧑🏻‍🏫 Why Every CEO Needs Someone To Report To. An interview with Alexis d'Amecourt, CEO at Mochary Coaching.
🥣 Unfiltered: Granola Ate The Competition. A decade of AI note-takers, all beaten by the one that refused to join your call.
🪦 Substack Committed Creatorcide. A little part of the creator economy was murdered in cold blood.

PARTNERS 💫

Meetings aren’t the problem. The challenge is everything that comes after them. With Granola, the AI Notepad for people with back-to-back meetings, it's no longer about scrambling to keep up.

Take notes the way you always have. Granola works in the background, turning conversations into clear summaries, action items, and next steps. Before, during, or after a meeting, you can chat with your notes to quickly understand what needs to happen, write follow-ups, or share context with others.

Airwallex's agentic business accounts instantly provision accounts with local bank details to receive payments around the world.

Orchestrate transfers and payouts to manage liquidity. Earn returns on your idle cash. Issue virtual cards with pre-set limits and enforce spend controls automatically. Automate your financial operations across entities, across currencies, continuously, 24/7.

_{Interested in sponsoring these emails? See our partnership options here.}

HOUSEKEEPING 📨

Last week was a bit emotional for us when our little man Ziggy had to have a bunch of teeth extracted. He’s almost 10 now, and it marks a sad day when you realize your best buddy is moving into his senior years. He’s doing great, but still, it’s sad to think about.

A friend of mine recently raised some money for her startup Elita, which focuses on dog longevity. What an awesome thing to dedicate your life to. Now, let’s jump into today’s piece.

INTERVIEW 🎙️

Shelley Copsey, Founder & CEO at FYLD

Shelley Copsey is the Founder & CEO of FYLD, an AI platform created in 2019 that makes infrastructure field work safer and more productive. In February 2026, the company closed a $41M Series B led by Energy Impact Partners, with Partech joining through its Growth Impact Fund, after finishing the year at 82% growth. Customers now include Southern Water, Kier, Amey, Halliburton, and TC Energy, with some putting 30k to 40k people into the field every day. She has more than 20 years of experience working at the intersection of physical infrastructure and digital technology. Shelley began as a tax consultant at KPMG in Australia, then led innovation and digital-ventures work at PwC and commercialization at CSIRO's Data61. Before FYLD, she was CEO of 3D mapping company GeoSLAM and a board director at telehealth startup Coviu.

FYLD launched a platform for field crews in April 2020, at the exact moment nobody could visit a site, in an industry where workers couldn't stay home because gas, water, and power had to keep running. What survived is a company built on an unfashionable premise: the frontline worker is the real customer. FYLD holds an 8/10 satisfaction score with the people actually holding the phones, because Shelley believes a field platform dies the moment field workers reject it. She sold in through safety rather than productivity, on the logic that everyone already owns a productivity tool. Almost nobody had seen anything like this on safety, and her case studies now run to a 48% drop in injuries and incidents. The harder problem is the one in front of her. Coming off the Series B, she's trying to price an AI product on consumption and outcomes in a sector where neither vendors nor buyers have worked out yet what those deals should look like.

What problem is FYLD solving for infrastructure teams today?

Our customers send anywhere from two or three thousand right up to thirty thousand people a day into the field, and the typical problems with field service are pretty consistent. Somewhere between 35 and 50% of the budget in the organizations we work with goes on human labor. Then those people head out, and nobody really knows what actually happens in the field. Nobody knows which jobs are going well and which ones are struggling. Nobody knows which field workers are operating in a high-risk environment. There's almost zero visibility from the moment they hit the field. So project costs blow out, you get quality or schedule problems, and people get hurt.

FYLD brings real-time visibility into what happens once those workers go out. We crowdsource the data. We get field workers to take short videos and tell us what they're up to, and then we use AI to analyze what's going on across all of those two thousand to thirty thousand people. A remote manager might run anywhere from five to twenty crews. We help them figure out which jobs in their patch they should focus on, and how to shift from reactive decision-making to being proactive, so they drive every day's operations through to safe completion.

What's a real-world example of FYLD making the work safer for a customer?

