<p dir="ltr">The Environment Agency has introduced a new citizen science initiative at Par Beach in Cornwall and Combe Martin in Devon, encouraging beachgoers to engage with AI to monitor water quality and beach conditions.&nbsp;</p>
<p dir="ltr">This 12-week trial, starting in October, aims to involve the public in gathering valuable data while keeping them informed about the state of their local environment.</p>
<p dir="ltr">At both beaches, visitors will find QR code signage prompting them to start a conversation with the Hello Lamp Post AI platform. This system will share the latest information from the Environment Agency&rsquo;s water quality monitoring efforts, but the interaction is designed to be more than just a one-way update. The agency hopes that beachgoers will contribute their own observations by answering questions about the number of people or dogs on the beach, the clarity of the water, or any signs of pollution.</p>
<p dir="ltr">The project has gained strong local support. Barbara Martin of Combe Martin Water Watch praised the agency&rsquo;s continued focus on water quality, stating, "We are proud of the 'Good' water quality rating for Combe Martin, and this AI scheme will empower both residents and visitors to report concerns directly and instantly." Similarly, Jenny Tagney from Friends of Par Beach, which participates in water quality monitoring projects, expressed excitement about the initiative. "We are really pleased to support the Environment Agency in its efforts to understand the key factors affecting Par Beach's water quality."</p>
<p dir="ltr">This AI-driven citizen science trial is expected to complement the Environment Agency&rsquo;s seasonal sampling efforts, which concluded on September 30. By maintaining an ongoing flow of data beyond the regular monitoring season, the project will help create a fuller picture of the environmental conditions at both beaches, involving the community directly in safeguarding their coastal areas.</p>
<p dir="ltr">This initiative signifies the integration of AI into environmental monitoring - one of many industries that will increasingly rely on AI expertise. To explore a range of diverse sectors where your skills can make an impact, <a href="https://www.generativegroup.ai/contact-us/">contact us today.</a><strong id="docs-internal-guid-d4c4296c-7fff-da99-24c6-c13cc426c710"></strong></p>

The Environment Agency has introduced a new citizen science initiative at Par Beach in Cornwall and Combe Martin in Devon, encouraging beachgoers to engage with AI to monitor water quality and beach conditions. 

