Dr Kate Jarvis has spent her career bringing advanced software to industries that technology had largely passed by. After a PhD at Stanford and years building data-driven products in edtech, followed by a spell mentoring founders at Seedcamp, she turned her attention to real estate, an industry running billion-dollar portfolios on spreadsheets and scattered documents. In 2023 she and co-founder Johnny Morris launched Fifth Dimension, a decision intelligence platform that unifies fragmented institutional data and unstructured documents into a single layer, with agentic AI that prepares analysis, surfaces risk and readies decisions before a team has even thought to ask.

The approach has won the confidence of some of the sector’s largest names, among them BXP and Realty Income, with trillions in assets now running through the platform across offices in London, New York and Singapore. A recent $26M Series A, led by HV Capital, takes total funding beyond $40M and supports expansion across the US, already the source of 70% of revenue, and Asia Pacific. Here, she talks us through the path from a business run off a spreadsheet to a company spanning three continents, what a purpose-built platform can do that general tools cannot, and where she expects decision intelligence to take the industry next.

You spent years in edtech and data-driven product leadership, and later mentored founders at Seedcamp, before launching Fifth Dimension in 2023. What was the gap in the market you recognised and how did your earlier experience help you to shape the right solution?

“Before we started our own business, Johnny and I worked together trying to get more first-time buyers onto the housing ladder. But we kept losing deals and clients because our underwriting was too slow. Picture 30 people in the office copy-pasting comparables off the internet, screenshotting photos into PDFs, building a 660-page investment committee memo. For a 240k home. We couldn’t help but wonder: if this was what it was like for us, what were all the top investment and asset managers in the world doing? Were they really running billions of dollars of transactions through Excel? They were, and many still are. The only reason it wasn’t catastrophic was that the returns were high enough to absorb some of the opex and opportunity cost. But it was bananas. 

“If you peel back the layers of the problem, we know that real estate sits on the world’s biggest gold mine of data. It’s what gives each business its edge – that historic transaction and performance data. How did my data centers perform last year? What were the ten year returns on our urban Class A office portfolio? But these data are locked away in 1000s of different files and storage systems. We knew there had to be a better way.

“My experiences getting my PhD at Stanford, then working at startups in Silicon Valley and London, and mentoring other founders, were a fifteen year build-up to that realization. I’ve always focussed on how to build and deploy cutting edge software for the least tech savvy industries and consumers; in ed tech, I built machine learning-based products to teach kids math and literature. But I onboarded teachers to our platforms who needed to be shown how to open a web browser. I’ve always been motivated by bringing the best technologies where they’re needed most, where the power of digital transformation can help people leapfrog decades of incremental progress. 

“That’s exactly what real estate professionals need, paired with a deep empathy for the workflows these analysts and managers complete every day. (My cofounder, Johnny, spent many years bringing data warehousing to traditional real estate businesses before we joined forces.) 

“When we got early access to GPT-3 and started to build with and around it, we had the same lightbulb moment everyone did of, “Wow, this is going to change the world.” It was the start of a solution to our problems. But I also knew how much superficial, bad software was about to be sold into our industry by people with no grasp of the technology. We weren’t gonna let that happen unchecked! 

“Fifth Dimension was years ahead of the times, we were building harnesses in 2023. But everything 5D is today traces back to us being furious we were running a business off a spreadsheet.”

The platform brings together fragmented institutional data and unstructured documents into a single intelligence layer. What does that change about the way a firm decides where to deploy capital?

“Three major changes take place in the businesses we work with. 

“First, the deployment of internal resources shifts from the doing of rote, repetitive, error-laden work like updating waterfalls manually in Excel, to sourcing more new deals or spending more time at a development site, actually checking up on progress. 

“This also means more data are actually produced for later use throughout an asset’s lifecycle. Today an acquisitions team ends up screening a fraction of the deals actually hitting their desk, because screening one properly is such heavy work. So capital gets committed to the deals that happened to get an analyst’s attention, not necessarily the best ones. And the deals that are passed on often just live in deleted email attachments, which is a wasted opportunity to better hone the business’s buy box. With 5D, more deals get screened and logged and with all that data in one intelligence layer, the team gets to bid faster, catch problems earlier, and walk away from bad deals sooner.