Think about the people running electricity down the wires to our houses, or the crews working on the water pipes. A common situation is that they head out into the field and use heavy machinery. Say they need to do some digging to get down to a pipe. We see two hazards there on repeat.

The first is overhead wires. These workers are no different from people who sit in offices. You get what we call industrial blindness, where you stop seeing the things that are there every single day. They frequently forget to look up, and then the machinery clips the overhead wires. That's a very big safety risk on site. The second is the deep excavation itself, and whether there are enough boards and barriers around it to stop the crew from falling in. Again, it feels like something that should be obvious. But think about how busy these sites are, with cars driving past at high speed.

It's easy to assume they're experienced, so surely they should be safer. The onus should really be on us to make sure they actually are. They get distracted. They stop seeing things. So with FYLD, we get them to tell us about the risk environment on site. There's a fundamental difference between ticking boxes on a piece of paper and actually vocalizing what's around you, and we know we can drive a behavioral change in safety that way. Then if you think about computer vision, we can spot the things they don't. The platform will ask whether they've noticed the overhead cables and whether they've put anything in place to protect themselves against them. That's FYLD at its most rudimentary, but the impact on those field workers is very substantial.

And what's a good example of somebody being able to move faster?

This won't surprise anybody, but we track around 30% standing time on site. When you drive past a job like that, you frequently see people standing around. I think the general public has a habit of blaming the field worker, but when it's 30 to 35% of the time, you have to think a bit harder than it being his fault for wanting a cigarette or a sausage roll.

What we actually see is that the wrong skills or the wrong kit got sent, or any number of other process problems. These organizations run on very tight budgets. The regulator gives them five years to complete a plan, and they've got a cost cap to work within. So people know in their gut that some of these problems exist. It's just very hard to act on them systemically, because you don't have the data to understand them in any depth.

From a productivity point of view, one of the things we do is work out in real time where standing time is happening. We can tell either because people tell us it's happening, or because we see abnormalities in the data. So first, in the moment, let's get them back to work. And second, let's help our customers understand at scale where the problems in their operational processes sit, so they can fix them through a data-led angle.

How does FYLD actually see what's going on and get people back to work?

Let's go to the remote command center. Say I'm Shelley, the remote manager, and I've got twenty crews. Before FYLD, my phone was ringing; I'm driving site to site, visiting people, getting WhatsApps. Managers like that are inundated with unstructured information. Instead, we have the field workers film what's actually happening on site, and we triage from there. We've got a very large proprietary data set, so as each new piece of data comes in, we can start checking it against that. Does this job look normal? Is it progressing to plan? Two hours ago, did we predict they might need a permit to continue, and was one actually put in place?

You go from sampling to letting the data point you at the problem. So as that remote manager, I stop chasing whoever is shouting loudest and go to the jobs where the data says something serious is happening. That's typically what lets you unblock a job and get people back to work.

What was the hardest part of getting FYLD off the ground?

They were getting sick constantly because they were out in the world with no immunity to it. So actually finding customers who would talk to us and try the platform was hard. You need very deep user understanding when you're working with field workers. Most platforms like FYLD fail because the field workers reject them. That real hands-on time was the first thing.

The second problem has the same root. You can't turn off the energy network, so these organizations have to be extremely deliberate about introducing new technology. They need to be sure it won't affect the communities they serve. That makes for a lot of hesitation about moving at all. It has eased a lot in the last three years, because AI is common enough in everyday life now that our customers are adopting it much faster. Five years ago, that wasn't the case at all.

How did you land your first big customer?

It was interesting. One thing we realized early on was that the go-to-market had to focus on wherever we could find an insertion point for our product, which is very broad. Every one of these companies has an ERP and an asset management system. Nobody has an operating system for the field, and that's what we're ultimately building. When you're a young, tiny company walking into these behemoths talking about an operating system for anything, it's fairly easy to get laughed out of the room. You don't have proof points. You've got a big vision. So we had to dumb it down to the smallest part of the platform where the utility would be immediately obvious, and where there was a huge problem in the market. We went in through the safety lens.