AI Signs Set to Improve Beach Conditions in the South West of England

<h4 dir="ltr">Data centres are under immense pressure, from power constraints to sustainability mandates. Cooling systems are one of their biggest energy drains &ndash; and OctaiPipe is using AI to make them significantly more efficient.</h4>
<p dir="ltr">But this isn&rsquo;t your typical AI startup story. Born from a spur-of-the-moment conversation on a train and shaped by years of deep R&amp;D, OctaiPipe has evolved from consultancy roots into a product-first company focused on federated learning &mdash; an approach that keeps data local and private, while still enabling powerful AI training at the edge.</p>
<p dir="ltr">We caught up with co-founders George Hancock, Ivan Scattergood and Eric Topham to talk about privacy-preserving AI, federated learning, and building for the market before the market&rsquo;s ready.</p>
<h2 dir="ltr">What specific problem is OctaiPipe solving, and why does it matter now?</h2>
<p dir="ltr"><strong>George:</strong> As AI demand grows, so does the pressure on the infrastructure that supports it. Data centres are facing increasing constraints around power, energy use, and sustainability &ndash; and cooling systems are one of their biggest energy drains. That&rsquo;s where we come in.&nbsp;</p>
<p dir="ltr">Our focus is to optimise energy efficiency in data centre cooling systems using AI, aiming to deliver up to 30% savings.</p>
<p dir="ltr">What makes our approach different is the use of federated learning. Our platform allows AI to be trained and deployed securely at the edge, maintaining privacy. This design has attracted attention from both customers and investors, which has naturally drawn us deeper into the data centre space. We offer a timely solution to a growing challenge.</p>
<h2 dir="ltr">Tell us about the idea behind your startup. What&rsquo;s the story there?</h2>
<p dir="ltr"><strong>Ivan:</strong> It all started on a train, really. Eric met Paul Carver while geeking out over weather maps &ndash; Eric was racing professionally at the time, and Paul was about to set sail. One thing led to another, and Paul, excited by Eric&rsquo;s background in data science, suggested they start a company.</p>
<p dir="ltr">That company turned into a successful data science consultancy. George came on board to lead commercial efforts, and I joined to help deliver projects. As part of our work, we ran an &ldquo;innovation sandbox&rdquo; where MSc students could complete their theses by working on real R&amp;D challenges for our clients.</p>
<p dir="ltr">It was during one of these projects that we began exploring federated learning &ndash; driven by a use case involving factory-floor devices producing a terabyte of data per week each. Uploading that volume to the cloud was tough, so instead of taking the data to the code, we brought the code to the data. That&rsquo;s when the concept behind OctaiPipe, our federated learning platform, was born.</p>
<p dir="ltr">The market wasn&rsquo;t quite ready for a federated learning platform on its own. What it was ready for was a compelling application of that tech. Our work with edge devices in manufacturing pointed us toward data centres.&nbsp;</p>
<p dir="ltr">Several clients nudged us in that direction, and we realised we could simulate data centre energy use and train reinforcement learning agents to optimise cooling. That&rsquo;s how OctaiPipe became what it is today: a platform quietly powered by federated learning, driving real-world energy efficiency.</p>
<h2 dir="ltr"><img src="https://cdn.sourceflow.co.uk/__gallery/1ad4899e-4079-46fd-b027-5312fccfc353" alt="" width="600" height="398"></h2>
<h2 dir="ltr">There&rsquo;s a lot of uncertainty around AI in the wider population. What are some of the common myths or misconceptions you hear?</h2>
<p dir="ltr"><strong>George:</strong> We often hear hesitancy to the tune of "I don&rsquo;t know it, I don&rsquo;t trust it, I don&rsquo;t want to use it"&nbsp; &ndash; especially in manufacturing, where people have real concerns about how it fits into their work.&nbsp;</p>
<p dir="ltr">There&rsquo;s also this idea that AI is something you just plug in and it works. But without upskilling and buy-in from the people using it, even the best solutions won&rsquo;t stick. Our approach with clients is always: do you actually need an AI solution? If we can&rsquo;t see a clear business case or ROI, we won&rsquo;t build it. We&rsquo;ve walked away from projects that didn&rsquo;t stack up, and some of those clients have come back later with a more focused brief.</p>
<p dir="ltr">And honestly, AI still exists in a bit of a bubble. To people outside the space, it often just means "ChatGPT" now. But there&rsquo;s a whole world beyond that. Ivan, for example, was working on ML way before it was cool.&nbsp;</p>
<p dir="ltr"><strong>Ivan:</strong> I studied cybernetics and control engineering back in 1994 and built a neural network to play the Prisoner&rsquo;s Dilemma. The maths has been around for decades. What changed was the arrival of big data and greater computing power about ten years ago. That's why there's been this massive explosion.</p>
<p dir="ltr">Back when we were a consultancy, we used to avoid saying "AI" at all. We&rsquo;d say ML, because that&rsquo;s what we were actually doing. We even had a running joke: if you&rsquo;re talking to investors, it&rsquo;s AI. If you&rsquo;re hiring, it&rsquo;s ML. And when you&rsquo;re actually building the thing, it&rsquo;s logistic regression.&nbsp;</p>
<p dir="ltr">There's a lot of hype about it but, in most situations, you don't need anything particularly complicated to help you. You just need something that works.</p>
<h2 dir="ltr">&lsquo;AI safety&rsquo; can mean very different things depending on who you talk to. What does AI safety mean to you, and how does it show up in your day-to-day?</h2>
<p dir="ltr"><strong>George:</strong> AI safety is a broad term, and its meaning really depends on who you&rsquo;re talking to &ndash; practitioners, users, commercial teams, or policymakers will all define it differently. For us, it&rsquo;s about building AI systems that are explainable and trustworthy.</p>
<p dir="ltr">Last year, we worked on an Innovate UK initiative to unpack this idea. We looked at how terms like &ldquo;AI safety&rdquo; and &ldquo;AI ethics&rdquo; are often used interchangeably, and we built a comprehensive framework around the principles that matter most to us &ndash; things like transparency, interpretability, and user confidence.</p>