“So that’s the second change. The assets firms choose to buy and sell, when, and on what terms, shifts as data are more accessible and intelligible – so that Japanese knotweed near a development is noticed sooner, knocking $10M off the price, and the nasty sub-clause that makes a logistics facility’s partial triple net lease even more partial is discovered before you purchase. And these discoveries, too, are logged as data. So as the team screens every deal against their target IRR, yield on cost and cash-on-cash, etc. and stress-tests it against properties already owned in the same submarket, 5D is creating a fully auditable trail for every IC and for public companies, for their auditors, too! Every decision is fully defensible, because every figure traces back to where it came from.

“Third, as businesses work with us year on year, multiplying their data each year, this intelligence layer  enables them to be more proactive than ever before. Suddenly, teams don’t just react to inflation and rising fuel prices. Instead, the platform warns you of red flags that mean your hospitality portfolio may bleed 80bps next quarter, recommends actions to take, and starts working on those actions in the background, using a combination of your own data and real-time market analysis. Maybe you’re even able to capture higher rents per square foot in a retail centre near where several big name bands are set to play this year, because you’ve got the data and scenarios mapped out to show how much more business they’ll do as a result.  

“In sum, businesses working with Fifth Dimension stop deploying capital on today’s whims, constrained by today’s small teams, and start deploying it on the full picture – faster, and on evidence the team can stand behind. In the end, that is how you turn the world’s biggest gold mine of data into alpha: capital goes into the right deals, the bad ones get caught before they cost you, and the best ones get won while a slower competitor is still diligencing them.

Real estate is one of the most detailed asset classes there is, with a single asset carrying years of financing, dozens of stakeholders and deep layers of compliance data. You built for that complexity from the start, and it gives the platform a real edge. What can a purpose-built approach achieve here that broader solutions cannot?

“We’re a data intelligence system of insights and actions, not an AI chatbot. The platform connects the firm’s shared document estate and systems, extracts and structures commercial real estate data via concepts specific to your business, reconciles conflicting data points, cites the evidence, and writes the result into a shared source of truth for the relevant deal, asset or portfolio. The team gets faster answers because the data is already structured. The firm gets safer answers because the evidence is inspectable. And better decisions are made because the right insights are surfaced by 5D at the right time, before others have a chance to act.

“When a new data room lands, it comes with hundreds of documents, and the old way is that a person trawls through it and hopes nothing gets missed. The platform reads every file before anyone opens one, understands the financial differences between a hospital and a wind farm, and gives you back a structured picture: what’s there, what’s missing, what needs your attention. It will flag that the environmental report and the title documents aren’t in the room, and that’s a conversation with the vendor on day one, not week three. It finds the four versions of the rent roll, reconciles the numbers across all of them, and cites exactly which file each figure came from. Then it analyzes the deal, reading the financials and the leases together, because in a real underwrite they are the same problem, and comes back with a verdict every figure can be traced to.

“And it keeps working when the team logs off. Overnight, it checks the data room against your diligence checklist, flags what’s still outstanding, and drafts the email to chase the broker, so the team walks in to decisions, not to-do lists. That is what LLM harnesses (“doing systems”) with multiple specialised, real estate-tailored models achieve. 

“You can pipe data into Claude naively and all you get is “garbage in, garbage out.” Congratulations, you’ve given an LLM more data on an asset or your portfolio than it can parse, so be prepared for it to lie to you, fluidly and silently. And there’s no structure of intelligence around the data, so there’s no way of the LLM knowing whether the valuation from your CRM or the valuation from your broker is the correct one, even if you manage to join those data together yourself. 

“It’s easy to put a wrapper around an LLM that makes the industry feel spoken to (“look I can see my lease and rent roll side by side!”). It’s a lot harder to deliver quality, accuracy, and consistency at scale – across a whole, complex, multi-faceted business. That’s why we win. 

“What if Claude goes down today, as it does often? Do you want to tell your Board your sustainability plan doesn’t include a provision for multiple LLM tool use in case one goes down? Or do you want to spend most of your time with your CFO talking about controlling for Anthropic or OpenAI token costs because it turns out these tools make you pay 5-10x what you expect for any “real” data-driven work? No, let an actual software business that worries about problems like these, and only problems like these, take care of that for you. 

“You bring the data and documents, we bring the tech (that helps you create, log, and cleanse more data), and the LLMs are just tools under the hood we can trade in and out if needed. The firms that have tried to build tech like ours themselves spent tens of millions and still can’t get their own people to use what they’ve built. Why? It doesn’t work well enough. 

“Plus, these businesses don’t have large teams who have worked across 100s of businesses like yours, specifically to help you transform your business. We do. Our users are in Fifth Dimension every day.”