Our platform has incredible productivity uplifts, but everybody already has a productivity tool. I think most of them don't work, but that doesn't matter. The box is ticked, and that's very hard to go head-to-head with. Safety was different. Nobody had ever seen anything like what we were doing, and it's an industry that still hurts people at a rate greater than most people realize. So once you start showing a new way of working, things move.

We were very lucky to be built in conjunction with SGN, one of the UK's biggest gas networks. We had a case study showing a 20% reduction in incidents and injuries in the first year. Once we really focused on that insertion point, we found our way into customers. And then we could run the typical startup play from there, land and expand, showing the new features and functionality and pushing up the utility and ROI.

What other data points can you share on the impact FYLD is having?

Safety is always our top priority. In the last twelve months, we've built several case studies showing up to a 48% reduction in incidents and injuries in the field. That's huge. That's a lot of people walking out their front door knowing they're coming home for dinner with their husband or wife and their children. That one we're massively proud of.

Our customers also operate on thin margins, or inside regulated cost constraints, so the productivity piece is massive. Time to value matters enormously to us, and right now it typically sits between four and six weeks. On productivity uplift, if you think of it as capacity increase, we're getting as high as 8 to 12% over that same four to six weeks. What we love there is that our customers aren't deploying transformation projects. You're just giving them a natural new way of working that's intuitive. They're not sitting there thinking they have to invest a hundred million pounds, Salesforce style. Very low barrier to entry, very high outcomes.

For a use case-specific outcome, take the water sector here in the UK. Think about the sheer amount of water that leaks out of the networks. A typical leak runs for five days before it's fixed. We're not in leak detection. We're in what happens once you've found one, which is showing customers how to use their data to get the repair done better and faster. We've dropped those five days to three and a half, so a 30% reduction. The impact is really very substantial.

What does your day-to-day look like as CEO?

First and foremost, I'm not a technical founder. So you'll find me leaning very much into the commercial parts of the business and the customer piece. My starting hypothesis is that happy field workers give you more, better data. So the thing you probably see me obsessing over most is what our daily active usage looks like. When we're deploying, are those activations going well? Are we getting positive CSAT, and if not, what's coming through? We're running an 8 out of 10, which is great, but keeping it there is no mean feat.

Then, in field service, your users are at the front line, but your economic buyer is usually sitting in an executive seat. Those executives aren't obsessing over the future of data-led field operations the way I am, because their days are very, very busy. So the other thing I focus on is putting that future in front of them and validating it. You can tell quickly whether they think you're completely off the mark, or whether they hadn't considered it but there's something in it. That validation of the big picture matters a lot to the roadmap.

On a daily cadence, the thing I always come back to is that I can do my job in the evening. What I can't do in the evening is keep the team moving at pace, so removing friction points is critical. I never want a world where people think decisions have to come up to me, and we actively push against that culture. But sometimes you can see someone ruminating on a reversible decision for too long, and it's within my power to nudge them along.

The Series B earlier this year fundamentally changed the nature of the organization. All of a sudden, you've got very different stakeholders thinking about what the future looks like. I'm spending more time on future strategy and the path to a monetization event, and working out what that path looks like has been quite a change-up.

Do you have any frameworks for goal setting?

We're reasonably top-down in how we set goals, and I focus on an OKR framework. What I love about OKRs is that anybody on the team would tell you I'm the furthest thing from a micromanager you'll ever meet. It bores me, and we hire great people. With OKRs, they know where I want them to get to, and they have swim lanes to get there, but they can do their own thing, and I can just bump them back in if I see they're drifting off-kilter. So we do a very thorough OKR-setting at the company level.

We've done a lot of three-year deals in the past, which is big for a company of our age when you're doing them with major enterprise. But the market is shifting. People are getting nervous about vendor lock-in and all sorts of things. So you still need your North Star, the growth rates you're after, your active users, the types of logos you want. You just have to be very flexible about the path to get there, what the business model looks like, and how you iterate on it. That's why we sit down every quarter and get clear on all of it. If something has become irrelevant and needs throwing out, that's always been fine at FYLD. We don't keep chasing goals that have gone out of date.