<p dir="ltr">That thinking carried through into a project with P&amp;G, where we focused on practical ways to make AI outputs understandable and reliable. Not just technically sound, but usable and trusted by the people interacting with them.&nbsp;</p>
<h2 dir="ltr">What&rsquo;s been the biggest challenge in bringing an AI product like yours to market at scale?</h2>
<p dir="ltr"><strong>George: </strong>Market maturity. We&rsquo;ve always been slightly ahead, working on solutions 18 months before the market&rsquo;s really ready for them. That edge has come from staying close to research and applying data science deeply, but it&rsquo;s also meant we&rsquo;ve had to shift focus to more immediate, tangible applications.&nbsp;</p>
<p dir="ltr">The challenge isn&rsquo;t whether AI can help. It&rsquo;s whether the market recognises the problem it&rsquo;s solving, right now.</p>
<p dir="ltr"><strong>Eric: </strong>I&rsquo;ll put it another way: 99% of people talking about AI don&rsquo;t really know what they&rsquo;re talking about. That makes scaling difficult, because most organisations don&rsquo;t know what they&rsquo;re buying or how to be ready for it.</p>
<p dir="ltr">It&rsquo;s not just about technical integration. It&rsquo;s about embedding AI into business processes. You can build a brilliant churn prediction model, but if the business doesn&rsquo;t know what to do with that prediction, it&rsquo;s useless. And too often, rather than saying &ldquo;we weren&rsquo;t ready,&rdquo; people blame the application.</p>
<p dir="ltr">So, yes, it&rsquo;s a maturity problem. Even now, as awareness grows, most people still struggle to frame the use case, or measure its impact in commercial terms.</p>
<p dir="ltr"><img src="https://cdn.sourceflow.co.uk/__gallery/7edcc2f1-801c-4b66-bcaa-0f4846528d41" alt="" width="600" height="312"></p>
<h2 dir="ltr">Is it possible to have the traditional 3&ndash;5 year business plan in a sector like AI?</h2>
<p dir="ltr"><strong>Eric: </strong>I&rsquo;d say the idea that &ldquo;everything changes too quickly for long-term planning&rdquo; mainly applies to the world of LLMs, where there&rsquo;s a well-funded arms race driving rapid iteration.</p>
<p dir="ltr">But outside of that bubble, there are slower-moving, longer-term trends &ndash; like edge computing and the ability to run machine learning on small, distributed devices. Those shifts take time because they rely on complex systems maturing and integrating together.</p>
<p dir="ltr">When George says we&rsquo;ve been operating 18 months ahead of the market, I&rsquo;d argue it&rsquo;s often further. We&rsquo;ve been building toward a distributed machine learning paradigm for years: models that don&rsquo;t rely on massive, centralised datasets in the cloud, but instead train locally on devices like mobile phones. It&rsquo;s a direction that supports data privacy, sovereignty, and trust. These are all issues that matter increasingly to individuals, businesses, and governments alike.</p>
<p dir="ltr">So yes, some areas of AI are evolving quickly.&nbsp; LLMs release every few weeks, but even those models are borrowing ideas like quantisation from edge AI, where constraints like limited memory have been long-standing design challenges.</p>
<p dir="ltr">What we&rsquo;re seeing is the growing viability of scalable, distributed AI &ndash; something we&rsquo;ve been ahead of the curve on for some time. And in that space, yes, you can plan for the long term because you're building on maturing trends, not chasing hype.</p>
<h2>You&rsquo;re now a team of 15. What has the talent attraction process been like, and how has it evolved?</h2>
<p dir="ltr"><strong>Eric: </strong>We&rsquo;ve shifted a lot as a company. In the early days, we were more service-based &ndash; building bespoke solutions for others &ndash; so we hired a lot of sharp, junior talent straight out of top universities. The recruitment focus then was all about raw technical ability.</p>
<p dir="ltr">Now, as we&rsquo;ve transitioned into a deep-tech, product-led company, we look for experienced specialists with domain expertise. The process has become more about understanding how someone approaches design and problem-solving. Can they navigate complexity, not just code?</p>
<p dir="ltr">If someone shows up and can&rsquo;t code, we&rsquo;ll figure that out pretty quickly and they won&rsquo;t be around for long. So it&rsquo;s less about testing fundamentals and more about bringing in people who already know how to apply them at a high level.</p>
<h2 dir="ltr">Finally, you're obviously very busy people. What can we find you doing in your downtime?</h2>
<p dir="ltr"><strong>Eric: </strong>Thinking about work.&nbsp;</p>
<p dir="ltr"><strong>George: </strong>Yeah, ringing each other to chat about what we need to do next!&nbsp;</p>
<p dir="ltr"><strong>Eric:</strong> But when I do switch off, I like exploring Paris with my wife or going climbing.</p>
<p dir="ltr"><strong>Ivan:</strong> I train in kung fu two or three times a week &ndash; I&rsquo;ve been training with my current Shifu for eight years. Before that, it was Taekwondo, Wing Chun, boxing&hellip; martial arts have been a big part of my life since my twenties. I do a fair bit of mountain biking as well.</p>
<p dir="ltr"><strong>George:</strong> I used to be big on gym, swim, and exercise. But since my daughter arrived last July, most of my time off is spent on daddy daycare duty to give my wife a well-earned break!</p>
<hr>
<p dir="ltr"><img src="https://cdn.sourceflow.co.uk/__gallery/1993f6aa-ebed-4306-a563-2ddf177443ce" alt="" width="600" height="322"></p>
<p class="article-editor-paragraph"><strong>OctaiPipe is one of the 54 innovative AI organisations featured in our 'Faces of AI 2025' report, which is available to </strong><a class="article-editor-link zSvKgyodXUoGbhUhxSWpxUXaFuhWtTKexrM " href="https://www.generativegroup.ai/reports/faces-of-ai-2025/?utm_source=linkedin&amp;utm_medium=social" rel="noopener noreferrer"><strong>download for free now</strong></a><strong>.</strong></p>
<p class="article-editor-paragraph article-editor-content__has-focus"><strong>At Generative, we work with AI-driven startups to scale innovation, refine strategy, and accelerate growth. If you're building in AI and looking for the right talent to take your company to the next level, </strong><a class="article-editor-link zSvKgyodXUoGbhUhxSWpxUXaFuhWtTKexrM " href="https://www.generativegroup.ai/contact-us/" rel="noopener noreferrer"><strong>let's talk.</strong></a></p>

Cooling systems for data centres consume huge amounts of energy. OctaiPipe is taking on the challenge with AI-powered federated learning

Cool Heads, Cooler Systems: The Startup Making AI Work for Energy Efficiency

<h3 id="ember219" class="ember-view reader-text-block__paragraph"><a id="ember220" class="ember-view" href="https://www.linkedin.com/in/jonathanrhyslow/">Jonathan Low</a> <strong>and </strong><a id="ember221" class="ember-view" href="https://www.linkedin.com/in/jakobcassiman/">Jakob Cassiman</a><strong>, co-founders of </strong><a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.linkedin.com/company/scopeinspection/" data-test-app-aware-link="">Scope</a><strong>, were outsiders to the Testing, Inspection, and Certification (TIC) industry &ndash; and that&rsquo;s exactly what gives them their edge. </strong></h3>
<p id="ember222" class="ember-view reader-text-block__paragraph">They&rsquo;re redesigning the inspection process for industries still reliant on pen, paper, and outdated tech. Scope builds AI-powered tools that cut cycle times by up to 45%, freeing inspectors to focus on decision-making over admin.</p>
<p id="ember223" class="ember-view reader-text-block__paragraph">They&rsquo;ve only just begun, but the potential to disrupt the sector is huge. We caught up with Jonathan to hear how Scope went from an idea to a fast-growing company working with some of the biggest names in TIC.</p>
<h3 id="ember224" class="ember-view reader-text-block__heading-3">Could you be a key player in Scope&rsquo;s founding team?</h3>
<p id="ember225" class="ember-view reader-text-block__paragraph">Generative has partnered exclusively with Scope to hire for key London-based roles in AI and engineering. If you're ready to help reshape the $250bn inspection industry, get in touch with <a id="ember226" class="ember-view" href="https://www.linkedin.com/in/dan-clouting/">Dan Clouting</a>.</p>
<h2 id="ember227" class="ember-view reader-text-block__heading-2">Let&rsquo;s talk about the origins of Scope. How did you and Jakob come together?</h2>
<p id="ember228" class="ember-view reader-text-block__paragraph">Both of us are really ambitious people from technical backgrounds. I used to work at Conjecture, focusing on document processing and building tools to make AI systems more controllable. And Jakob was a machine learning engineer working with physical data.</p>
<p id="ember229" class="ember-view reader-text-block__paragraph">We both knew we wanted to start a company, so we left our jobs and joined a start-up program called Entrepreneur First in London; they introduce you to other aspiring founders and you work on exploring problems and developing products together.</p>
<p id="ember230" class="ember-view reader-text-block__paragraph">We started building AI tools to search documents like PDFs, Excels and other structured data formats, before leaning into the TIC market.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/dbd68f25-0a63-497f-b551-34e08b787274" alt="" width="600" height="568"></p>
<h2 id="ember232" class="ember-view reader-text-block__heading-2">How did Scope end up getting into the TIC industry?</h2>
<p id="ember233" class="ember-view reader-text-block__paragraph">One of our early customers was a large inspection company and the positive reaction we got after working with them was <em>incredible.</em></p>
<p id="ember234" class="ember-view reader-text-block__paragraph">We spoke to an inspector on their team who carried a hefty manual of standards she had to reference during every inspection. After trying the tool for the first time, she said, &ldquo;Wow, I don't need this anymore&rdquo; and threw the book away.</p>
<p id="ember235" class="ember-view reader-text-block__paragraph">It was a small moment that spoke volumes.</p>
<p id="ember236" class="ember-view reader-text-block__paragraph">After this experience, we started talking to more people in the TIC space, and realised it's an industry that's been stuck with dated systems for far too long. We saw an opportunity to create value and solve real problems &ndash; so we decided to double down on TIC.</p>
<h2 id="ember238" class="ember-view reader-text-block__heading-2">What&rsquo;s the biggest problem that Scope helps inspectors solve?</h2>
<p id="ember239" class="ember-view reader-text-block__paragraph">Inspectors spend most of their time on planning, creating reports, and data retrieval for their next inspection.</p>
<p id="ember240" class="ember-view reader-text-block__paragraph">We build tools that automate the time-intensive parts of the process, enabling inspectors to spend more time on decision making. This pays off for our customers: we reduce their cycle time by 45% and get average reporting times down to 12 minutes. Without compromising on the quality of the work.</p>
<p id="ember241" class="ember-view reader-text-block__paragraph">We talk to companies with 100,000 people, who'll admit, &ldquo;Yeah, we still do everything on pen and paper&rdquo;. The reality is, the software that inspectors are using is built for other people. It's built for people in factories, it's built for maintenance experts. Inspectors have been incredibly underserved by tech solutions.</p>
<h2 id="ember243" class="ember-view reader-text-block__heading-2">What's the most exciting technical challenge that you're dealing with at the moment?</h2>
<p id="ember244" class="ember-view reader-text-block__paragraph">Something that&rsquo;s difficult in the inspection space is dealing with a lot of different workflows across customers. Inspecting a t-shirt on a production line is very different from inspecting a power plant. So, we build software that can work for different types of customers. That's a challenge that's been pretty interesting for us to adapt to.</p>
<p id="ember245" class="ember-view reader-text-block__paragraph">We make bespoke systems for each customer. We've got machine learning engineers who ingest everything from PDF data to Excel data; we use that to automate their workflows. That&rsquo;s why the companies we work with love our solution. We're embedded in their systems.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/8eec0f8d-22d8-4480-807f-f71d0ffbcfec" alt="" width="600" height="400"></p>
<h2 id="ember248" class="ember-view reader-text-block__heading-2">As a startup what are you finding the most difficult at the moment?</h2>
<p id="ember249" class="ember-view reader-text-block__paragraph">Good question. At the moment it&rsquo;s hiring&hellip; we have a really high bar. We've hired people from DeepMind, Microsoft and Amazon and want every new hire to make the whole team better.</p>
<p id="ember250" class="ember-view reader-text-block__paragraph">We're ambitious, doing big things, and striving for excellence. The people that join Scope are motivated and intelligent, so it&rsquo;s a challenge to find people with the right skills and mindset.</p>
<p id="ember251" class="ember-view reader-text-block__paragraph">The people we interview are always really excited about working with us. Obviously there&rsquo;s the technical challenge of building these tools, which is attractive to lots of people. But it's about more than that: our users say that it saves them their weekends and gives them more time back with family. Making that difference is something we all find incredibly fulfilling.</p>
<h2 id="ember253" class="ember-view reader-text-block__heading-2">In terms of scaling the product that you've built, what's the main challenge at the moment?</h2>
<p id="ember254" class="ember-view reader-text-block__paragraph">We've taken a very modular approach to development to solve different issues. It can be difficult because most inspection workflows are unique. So, we've been through many iterations.</p>
<p id="ember255" class="ember-view reader-text-block__paragraph">Now, as we scale, the challenge is to make sure that the data layer stays consistent so we're able to onboard customers faster and make it easier for them to make adjustments to their reports and workflows.</p>
<p id="ember256" class="ember-view reader-text-block__paragraph">As a small team, we've been able to stay very customer focused. As we scale, we want to keep that mindset &ndash; ensuring that everyone on the team gets the same level of customer involvement as we do now. It&rsquo;s something that we're going to keep at the core of our business as we grow.</p>
<h2 id="ember257" class="ember-view reader-text-block__heading-2">How do you see your technology evolving over the next few years? And how will it change what you've built so far?</h2>
<p id="ember258" class="ember-view reader-text-block__paragraph">We know that language models and AI systems will continue to get better. We've seen a steady increase in how these models perform, so I expect that more of our work can be automated in the future. That&rsquo;ll create opportunities for the businesses we work with.</p>
<p id="ember259" class="ember-view reader-text-block__paragraph">I see Scope evolving into a platform where businesses can take the data that we gather like all the instructions for a product, and then build their own custom analysis and calculations on top of the data.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/704dc4fe-23e0-4ea3-a73b-60b851973353" alt="" width="600" height="400"></p>
<h2 id="ember260" class="ember-view reader-text-block__heading-2">You&rsquo;ve got a clear vision for how the product will evolve as you scale - but how do you see the company changing?</h2>
<p id="ember261" class="ember-view reader-text-block__paragraph">At the moment, we&rsquo;re looking to make a few key engineering hires to scale the team. We&rsquo;re on the lookout for people with the right AI skill set, but mindset is just as important for us.</p>
<p id="ember262" class="ember-view reader-text-block__paragraph">We know nothing worth doing is easy, so everyone on the team needs to be passionate about solving real world problems.</p>
<p id="ember263" class="ember-view reader-text-block__paragraph">While we&rsquo;re 10x oversubscribed on our Pre-Seed funding, we predict we'll be raising a Series A in the next 9-to-12 months. We're aiming to have more than a million dollars in revenue in 2025, and &ndash; if we hit our growth aims by mid-2026 &ndash; we&rsquo;ll be on track to be doing 2-to-3 million dollars in recurring revenue, which is so exciting.</p>
<p id="ember264" class="ember-view reader-text-block__paragraph">As a part of the Series A fundraising, we'll be expanding to the US. We have a couple of customers out there who are eager to get involved in what we're doing. One day, we'll have deployment engineers in the Middle East and Asia too. Our ultimate aim is to be working with inspectors around the world.</p>
<h3 id="ember265" class="ember-view reader-text-block__heading-3">Scope is hiring!</h3>
<p id="ember266" class="ember-view reader-text-block__paragraph">Work on real-world AI with a pre-seed startup backed by tier-one investors. Contact <a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.linkedin.com/in/dan-clouting/?lipi=urn%3Ali%3Apage%3Ad_flagship3_company_posts%3BXL%2Fa1Xp%2FSt2h2foMBgxJ6w%3D%3D" target="_self" data-test-app-aware-link="">Dan Clouting</a> to learn more.</p>
<h2 id="ember267" class="ember-view reader-text-block__heading-2">When it comes to AI trends, what are you most excited about?</h2>
<p id="ember268" class="ember-view reader-text-block__paragraph">What excites me? This is really niche, but I'm quite excited by OCR (Optical Character Recognition).</p>
<p id="ember269" class="ember-view reader-text-block__paragraph">We&rsquo;re getting really good at being able to interpret really ugly hand-written documents, which means we can take a lot of data that's just written down and locked away in a filing cabinet somewhere and turn it into really useful data.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/9567ea09-3068-4140-969b-fd62acf49b2e" alt="" width="600" height="400"></p>
<h2 id="ember271" class="ember-view reader-text-block__heading-2">And what do you feel is overhyped?</h2>
<p id="ember272" class="ember-view reader-text-block__paragraph">I think agent frameworks are overhyped. A lot of people talk about AI agents and making them work, but so far, we haven't really seen enough reliability to make it work in an enterprise setting.</p>
