In his influential book ‘Crossing The Chasm’, Geoffrey Moore defined new labels for innovators and early adopters ‘technology enthusiasts’ and ‘visionaries’. I use the blanket term ‘geek’ and I’m proud to call myself one.
In my long career with software product companies, I’ve been involved in many ‘innovative’ new product launches. I often think about this because it highlights the undue emphasis we can place on new technologies, perhaps neglecting opportunities to find new ways to create value using our existing technologies.
If we look at the ‘proptech’ space at the moment, there’s a plethora of exciting, upcoming vendors doing amazing things with AI, IoT, VR/AR, data analytics, cloud native, integration and mobile tech. Mmmm, lovely tech.
AI is all the rage, still
If you read any technology news then you’ll be aware that the ‘hyperscalers’ (Google, Meta, Microsoft et al) are engaged in an arms race of investment into ever-more powerful large AI models. Microsoft, Amazon and Meta each have plans to secure nuclear power plants (yes, really) specifically for powering datacentres building AI models. Of the 25 technologies listed on this year’s Gartner Hype Curve for Emerging Technologies, 12 are AI technologies and half of the others rely on AI. Agentic AI (i.e. AI that can work out how to do what you want and then get on with it) supported by the development of ‘large reasoning models’ is the NKOTB (Google late 1980s/early 90s boy bands) and promises to revolutionise the use of AI in many contexts.
As Wharton professor and AI thought-leader Ethan Mollick recently put it, “…for better and for worse, we’re far from seeing the end of AI advancement.” Make no mistake, the amazing breakthroughs of recent years aren’t about to dry up. Google has just announced that its NotepadLM tool now lets you interact verbally with the podcast it creates from your documents, so how long will it be before we’re providing customer service via life-like, always-on avatars that never have a sick day or a bad mood?
However, I’d say that the most useful use-case for these exciting developments is possibly also the most boring.
It’s the data, stupid
These days we like to describe ourselves as working in ‘tech’, ‘digital’ or ‘innovation’. These cool-sounding epithets mask our true raison d’être; providing easy access to reliable information in order to improve efficiency and effectiveness.
Adding any amount of cool-kids’ tech, slick apps or eye-candy visualisation tools means diddly squat if your underlying data is fragmented, contradictory, inaccessible, out of date, miscategorised or just plain wrong. It’s arguably ‘jumping the shark’ to try to predict that your tenant might suffer from mould due to damp if you can’t respond to the fact that they’ve already told you about it many times.
“The beginning of wisdom is the definition of terms.”
Socrates
We’ve been trying to solve this problem for some time. My whole career has centred around it, whether it be attempts to create monolithic database systems, business semantic layers for reporting, open integration platforms, data standards, data-quality tools, design standards or business frameworks. Today, I’m working with Esuasive whose cloud-native platform is built around the principle of well-defined, authentic and up-to-date data integrated across the enterprise.
Today, it’s sadly still common that housing providers have yet to achieve a single, trusted source of information for their businesses. In the end, the problem has always boiled down to the fact that it’s human beings who do the important stuff, and humans generally aren’t deterministic when it comes to describing or communicating things. IT leaders face a constant battle to provide solutions for their organisations that people can actually use, and with which they’ll engage.
Not just the UI for AI
In a recent partner briefing, Microsoft’s CTO for housing Kishore Rajendran described how it wants Copilot to be ‘the UI for AI’. I’d go one step further and suggest that LLM-powered interfaces and agent technologies such as Copilot will become the UI for everything. Think ‘Jarvis for Housing’.
The point I’m trying to make is that today, we expect people to type information into screens or phones. This is an incredibly inefficient and error-prone method of information transfer. If you’ve used dictation when writing emails or texts, you’ll know that it’s now very accurate and much faster than even a seasoned keyboard jockey like me can do.
Taking this a step further, I’d say that the single most useful capability of LLMs isn’t to generate rubbish marketing images or write banal social media posts but to interpret what you say to them or show them. There’s a massive potential to massively improve data quality (dare I say, information quality) by using Copilots to capture information from humans, interpret and classify it and sense-check and validate it before committing it to the corporate corpus. Arguably some of the main causes of data inaccuracies are:
- Effort required from, and friction encountered by, people typing stuff into screens, causing them to make mistakes or cut corners;
- Differences in interpretation or expression of facts or classifications between humans;
- Differences between product vendors in methods of data capture, storage and transmission.
To me it seems obvious that if people could capture information into a system simply by verbally describing or showing an intelligent assistant, with confidence that it will be interpreted into a commonly-accessible form, then this could dramatically reduce errors and at the same time hugely increase the richness of the information captured. To do this doesn’t even mean having to replace all of your existing systems; after all, it should be possible without huge difficulty to integrate Copilots and other AI tools into existing UIs.
However, it does raise the question: if you could replace your various UIs with a common, intelligent conversational interface, why wouldn’t you then insist on your organisation having a single source of trusted information? The case for tolerating disparate and/or legacy systems will be eroded further. I think we can now finally see how the dream of a ‘single version of the truth’ can be a reality, and the future is bright for platforms such as Esuasive that actually deliver this.
Aidan Dunphy is the chief product officer at Esuasive.