Who's on your team, and what's your philosophy on managing them?

My direct reports are a Chief Revenue Officer, a Chief Product & Technology Officer, a Finance Director, a Head of Strategic Enablement, a VP of Customer Success, and a VP of Growth & Strategy we've just appointed. The whole marketing team also reports to me, because as a founder, the day you start putting someone between you and that team, you've made a fatal mistake. So that's eight or nine direct reports.

I’m so not a fan of a standing thirty-minute meeting with everybody on a Monday. I'm always available on Slack and WhatsApp, and if something is transactional, let's just keep the business moving. When something pops up that needs talking through, let's get it done today in fifteen minutes rather than waiting until next Monday. Either I or someone on the team will put thirty minutes in here and there for a more general catch-up, but I'm definitely not a rigid schedule setter.

What I do like is a Monday sales meeting where everybody gets visibility. There are functions where you need to be able to say what will look different by next week, so the whole sales team sits down with the CRO. I meet the marketing team every Monday for the same reason, working out the tactical things we'll do this week off the back of what we learned. It's a bit horses for courses. But mostly I lean towards talking when we need to and keeping the pace up.

What's your go-to market for securing logos everybody would be jealous of?

It's really evolved over the last twelve to eighteen months. Back then it was almost entirely founder-led and network-based. Then we tried a couple of things like hiring a straight sales professional—someone with a really good black book of contacts. We've found time and time again that people with good contacts in our industry, who have sold into it on repeat, do better than cold starters, but a black book doesn't last forever.

We're also very focused on building up the partnership channel. We've got an amazing partner over in Canada in Bravo Target Safety. They're so well recognized in turnarounds for companies with safety challenges that they can walk in and talk to people about us. We've got another three or four partnerships really getting moving, including a technology partnership that should drive its first income next quarter. And we forecast that by the back end of next year, that channel will be a very significant ARR driver, more so than direct sales alone.

What's your biggest problem to solve coming out of the Series B?

With some of our customers, we're across 20% of their business and doing really, really well, but these are big customers. So how do we build a rinse-and-repeat mechanism to get from 20% to 100% of their business? And I say that in light of changing business models, of the shift to consumption and outcome-based pricing. It's quite a difficult problem to maneuver. Two years ago, the math was simple. One user costs X; a hundred users cost some percentage less than X per head. We all knew those discount parameters to get to scale. I don't think we're doing those deals anymore. Now the questions are whether the current fee becomes a platform fee, whether we shift to usage and outcome, and how we account for the different parts of a customer's business that use the product. Getting that right is going to be absolutely central.

Meanwhile, our customers are thinking differently about their labor force. They're starting to see that data will let every person be more productive, and maybe more cross-functional. So two things are coinciding: their rethink on people and ours on consumption and outcome. And nobody really knows how to do these deals in our industry, not us and not our customers. On top of that, you need to drive very high ARR growth, which our investors and we require, as it should be. So how do you keep that pace of deal-making going in a complex and unknown environment?

How do you test pricing in that complex and unknown environment?

You need a reasonable level of transparency with your customers about your known knowns and known unknowns. One nice thing for us is that we've got five or ten customers who have been with us for five years. They're great customers, and we trust them to tell it like it is. So both sides are willing to put a price reset mechanism in at twelve months if the deal isn't working for everybody. But as the vendor, you go in knowing that if it's working for them and not for us, we'll probably swallow the bitter pill for another three years before we can reprice. And they need to know we're not taking them for a ride either.

The other piece is smaller expansions inside the customers we already have. Rather than aiming to go from one business line to five, we could take on half of another business line for six or nine months and run it together, so everyone builds some knowledge. Then it's deal by deal. I'm hunting for the friction points, asking what's actually bothering this customer when they're buying in an environment nobody understands yet. And then, within our own risk tolerances, how do we take that friction away so the deals keep moving? You have to be more agile now than ever.

On the product side, how do you train your data sets?

A lot of human-in-the-loop, given the types of environments we work in. There’s nothing black box at FYLD. We're always looking at outcomes and validating them, and we get customers to validate too, which I think is super important. It's the same question with the tools we use. Anything third-party that we can't see inside, we stay well away from. In product, you need real transparency with your customers about what you're running.