<p id="ember273" class="ember-view reader-text-block__paragraph">While it may work for simple tasks, there's a lot of steps in more complex tasks where things can go wrong.</p>
<h2 id="ember275" class="ember-view reader-text-block__heading-2">What's the best advice you've ever received about leading a tech company and a startup?</h2>
<p id="ember276" class="ember-view reader-text-block__paragraph">The best advice I&rsquo;ve had while building Scope is to build a product that people actually need. Something that <em>genuinely</em> fixes a problem for people.</p>
<p id="ember277" class="ember-view reader-text-block__paragraph">It&rsquo;s easy to think about solutions and then try to find people that need that solution. But you should speak to people, find real problems, and <em>then</em> think about how you can build a solution to fix that problem.</p>
<h2 id="ember278" class="ember-view reader-text-block__heading-2">You&rsquo;re obviously incredibly busy. How do you switch off and relax outside of work?</h2>
<p id="ember279" class="ember-view reader-text-block__paragraph">Jakob and I do a lot of running. He recently did a marathon in Europe and I&rsquo;d like to have a go at the London Marathon one day, but it&rsquo;s just so hard to get into.</p>
<p id="ember280" class="ember-view reader-text-block__paragraph">I also love swimming. I do a lot of in-pool swimming, but when I can, I&rsquo;ll go open-water swimming. I&rsquo;ve also tried my hand at some long distance swimming when I&rsquo;ve had the opportunity. A few months ago, I went for a swim across a huge lake in Wales in the middle of winter, which was incredible, but also a <em>big</em> challenge and pretty scary at times!</p>
<h2 id="ember281" class="ember-view reader-text-block__heading-2">Finally, if you weren't working at Scope, what do you think you would be doing now?</h2>
<p id="ember282" class="ember-view reader-text-block__paragraph">Before founding Scope, I knew that I wanted to start a business. That&rsquo;s why I left my job and joined the Entrepreneur First program&hellip; but it&rsquo;s interesting to think where I&rsquo;d be if things turned out differently.</p>
<p id="ember283" class="ember-view reader-text-block__paragraph">If I wasn&rsquo;t building Scope and solving problems for the TIC industry, I'd probably be building another company doing something with AI. There's a lot of interesting problems out there that can be solved with technology and AI, so I&rsquo;d like to think if I wasn&rsquo;t fixing problems in the inspections process, I&rsquo;d still want to be fixing problems elsewhere.</p>
<h2 id="ember284" class="ember-view reader-text-block__heading-2">Want to be part of the team revolutionising TIC?</h2>
<div class="reader-image-block reader-image-block--full-width">
<p><img src="https://cdn.sourceflow.co.uk/__gallery/be8d77f4-5c44-4503-b297-585b45ea0490" alt="" width="600" height="600"></p>
<p id="ember286" class="ember-view reader-text-block__paragraph">Are you a driven AI Engineer, looking to make an impact and solve problems with technology? <a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.linkedin.com/company/generativegroups/" target="_self" data-test-app-aware-link="">Generative</a> has partnered exclusively with <a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.linkedin.com/company/scopeinspection/" target="_self" data-test-app-aware-link="">Scope</a> to find top talent for three critical roles:</p>
<p id="ember287" class="ember-view reader-text-block__paragraph">👉 Applied AI Engineer (&pound;90K&ndash;&pound;130K + equity)</p>
<p id="ember288" class="ember-view reader-text-block__paragraph">👉 Founding Full Stack Engineer (&pound;80K&ndash;&pound;120K + equity)</p>
<p id="ember289" class="ember-view reader-text-block__paragraph">👉 Forward Deployed Engineer (&pound;80K&ndash;&pound;120K + equity)</p>
<p id="ember290" class="ember-view reader-text-block__paragraph">These are permanent, London-based opportunities (Covent Garden) and offer a rare chance to get in early at a 10x oversubscribed pre-seed success story, backed by tier-one investors.</p>
<p id="ember291" class="ember-view reader-text-block__paragraph">📩 Get in touch with <a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.linkedin.com/in/dan-clouting/?lipi=urn%3Ali%3Apage%3Ad_flagship3_company_posts%3BXL%2Fa1Xp%2FSt2h2foMBgxJ6w%3D%3D" target="_self" data-test-app-aware-link="">Dan Clouting</a> / <a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="mailto:dan.clouting@generativegroup.ai" target="_self" data-test-app-aware-link="">dan.clouting@generativegroup.ai</a> to find out more.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/1993f6aa-ebed-4306-a563-2ddf177443ce" alt="" width="600" height="322"></p>
<p id="ember293" class="ember-view reader-text-block__paragraph"><strong>Scope is one of the 54 innovative AI organisations featured in our 'Faces of AI 2025' report, which is available to </strong><a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.generativegroup.ai/reports/faces-of-ai-2025/?utm_source=linkedin&amp;utm_medium=social" target="_self" data-test-app-aware-link=""><strong>download for free now</strong></a><strong>.</strong></p>
<p id="ember294" class="ember-view reader-text-block__paragraph"><strong>At Generative, we work with AI-driven startups to scale innovation, refine strategy, and accelerate growth. If you're building in AI and looking for the right talent to take your company to the next level, </strong><a class="UCfvfNsyTpsPQidhUnqqzZLAaQtxAVUA " tabindex="0" href="https://www.generativegroup.ai/contact-us/" target="_self" data-test-app-aware-link=""><strong>let's talk.</strong></a></p>
<figure class="reader-image-block__figure"><a href="https://www.linkedin.com/posts/generativegroups_scope-x-generative-activity-7323270478075232256-AP7q?utm_source=share&amp;utm_medium=member_desktop&amp;rcm=ACoAABgfqgoBCfoqMBKUwKk7zyapkunRbJbrdKY" target="_blank" rel="noopener noreferrer">
<div class="ivm-image-view-model    reader-image-block__img-container">
<div class="ivm-view-attr__img-wrapper
        