Firms currently using the platform are deploying five times more capital with the same team, lifting net operating income by 5% and underwriting complex deals in days rather than weeks. What is it about the way the platform works that allows firms to unlock gains on this scale?

“We know what work you actually do, and what your metrics are for success. So we’ve mapped out 75+ common processes for automation that take you from AI 0 to hero – better operational efficiency, faster action, and better financial performance. We tell you where to start, train you on the tool to get comfortable, and help you progress through the stages of AI adoption until the gap between you and your competition is so wide, they can’t even see you anymore. 

“Today, a firm’s most experienced people spend far too much of their time on data wrangling rather than judgment and analysis. The bottleneck was never the thinking, it was the assembly, finding the documents, reconciling the versions, validating and formatting, and that is exactly the part the platform takes off them. Free up that time and the same team covers far more ground, which is where deploying five times the capital comes from. 

“The work also stops disappearing into one person’s email inbox or chat thread. Fifth Dimension organises your data into a single source of truth per asset or portfolio, called a workspace, where people can collaborate. So the fiftieth deal this year is diligenced to even better quality than the fifth, because the context compounds as you go. 

“That’s also why underwriting that used to take weeks now takes 1-2 days: the mishmash of people, tasks, and conflicting data that caused the delays is transformed into clear, simple to review outputs and suggested actions. The same engine, turned on the assets a firm already owns, is what lifts net operating income. It catches the lease renewal option, the expense spike, the variance a stretched team would otherwise only catch at the quarterly review, when the window to act has already narrowed. Catch those potential issues early, act early, and the cash flows back to you.”

Some of the largest names in the sector, including BXP and Realty Income, now run on the platform, and you hold ISO 27001 and SOC 2 Type II certification. How important has earning that institutional trust been to the company’s growth?

“It’s been fundamental. In our market, the biggest firms simply won’t move first on an unproven vendor, so earning their confidence is what lets you grow at all. Honestly, that trust starts with two things:

“A secure product. That’s more important than just achieving the certifications. You actually have to build product and processes that are secure by design. That’s what we do. Every customer we work with has their own instance of the product. Your data stays yours. 

“A team full of safe, competent hands. When BXP and other early customers chose to work with us, they weren’t just choosing a software product, they were choosing a team. A strategic partner they could open their critical business processes and data to for automation, with hope and trust. It’s the people at 5D who matter most here.  

“With that as the bedrock, names like BXP and Realty Income open doors with others, because in this industry businesses watch closely who their peers rely on. Again, it’s all about trust in people, not just products.”

Your recent $26M Series A, led by HV Capital, takes total funding beyond $40M and supports expansion across the US and Asia Pacific. With the US already accounting for 70% of revenue, where do you see the next phase of growth coming from?

“It will come from going deeper where we already are. The US is 70% of our revenue, and the plan is to keep doing what’s working there, bigger and better. We tend to land with one team and then expand, getting so deep in a firm’s data stack and producing more and more useful insights, that the work just keeps spreading, from one function to the next, from a few seats to the whole business.

“Then we’ve just started cracking Asia Pacific, with Singapore as the launchpad. Real estate operations are locally interconnected, so you have to go where the operators are, which is why we opened an office in Singapore. Like the US, APAC moves faster than Europe as there’s an appetite for urgent digital change and excitement for continuing economic growth, so we expect that market to overtake our European revenue in the coming years.

“Fifth Dimension’s expansion will be fueled by a further investment in community building, as well as growing our global R&D team. Just as it was once really cool to bank with Monzo, with their brightly coloured cards and revolutionary energy in an old school industry, we want people to be excited to be a part of 5D’s global community. Whether it’s a hackathon with analysts in Tokyo or an exclusive dinner with yours truly and some of the tech industry’s biggest luminaries in San Francisco, we want to make Fifth Dimension an experience of change, not just of software.

“But of course, we’ve gotta stay ahead of the foundation model labs, too, which requires us to keep building world class technical teams across geographies.” 

You have built a presence across London, New York and Singapore in a short space of time. For leaders looking to bring AI into established, complex industries, which qualities do you think matter most?

“The biggest one is a willingness to go genuinely deep, and to resist the quick buck. It’s tempting to put a nice wrapper around an LLM, make the industry feel spoken to, and sell. That doesn’t create lasting customer value, so it’ll never build a lasting company (ie create lasting shareholder value). 