Customers have deep concerns, of course, because we train on their data. It's a very big proprietary asset. So a lot is going on around how we anonymize it and how we store it. We've also invested heavily in extremely capable engineers. We're working with critical national infrastructure, so we take a lot of care to ensure our data engineers are careful and respectful and get the law right, from GDPR to whatever else we come across.

How are you thinking about AI internally?

Twelve months ago we were stuck in a whole stack of tools. We were very, very careful about data security and who the vendors were, as we've always been. But once something passed security, anyone could try it. The thinking was that if people are interested, they'll have fun with it, and they'll build confidence and skills along the way. Then you have to shift rapidly into how we drive results. We're all watching the cost of tokens hit the P&L, and neither I nor our CFO can responsibly avoid asking what outcomes we're getting for it. So this year we've narrowed the tools people can use down to the ones we can see having an impact and that actually automate workflows. That has been key.

We also look at who is and isn't using them. We've got great people, so if someone isn't, it's usually because they sit in a commercial rather than an engineering role, and they just don't have the confidence yet. We've done a lot of partnering from product and engineering into the commercial team. Help someone automate their first workflow, and they'll find the confidence for the second. The next change is probably bringing in an internal AI expert. We run a really efficient headcount of around fifty people, but I want to push up the output-to-headcount ratio.

The other thing I always come back to is that we're five years old. The product is heavy AI, but our internal operations aren't AI native. If we raise again, or exit, we're competing with companies built on an AI native cost base. That transition has to be a real focus with the existing team, and it's now absolutely core to how we interview. We want to hear the actual things someone has put into production, and whether they've driven new revenue or driven down a cost curve. We're not bringing in capability that doesn't already work that way.

You've probably saved a lot of lives. Do you ever talk about it that way?

It is a real thing. I love that people are walking out the front door knowing they're coming home safe. My grandfather was from the generation in Australia that left school at eleven or twelve to shovel coal into the electricity generators. As a kid, he'd tell me stories about what went on out there. When I came into this role, I remember thinking that at least I had a grandfather with four limbs.

I frame it this way: these people are entitled to the outcomes we bring them. I'm proud we deliver them, but I think we should all ask why construction is still dangerous in the first place. It should be table stakes.

How do you get the best out of yourself?

I love working with the team and letting them do the best work of their career. But I also know that means being ruthless about hiring, about probation, and about exiting people. The worst of me comes out when I start seeing non-performers or B-grade people. They're great for some companies, but in a company like ours, you need your solid rocks to be A-grade solid rocks, and you need people shooting for the stars with you. Either is fine.

There's maybe one point I haven't covered. Field workers are invisible to head office. You send them out, and that's it. But go and talk to them, and they'll tell you the COO dropped virtually into their job and told them the site looked great. They'll tell you they used to feel invisible, and now the platform lets someone see what they actually do day to day. I need to immerse myself in that. Then the tougher parts of the job get done, because I've got that behind me.

Extra reading

• FYLD: Interview With Founder And CEO Shelley Copsey About AI-Powered Infrastructure Field Work - September, 2026

• Meet the innovator making construction sites run smoother - July, 2026

• How AI Is Preventing Field Failures - June, 2026

And that’s it! You can also follow Shelley on LinkedIn and check out FYLD on their website to keep up with what they’re building.

BRAIN FOOD 🧠 

TWEETS OF THE WEEK 🐣 

TOOLS WE RECOMMEND 🛠️

Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. Today, we are highlighting Framer*—the site builder trusted by startups to Fortune 500.

See the full set of tools we use inside Athyna & Open Source CEO here.

HOW I CAN HELP 🥳

P.S. Want to work together?

1. Hiring global talent: If you’re hiring tech, business, or ops talent and want to do it 80% less, check out my startup, Athyna. 🌏

2. See my tech stack: Find our suite of tools & resources for both this newsletter and Athyna here. 🧰 

3. Reach an audience of tech leaders: Advertise with us if you want to get in front of founders, investors, and leaders in tech. 👀