        ">&nbsp;</div>
</div>
</a></figure>
</div>

Jonathan Low and Jakob Cassiman, co-founders of Scope, were outsiders to the Testing, Inspection, and Certification (TIC) industry – and that’s exactly what gives them their edge. 

Disrupting a $250bn Industry: How Scope is Helping Firms Ditch Pen and Paper for AI

<h2 id="ember2727" class="ember-view reader-text-block__heading-2">While AI continues to evolve, Eduardo Candela is on a mission to ensure it&rsquo;s used with care, clarity, and confidence.</h2>
<p id="ember2728" class="ember-view reader-text-block__paragraph">As the co-founder and CTO of Maihem, his priority is clear: help enterprises embrace AI with confidence, bridging the gap between cutting-edge technology and real-world deployment.</p>
<p id="ember2729" class="ember-view reader-text-block__paragraph">We sat down with Eduardo&nbsp;to hear more about his mission, the challenges of building trustworthy AI, and how his PhD and industry experience laid the perfect foundation for his entrepreneurial journey.</p>
<hr>
<h3 id="ember2731" class="ember-view reader-text-block__heading-3">Let&rsquo;s talk about the origins of Maihem &ndash; how did the idea come about?</h3>
<p id="ember2732" class="ember-view reader-text-block__paragraph">I started my career in the automotive industry, working in Silicon Valley at Tesla, and then later at Bosch Centre of AI. That&rsquo;s where I saw firsthand what we call mission-critical AI: systems being deployed in cars, where safety and performance are essential. It became clear how important it is to not just build powerful AI, but to keep it safe and under control.</p>
<p id="ember2733" class="ember-view reader-text-block__paragraph">Before I worked in Silicon Valley, I did my Masters at MIT, where I also conducted research on using AI in mission-critical decision-making processes. Five years ago, I moved to London to pursue my PhD at Imperial College London to research safety in self-driving cars, specifically how to measure risk and train models to be risk-aware and reliable using reinforcement learning.</p>
<p id="ember2734" class="ember-view reader-text-block__paragraph">Max, my co-founder, took a different path. He was at McKinsey before starting a PhD in Natural Language Processing, where his research focused on how to make generative AI safer &ndash; specifically, how to detect and prevent the spread of conspiracy theories.</p>
<p id="ember2735" class="ember-view reader-text-block__paragraph">We met while still in our PhDs and we immediately connected over a shared passion and a shared question: how do we close the gap between AI research, and real and responsible deployment? <strong>Maihem</strong> became our answer. It&rsquo;s a way to take everything we learned in research and apply it to help companies deploy AI with confidence.</p>
<h3 id="ember2737" class="ember-view reader-text-block__heading-3">How do you think your background in academia has shaped your startup journey?</h3>
<p id="ember2738" class="ember-view reader-text-block__paragraph">When I left my job in Silicon Valley to pursue a PhD, people thought I was crazy&hellip; walking away from the money and the title. But I knew I wanted to live in the future, to go deep on the questions I cared about. A PhD gave me that.</p>
<p id="ember2739" class="ember-view reader-text-block__paragraph">It&rsquo;s still not common for people in academia to start startups. But for me, it makes perfect sense. You're using science to test ideas, you&rsquo;re running experiments, you&rsquo;re validating things, failing, refining, and spending years going very deep to solve a very specific problem that nobody has solved before. That mindset is exactly what you need to build something from scratch.</p>
<p id="ember2740" class="ember-view reader-text-block__paragraph">Both my co-founder and I completed PhDs, and we&rsquo;ve found that background to be incredibly valuable. It&rsquo;s a huge investment of time and energy, so you have to be really clear about what you want to get out of it. But if you are, it&rsquo;s absolutely worth it.</p>
<p id="ember2741" class="ember-view reader-text-block__paragraph">So while it doesn&rsquo;t have to be a PhD, I really encourage anyone who&rsquo;s passionate about answering hard questions and uncovering truth to consider startups.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/b8136b3a-10eb-4cc5-bae4-b71178cc6d32" alt="" width="600" height="400"></p>
<h3 id="ember2743" class="ember-view reader-text-block__heading-3">There's a lot of uncertainty around AI within the wider population. From your perspective, what are some of the most common misconceptions out there?</h3>
<p id="ember2744" class="ember-view reader-text-block__paragraph">We see it across the board &ndash; people are very aware of AI&rsquo;s power, but there&rsquo;s still a lot of skepticism around it. One of the key things many people don&rsquo;t realise is that these new models are <em>probabilistic</em>. You can give them the same input twice and get two completely different outputs.</p>
<p id="ember2745" class="ember-view reader-text-block__paragraph">That unpredictability can be hard to wrap your head around when you&rsquo;re expecting consistency. Hallucinations are one of the biggest blockers to adoption right now. When an AI gives users a wrong answer, people often lose trust in the entire system and immediately disengage.</p>
<p id="ember2746" class="ember-view reader-text-block__paragraph">Of course, there&rsquo;s a more technical layer to this too, where models can be trained in the wrong direction. But for most users and businesses, the biggest challenge is just getting these systems to perform as<em> </em>expected.</p>
<p id="ember2747" class="ember-view reader-text-block__paragraph">There&rsquo;s a real gap between what people <em>think</em> AI should be capable of and how these technologies actually work. They&rsquo;re still not perfect. We want people to confidently harness the power of AI, while also being aware of the limitations. Bridging that gap between expectations and reality is key.</p>
<h3 id="ember2749" class="ember-view reader-text-block__heading-3">What&rsquo;s the most exciting technical challenge that your team has solved so far?</h3>
<p id="ember2750" class="ember-view reader-text-block__paragraph">One of the biggest challenges we&rsquo;ve been tackling is how to bring user feedback and preferences directly into how AI systems are built and behave.</p>
<p id="ember2751" class="ember-view reader-text-block__paragraph">Right now, AI gets developed by engineers, it&rsquo;s deployed, and then users interact with something that doesn&rsquo;t always meet their expectations. So we&rsquo;ve been asking: how do we close that loop? How do we make users feel in control, and actually get what they want from these systems?</p>
<p id="ember2752" class="ember-view reader-text-block__paragraph">It&rsquo;s really a human&ndash;AI alignment problem, and solving it is key if we want people to trust and adopt these tools.</p>
<h3 id="ember2753" class="ember-view reader-text-block__heading-3">&lsquo;AI Safety&rsquo; can mean very different things to different people. What does that term mean to you, and what does it look like in practice?</h3>
<p id="ember2754" class="ember-view reader-text-block__paragraph">That&rsquo;s a great question. For me, one of the key realisations &ndash; especially during my PhD research on self-driving cars &ndash; is that safety is about performance under constraints.</p>
<p id="ember2755" class="ember-view reader-text-block__paragraph">When I used to run simulations, the safest car was the one that just parked and never moved. Technically, that&rsquo;s &ldquo;safe&rdquo; but it&rsquo;s also useless. All of us, when we take a plane, when we drive a car, when we go out, we're always accepting some level of risk in order to achieve a goal.</p>
<p id="ember2756" class="ember-view reader-text-block__paragraph">The question then becomes: how do we achieve a goal without crossing certain boundaries? How do we maintain control? That&rsquo;s what AI safety looks like to me. The real challenge comes in defining and measuring those limits, and ensuring users feel they are in control.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/561c8268-1e75-422b-8392-81f08709d05a" alt="" width="600" height="400"></p>
<h3 id="ember2758" class="ember-view reader-text-block__heading-3">What is the biggest challenge in bringing AI products like yours to market?</h3>
<p id="ember2759" class="ember-view reader-text-block__paragraph">The biggest challenge (and the biggest opportunity) is the gap between the development of cutting-edge AI and real-world adoption.</p>