“We spent years getting into the weeds of the actual work, the thousands of subtasks in underwriting a single deal or managing an 80-property portfolio, because that’s the only way to build differentiated technology – an obsession with the actual work the software is replacing or augmenting, and a technical team behind it who is equally obsessed with the power and the limitations of LLMms and adjacent technologies. The combination of the two is the magic. The complexity on both sides – the industry and the tech – isn’t the obstacle, it’s the protection. 

“The second is humility about the industry itself. These are established, relationship-driven businesses, and you don’t win them with features. You win by earning the trust of the right people, because the industry is a network and word travels fast. You have to respect that the people who’ve done this for decades know things you don’t, and build with them, not at them.

“Third is conviction for the long game. Bringing AI into a conservative industry is never a flip of a switch, it’s a slow behavioural change, and there will always be a reason to wait or to take an early exit. The founders who win are the ones who hold the line on the bigger mission and keep building until they’re so far ahead that nobody can catch them.”

You have said that firms moving now will define their competitive advantage for the next decade. What practical first steps would you encourage other leaders to take as they weigh up adopting these tools?

“At 5D, we always say you have to learn how to walk before you can run. The most important thing is to believe in longterm transformation for your business, but to take small, actionable steps towards that vision now. The instinct in big firms is to commission a twelve-month AI roadmap and a strategy deck, and you’ll spend a year in meetings and ship nothing. Don’t boil the ocean. Pick one painful, real workflow, something your team does over and over and hates, and get a quick win there first. But ensure this fits into a broader plan for getting ROI from AI, or these little bits of value across your business will never aggregate to anything.

“That’s why, of course, we advocate for partnering with 5D; we’ve mapped out how to get from crawl, to walk, to run with AI, by doing this exact journey over and over before with businesses just like yours. And our Solutions Engineers and Customer Success team are there the whole way, to ensure your use of AI tools delivers those hard line returns. All you’ll get from Anthropic’s Account Manager is a blank look when you talk about NOI.

“And please, don’t assume you have to build tech yourself in order to maintain your special sauce. Your IP is your data; that stays yours. But it’s a fact that some of the most sophisticated firms in the world have spent millions to build internal AI solutions, and still can’t get their own people to use what they built. Their active user percentages are often as bad as Co-Pilot’s, topping out at 20%. The depth of technical knowledge it takes to turn processes into LLM powered software, and the fact that the best model changes every few months, make building in-house a trap. Buy what already works, and put your real energy into adoption.

“Then find your tinkerers. In every firm there are a few curious people already experimenting, often already quietly using AI tools on the side. Those people are your biggest competitive advantage, and put them together with a forward deployed team like ours, and they’re gold. Put a real tool in their hands, let them learn by doing, and let what works spread out from them. 

“Finally, start now. This is exactly why the firms moving today will define the next decade. The purchase moment for law already happened. The purchase moment for our industry is happening right now, and the firms that don’t move soon won’t be a few months behind, they’ll be three, five, even ten years behind, because the gap compounds. You don’t close that with a strategy document. You close it by getting started this quarter.”

Deal cycles are compressing and data volumes keep rising. Looking ahead, how do you expect the role of decision intelligence to develop across real assets over the coming years?

“The biggest shift is from reactive to proactive. Today, even with a good platform, you mostly ask questions of leases or Excel models and get answers. Where this is going is that the system tells you what you didn’t know to ask. It’s watching your portfolio in the background and saying, three of your tenants match the early-warning pattern of ones that churned last year, here’s what it will cost you, and here’s what to do about it. Teams stop arriving to a to-do list and start arriving to decisions.

“And it doesn’t stop there. The next step is the platform taking action, reaching out to the tenant who looks likely to leave, sourcing replacements, starting the lease process, so a person only steps in for the judgment call. As deal cycles compress and the data keeps compounding, this stops being a luxury. If your competitor walks into the investment committee with a fully screened, fully sourced decision that ran overnight while your team is still assembling the numbers, you’ve lost the deal before the meeting starts.

“Over time, AI stops being “just software” that users pop in and out of, and becomes the nervous system the whole business runs on, with every major real-asset decision powered by its intelligence. 

“And the part of the mission I care about most is what that does to access. I want real assets to be more equitable, open to everyone – to live in, to love, and to trade, not just institutional investors. When the data finally becomes alpha, a small, sharp team can take on the giants. My hope is that by the end of the decade, some ten-person outfit of people who didn’t go to college beats the returns of the biggest funds in the world for the first time. That’s the world I’m building towards.”