1998. The year I fell in love with a website!

Walking into the lab of my university computer science department, I sat down at a workstation and typed in for the very first time a URL that would become a beloved companion. Google.com.

It was immediately apparent that this was something different, something intriguing. The acres of white space and lonely search box. They stood there defiantly against the typical busy link lists of Yahoo!, Excite and AltaVista, which were the norm at that time (here for those not old enough to remember!).

Then the magic happened. I typed in a keyword and pressed search, and in what seemed like an incredibly brief moment I was presented with exactly the information I wanted. No waiting, no pop-ups, no irrelevant results.

It’s hard to look back on that experience now, 25 years later and convey how totally different it was to the status quo and how obvious it was that this website was going to change the Internet. Now 25 years later, a whole quarter of a century, it’s happened again with ChatGPT.

Following and analyzing commercial building technology as I have for 18 years, you get used to tech marketing hyperbole. Artificial Intelligence (AI) being the latest addition to my buzzword bingo card.

So, when I paid my $20 for ChatGPT pro subscription I was fully prepared to be underwhelmed.

That initial urge to play contrarian was misplaced, and I had to relent to the inevitable consensus conclusion… THIS. WILL. CHANGE. EVERYTHING.

In my little world of smart building market analysis, search and retrieval of information is a critical element of what we do at Memoori. Searching for information takes different forms.

The ‘hard yards’ of interviewing industry execs, conducting product briefing calls, even conversations on trade show floors and conferences, to of course (and to a large extent) searching the Internet across financial reports and many other types of documents.

The initial and inevitable question I then asked myself, after starting to use Large Language Models (LLM) like ChatGPT was “Are we witnessing a new era in search and retrieval of information?”

The lazy (and correct!) answer is yes. And don’t take my word for it, just look at the incredibly wide-eyed reaction from Google as they saw OpenAI coming to eat their breakfast, lunch and dinner!

The slightly less lazy answer is yes… but it’s more nuanced. And if you indulge me, I want to spend the rest of this post looking at that nuance and its impact on my industry, market analysis and competitive intelligence. The ‘What is’ vs the ‘What could be’.

Those different sources of information I previously described need to be collated, analyzed, and understood. This is where Memoori, as analysts, add our value. Then we output that knowledge in reports, articles, spreadsheets and presentations. This method of output has changed little in my 18 years, except for a lot less paper now!

We recognize that the traditional model of sifting through extensive reports for specific insights is becoming increasingly impractical. Today’s executives face time constraints and prefer concise, directly relevant information over navigating text-heavy documents. Acknowledging this shift, we are adapting our delivery methods to align with the changing preferences and expectations of our clientele.

And while we have written extensively about the impact of AI on smart building technology, we have also conveniently ignored the impact of AI on our own business. Until now.

We’ve developed AIM – an AI chatbot service designed to simplify the way our clients access and engage with smart building market analysis.

In the development of this solution, we have taken our complete body of work, including years of market reports, research notes and presentations and split them into text chunks. Vector embeddings are then created which are numerical representations of these text chunks and are stored in a vector database.

When a user then inputs a question (or task, like ‘write a report…’), we create another vector for that question and compare it against our database, returning the text chunks that are ‘closest’ (vectors that are most similar) to the question.

These most relevant text chunks, along with the question are sent to an LLM via an API, which is where the AI magic happens and the LLM formulates a coherent answer to the question.

Much like ChatGPT, AIM is a Chatbot. Once your question is answered you can ask follow-up questions or steer the conversation in the direction you want to go.

In summary, AIM represents a totally new way to access our research analysis and receive reasoned answers to questions or tasks, rather than thumbing through text-heavy documents. Whatever kind of smart building query you may have, AIM has you covered:

✅ Your manager wants a report on a competitor’s strategy?

✅ You want to quickly understand the vendor landscape of a particular technology?

✅ You want to understand the market dynamics affecting a particular vertical or geography?

Like any IT system, there are implementation challenges, anyone who has used ChatGPT will know that it can hallucinate. But with AIM it is less likely to do that when you ask it smart building technology questions because it has access to our sector-specific information.

The 1998 version of Google search is not the 2023 version of Google search. The rapid pace of improvement in AI will filter into AIM. The addition of complex data analysis tasks also won’t be too far away.

Our corpus of documents is of course not infinite and there are questions it won’t be able to answer. Integral to our system is the continued production of high-quality analysis and understanding of the smart building industry. Without feeding this analysis into our database it cannot produce the answers our clients would expect. Garbage in, garbage out as they say.

To that end, AIM will be provided to all Memoori premium subscriber companies in 2024, included in the subscription fee. Those that support our ongoing research and analysis will be the ones to benefit from the enhanced service of AIM.

We are excited to take this first step in our journey towards an AI augmented research service, and its potential to take market analysis and competitive intelligence to a whole new level.

I would love to hear your comments and feedback on what we have done so far and where we should take it in the future. Want to know more? We are holding a LinkedIN live stream where you will be able to get more details and ask questions.

Thanks for reading!

This article was written by James McHale, CEO at Memoori.