<p id="ember2760" class="ember-view reader-text-block__paragraph">You look at the headlines and see OpenAI, new models, new agents, new startups. But when you talk to enterprises &ndash; and the everyday users within their companies &ndash; there&rsquo;s often a huge disconnect. Most companies don&rsquo;t have the talent, the understanding, or the confidence to use these tools effectively.</p>
<p id="ember2761" class="ember-view reader-text-block__paragraph">I call this the &ldquo;chasm of adoption&rdquo; and that&rsquo;s where we come in to bridge the gap. It&rsquo;s not just about technology, it&rsquo;s about mindset. A big part of the work is education: helping people understand how these systems work, and how to use them safely and effectively.</p>
<h3 id="ember2763" class="ember-view reader-text-block__heading-3">How do you see your technology evolving over the next few years?</h3>
<p id="ember2764" class="ember-view reader-text-block__paragraph">AI will continue to improve and become embedded in every part of life. Our mission is to make sure that progress benefits everyone by making AI safe and trustworthy.</p>
<p id="ember2765" class="ember-view reader-text-block__paragraph">In the short term, we&rsquo;re seeing a flood of new models, agents, workflows &ndash; both open and closed source. But most of what&rsquo;s out there today is still what we&rsquo;d call level one or two autonomy. These tools are co-pilots. They support humans, augment their work, but don&rsquo;t operate independently.</p>
<p id="ember2766" class="ember-view reader-text-block__paragraph">At some point, there&rsquo;s going to be a shift. We don&rsquo;t know exactly when, but we&rsquo;ll move into true autonomy, where these systems don&rsquo;t just assist &ndash; they act on our behalf.&nbsp; That&rsquo;s the big opportunity for us.</p>
<p id="ember2767" class="ember-view reader-text-block__paragraph">Once agents become autonomous, you need to know: are they safe? Are they in control? Can they interact with each other reliably? We&rsquo;re not there yet, but the foundations are being built now. Our goal is to help organisations prepare &ndash; so when autonomy becomes the norm, they&rsquo;re ready.</p>
<div class="reader-image-block reader-image-block--full-width"><img src="https://cdn.sourceflow.co.uk/__gallery/30a048e8-1394-4323-9b1e-7feb14b08c89" alt="" width="400" height="400"></div>
<div class="reader-image-block reader-image-block--full-width">
<h3 id="ember2769" class="ember-view reader-text-block__heading-3">When it comes to AI trends, what are you most excited about?</h3>
<p id="ember2770" class="ember-view reader-text-block__paragraph">I did my PhD in reinforcement learning, so I&rsquo;ve been close to it for a while. It&rsquo;s exciting to see it finally gaining more attention; I find it very promising. Everyone&rsquo;s talking about it now because it mirrors how humans learn: through rewards and feedback.</p>
<p id="ember2771" class="ember-view reader-text-block__paragraph">With the rise of new models like DeepSeek and others, reinforcement learning is proving to be a really powerful and accessible way to improve agent behaviour. It&rsquo;s a current trend that&rsquo;ll only grow further.</p>
<h3 id="ember2772" class="ember-view reader-text-block__heading-3">And what do you feel is overhyped?</h3>
<p id="ember2773" class="ember-view reader-text-block__paragraph">We hear a lot of talk of AI being super dangerous or close to taking over the world. But that&rsquo;s just not where we are. Especially with LLMs, people sometimes forget that these are just models trained to predict the next token, the next word.</p>
<p id="ember2774" class="ember-view reader-text-block__paragraph">We&rsquo;re nowhere near Artificial General Intelligence (AGI). These systems are not self-improving, they use huge amounts of energy, and they&rsquo;re still far away from replicating human intelligence in many domains. So while LLMs are powerful tools, we&rsquo;ll likely need different, more powerful architectures to get anywhere close to AGI.</p>
<h3 id="ember2776" class="ember-view reader-text-block__heading-3">Maihem is currently a team of five. How has the talent attraction process been for you as a startup?</h3>
<p id="ember2777" class="ember-view reader-text-block__paragraph">It&rsquo;s really exciting, but challenging. Hiring for a startup is nothing like hiring for a big company. I&rsquo;ve done both, and they&rsquo;re completely different worlds. You talk to people coming out of Big Tech and often, they just don&rsquo;t have the mindset or culture fit for a startup.</p>
<p id="ember2778" class="ember-view reader-text-block__paragraph">What you&rsquo;re really looking for is your own SWAT team: strong generalists who are smart, driven, and deeply passionate about the mission. We&rsquo;ve been really lucky with the team we&rsquo;ve assembled. Most of them have been founders themselves, so they already understand what it takes to build something from scratch.</p>
<p id="ember2779" class="ember-view reader-text-block__paragraph">That founder mindset has made a huge difference. Instead of chasing the obvious CVs, we looked for people who genuinely <em>want</em> to prove themselves and make a difference. That&rsquo;s where we&rsquo;ve found the most success.</p>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/99c75b27-56cd-4e35-9240-86a784b30068" alt="" width="400" height="400"></p>
<h3 id="ember2781" class="ember-view reader-text-block__heading-3">You&rsquo;re obviously incredibly busy. How do you switch off and relax outside of work?</h3>
<p id="ember2782" class="ember-view reader-text-block__paragraph">For me, work-life balance isn&rsquo;t about separating work from life. Finding passion in what you do makes all the difference.</p>
<p id="ember2783" class="ember-view reader-text-block__paragraph">I&rsquo;ve always loved diving deep into tough problems; staying up late solving something challenging never felt like a chore. I love reading, especially about startups and tech. It&rsquo;s not work for me, it&rsquo;s real curiosity.</p>
<p id="ember2784" class="ember-view reader-text-block__paragraph">I also love music; I play piano and guitar &ndash; it puts my brain in a different mode and often helps me solve problems in unexpected ways. And of course, I try to stay active. I play football every weekend and tennis when I can. That's how I keep my energy up!</p>
<h3 id="ember2785" class="ember-view reader-text-block__heading-3">Finally, if you weren't working on Maihem, what other problems would you be trying to solve with AI?</h3>
<p id="ember2786" class="ember-view reader-text-block__paragraph">My passion and career have always been focused on one thing: how to build systems that make smart, automated decisions. Even before everyone was calling it &ldquo;AI,&rdquo; that&rsquo;s what I was drawn to.</p>
<p id="ember2787" class="ember-view reader-text-block__paragraph">So if I weren&rsquo;t focused on helping companies deploy AI safely and responsibly, I&rsquo;d probably be building something similar myself &ndash; maybe a vertical agent tackling a specific, meaningful and safety-critical use case.</p>
<p id="ember2788" class="ember-view reader-text-block__paragraph">I love the idea of building AI that helps people with their mental health, or invest their savings safely. But right now, my goal is to help all of these organisations first, to maximise my impact in the world.</p>
<hr>
<p class="ember-view reader-text-block__paragraph"><img src="https://cdn.sourceflow.co.uk/__gallery/1993f6aa-ebed-4306-a563-2ddf177443ce" alt="" width="746" height="400"></p>
<p id="ember2790" class="ember-view reader-text-block__paragraph"><strong>📥 Eduardo is one of the 54 innovative technical founders recently featured in our 'Faces of AI 2025' report, which is available to </strong><a class="HyRpFmCjAWIuRxNzmfOWKyciaWbRBoTwI " tabindex="0" href="https://www.generativegroup.ai/reports/faces-of-ai-2025/?utm_source=linkedin&amp;utm_medium=social" target="_self" data-test-app-aware-link=""><strong>download for free now.</strong></a></p>
<p id="ember2791" class="ember-view reader-text-block__paragraph"><strong>At Generative, we work with AI-driven startups to scale innovation, refine strategy, and accelerate growth. If you're building in AI and looking for the right talent to take your company to the next level, </strong><a class="HyRpFmCjAWIuRxNzmfOWKyciaWbRBoTwI " tabindex="0" href="https://www.generativegroup.ai/contact-us/" target="_self" data-test-app-aware-link=""><strong>let's talk.</strong></a></p>
</div>

Eduardo Candela is on a mission to ensure AI is used with care, clarity, and confidence.

AI Signs Set to Improve Beach Conditions in the South West of England

Cool Heads, Cooler Systems: The Startup Making AI Work for Energy Efficiency

Disrupting a $250bn Industry: How Scope is Helping Firms Ditch Pen and Paper for AI

Why Trust, Not Tech, Is AI’s Biggest Challenge: Eduardo Candela on Founding Maihem