Thoughts on Artificial Intelligence in Data & Analytics
TL;DR - The adoption of Artificial Intelligence in Data & Analytics is inevitable but the process will not happen overnight
If you haven’t already subscribed, join 605 smart, curious members of the Data & Analytics community by subscribing here:
Please do consider sharing this newsletter with friends or colleagues who you think may be interested. Thank you!
When the venerable Data & Analytics sage David Worlock invited Asymmetrix on to his Outsell FutureScapes video blog to discuss how Artificial Intelligence is impacting Data & Analytics valuations, it took some convincing for us to take part.
Like any fast-growing area, it can be hard to distinguish in AI what is real and what is snake oil;
When it comes to valuing businesses, disentangling how one factor (beyond revenues or profitability) impacts on the ultimate financial value an investor or acquirer ascribes to a business is not always possible.
But David Worlock is an old friend (we interned at his Electronic Publishing Services business over 20 years ago) and a conversation with him is always thought provoking.
(I also highly recommend his memoir of a postwar Cotswold childhood: Facing up to Father.)
Asymmetrix prepared for the conversation last week by digging into the AI strategies of the larger Data & Analytics companies:
And then ploughed back through recent Data & Analytics transactions, looking for interesting deals that could support arguments one way or another.
We will share the video conversation with you in due course, but thought it would be useful to put some of our thoughts in writing immediately.
Overall thoughts
William Gibson was right - the future is already here - it's just not evenly distributed.
Not all AI is created equal. Working out which AI does what it claims and which does not is hard. (Expect to see a Data & Analytics company focused on exactly this in the not too distant future.)
Technology is always changing - the current state of AI today won’t be the current state in six months’ time.
Some tasks are more readily suited to AI than others. In general, the simpler the task the more likely it is that AI could currently be deployed.
All technological innovation creates opportunities for humans to move up a level in thought work. This has an impact on the work people in different professions do, and who businesses hire, and the educational system to prepare them for work.
AI adoption is inevitable just as the adoption of all of the last cycles of technology was inevitable: blockchain, mobile, cloud, social, web, the PC…
Can you start a Data & Analytics business in 2024 without using AI? How influential is an AI component in giving credibility to investors?
Of course you can - we see Data & Analytics businesses launching that are not AI-driven.
But does AI make a business more investable? Yes - without a doubt. Investors, even if they don’t fully understand AI, see the recent stock market gains of Nvidia and want a piece of the action. So “what’s your AI strategy?” has become the battle-cry of the investor.
And if an AI-driven solution actually does what it’s supposed to and provides meaningful impact on professional workflows, then it will generate tangible revenue benefits for the business that can successfully monetise it.
Conversely, you may find that not having AI at the heart of your business may mean that your business is discounted or ignored by potential investors. We have heard of one large debt fund which uses an AI risk scorecard to assess all of their investments. If your business model is at risk of disruption from AI, you may find it harder to raise funding.
Comment on the following three areas in which Data & Analytics companies are deploying AI.
In general, the harder the human intelligence task the harder it is to use AI to replace it. We recently saw someone suggest that any task you would give to an intern could potentially be performed by Chat GPT. So that’s a handy way to think about where we are currently.
Start-ups where AI poses a complete challenge to established ways of formulating solutions or supporting decision-making in a particular market sector.
This is the dream for Data & Analytics and SaaS providers alike - to be able to provide an AI version of a job function, whether this is an Analyst at Investment Bank, a Junior Partner at a Law Firm, or a Purchasing Manager at a supermarket chain.
Look at the last $100m Series C fundraising last week by Legal workflow solutions provider Harvey. Their raise from Google Ventures, OpenAI, Kleiner Perkins, Sequoia Capital, Elad Gil, and SV Angel, valued two-year old Harvey at $1.5bn.
How much of this is hope value? Plenty. But a workflow solution that can truly replace an expensive junior lawyer is a truly valuable commodity. Time will tell whether or not Harvey fulfils this promise.
Situations where AI is used in product development cycles to create scale change for uses of existing knowledge products by virtue of a radical increase in the speed, comprehensiveness or creativity of the solutions derived;
Doug Peterson, CEO of S&P Global, explained how AI is transformative for Product Development for Data & Analytics businesses at the Stifel 2024 Cross Sector Insight Conference:
If you take different divisions - now you can pull out that information… and start extracting it for new solutions.
Data & Analytics businesses are sitting on a mountain of information that could be productised to solve different workflow needs for their clients. Data extraction using AI provides a way to develop products on a much more rapid basis than was possible before.
Situations where the required investment in AI will have a revolutionary effect on the cost base and productivity of the knowledge company concerned.
This is probably the most obvious opportunity. As mentioned above - the simpler the task the easier it is to use AI to automate that work. And a lot of Data & Analytics relies on good old grunt work, much of which is currently outsourced to low cost jurisdictions.
Baer Pettit of MSCI outlined the transformational impact this is having for their business at the Barclays Americas Select Franchise Conference 2024 in May:
The area where for sure we are having the most success is behind the scenes in data management, operations, quality control and even things like in making our coders more efficient. So generally, it's more getting a better outcome in an existing category rather than transforming the category, right? It's making a certain category of data more accurate, making it easier to get that data, making it faster to get that data, having fewer humans involved in cleaning it, things of that kind. And I think that that's 100% what AI is great at.
How much of what we are seeing at the moment do you think is about getting credibility with investors - “powered by AI” etc? How long do you think it will take for real change to take place?
There’s no doubt that there is some AI-driven froth in the markets currently, and that not every business claiming to be an AI business really is what it says.
But it is inevitable that all successful businesses will embrace AI eventually, just as successful businesses have embraced the other previous tech cycles: mobile, cloud, social, web, and the PC.
That change will take place gradually until we turn around and it is omnipresent. But by that time another tech cycle will have emerged. And other, more advanced forms of AI will come along too and make today’s AI look like the Microsoft Paperclip.
What do you feel about the reported sales of data by contact companies to AI technology developers? Is this a one off process or can repeat revenues be created? Does the sale of data in this way make the vendor more or less investable?
The following is a gut response and a full answer would require proper academic research.
But…
Intellectual Property is valuable because it is hard to create. Once you have it you can do powerful things with it. Data & Analytics businesses prove this out in spades.
So why would you sell your IP for cents on the dollar of its real value, just to provide a tiny bump of cash for your quarterly earnings?
The newspaper industry have learned the hard way that providing aggregators and search engines with access to their content both devalues that content and enables them to be disintermediated. Data & Analytics businesses should be careful not to find themselves in the same position.
We look forward to sharing the video content with you in due course. But, for now, please: