Tech, Data & Information Services: Six Big Topics for 2024
As we head into 2024, here at Outsell we’re planning for big, exciting things. We’re working on several new product and service capabilities to be unveiled in the coming months — including enhancements to our industry segmentation to make our coverage even more actionable, as well as new research formats, an improved headline service, and AI-enabled product features (yes, we’re leveraging it too, not just talking about it!).
2023 In Review
2023 will mark a pivotal moment in the future annals of technology, data, and information services. It was the year we reset expectations of what it means to interact with machines, and how information can be consumed, processed, and generated. AI was everywhere, all at once, and became our meta-theme for Outsell’s Annual Information Industry Outlook for 2024. Below are Philipp Mueller’s six key takeaways from 2023 and why they matter for the year ahead.
Top Six Topics for 2024
1. The Jury is Out on AI ROI
Many of us have spent 2023 trialing, piloting, testing, prodding, and coaxing generative AI in various shapes and forms, trying to make it do things that are not just interesting but also economically valuable. Despite the deluge of AI feature announcements, our clients consistently tell us that they’ve had mixed results in their endeavors — and that the verdict on where generative AI can add the most business value is still undecided.
One clear pattern has emerged. Leaders are finding faster and easier success in using AI to cut costs vs. driving revenue growth. Existing processes are well understood, and their automations readily explored. New ways to create value for customers are, by definition, less well understood, and so require experimentation and risk-taking. This is expensive, especially when staring at low-risk opportunities to cut the expense line.
However, we also hear often that when the AI does work, it can be surprisingly expensive to operate. Operating costs for generative AI add up quickly, especially in production scenarios where every click by a user can incur a non-trivial cost. It’s prompting a lot of leaders to explore less costly techniques in high-volume applications — typically either smaller LLMs, or other techniques altogether.
The bottom line right now is that companies — especially those in technology, data & information services — are mostly using AI to capture efficiency wins, up to 40–60% in some functions.
As for the people whose tasks are being automated: some of them will possibly be redeployed. However, we expect that most of them will not. We expect that the promise of smaller, leaner, more automated organizations with higher free cash flow is too tempting to pass up for operators and investors alike.
In Outsell’s Opinion: In some markets, we will see leaders pull ahead of their competitors in both profitability and growth as a direct result of their AI deployment. Best practices are emerging, but they vary by market — contact us for more detail for your markets.
2. Large Language Models Have Hit a Knowledge Ceiling — And AI Firms Are Starting to License Content
We spent much of 2023 learning about the limitations of LLMs. We learned that they can be biased, and that they’re terrible at abstract and multi-step reasoning. We also learned that they are imperfect stores of information, and hallucinate when they can’t remember facts — and that we need to spoon feed them facts alongside our questions (a.k.a. retrieval augmentation, or RAG) if we want to be remotely confident that we’re not getting baloney back.
In the last couple of months there have been more and more stories about LLMs hitting the limits of their faculties, even as the technology scales up. It emerges that at a certain point, better hardware makes less and less of a difference to model capability — and that the quality of the training data becomes a better predictor of model performance. In a way this is obvious: LLMs can’t know anything they haven’t been taught, ergo the bounds of their knowledge are defined by the bounds of their training corpus.
Some IP owners have already entered content licensing deals with AI companies. Axel Springer and the Associated Press are well known examples, and there are more. While it might seem counterintuitive for AI companies to license news content (since it is publicly available and can arguably be used under fair use terms), the deals cover not just historic content, but also a subscription to new content. This is critical, since it gives model operators the ability to generate more up-to-date answers, and address hallucination issues by injecting more authoritative data.
Such licensing deals are tempting for IP owners. It can be easy money — but they also raise existential questions. Among a wide array of risks, two stand out: intermediation risk — especially for ad-supported revenue models — and IP leakage risk. Once content has been used to train a foundation model, it can be impossible to remove from subsequent model generations: the Genie is out of the bottle, and can’t be put back in. Licensing opportunities therefore require careful calibration. Our experience with these kinds of arrangements is that while they can be lucrative, they require deliberate and surgical choices about what to license, when, to whom, and under what terms — and an eyes-wide-open approach to what can wrong. Caveat venditor!
In Outsell’s Opinion: AI companies will license more content as they look to differentiate themselves. This will create new revenue streams for licensors — but also cannibalize revenue streams that get displaced by AI, and materially weaken licensors who are not careful. We are working with a number of clients on licensing topics, and the dynamics vary substantially by market — contact us to explore them in more detail.
3. After a Year of Uncertainty, Pressure on AI Firms To Clean Up Training Data Is Rising
On the topic of training data, there was quite a bit of hubbub through 2023 about various people suing AI companies because they see their IP rights infringed in some way. The New York Times is the latest, and likely not the last. We published our first AI Legal & Regulatory Tracker earlier in 2023. Most of these cases are still running (look forward to our next tracker end of January.)
The quick recap for anyone who spent 2023 under a rock is that most generative AI training corpora use copyrighted material under fair use policies and are therefore not infringing any rights (in the US at least). As regards AI outputs, the US Copyright Office has asserted that copyright requires “human authorship”, and the UK supreme court has ruled that AI cannot be an inventor of patents. This all sounds appropriate, but it also offers a more cynical interpretation: AI models trained on rights-protected content can be used to generate rights-free training sets for other models in a practice known as data laundering.
Things are changing in Europe. After many months of drafting and deliberation, the European AI Act was finally ratified in December. It was already in the works before ChatGPT showed up, and language to cover LLMs was hastily added late this year. The legislation now mentions General Purpose AI (GPAI) systems and requires all such systems to comply with EU copyright law and publish detailed summaries of content used to train them.
This is good news. While improved transparency on training data will not by itself change fair use in the US, at least knowing what ingredients are in the AI sausage is an essential step towards a more balanced approach to LLM training.
AI firms are also coming under pressure from their commercial customers to clean up their training data. Enterprise buyers are increasingly concerned about compliance risks in their AI supply chain and want guarantees from their vendors that their AI models do not expose them to any compliance risks. This commercial pressure is likely to be a stronger force in cleaning up training data compliance than regulation, though the EU AI Act’s transparency mandates will accelerate the pace at which buyers can exert pressure.
In Outsell’s Opinion: AI companies will need to come clean at some point about using copyrighted content — and some will establish training data compliance as a competitive advantage. Fair use framing is unlikely to change. IP owners have other options, but the implications again vary by market — contact us for more details.
4. Open Access Is Grinding Ahead — And It Hurts
Speaking of information rights, scholarly content providers are at a critical juncture in the face of impending open access (OA) mandates from a consortium of European research funders and the US Office of Science and Technology Policy, which governs access to federal research funds. Over 50% of journal articles are now Open Access. The industry-wide shift, grinding along in 2023 and set against a 2024/5 deadline, is marked by varied expectations of impact and levels of readiness, with some stakeholders ready to embrace the changes, and others actively disengaged in the hope that mandates will be watered down. As plans evolve to address current criteria, mandates will continue to shift — making attempts at compliance more challenging.
Key insights from Outsell and Delta Think’s 2023 study highlight a looming financial challenge, particularly for smaller and society publishers. These groups, unlike their larger counterparts, face significant risks due to changes in revenue models and lack of capital.
Providers are responding, but also bracing for impact. They’re diversifying services, streamlining operations, and exploring new revenue models in an attempt to mitigate the financial blow and maintain relevance. However, the journey towards compliance with OA mandates is fraught with concerns.
The transition to OA is already reshaping the scholarly publishing landscape. Strategic adaptability, advocacy, and a keen eye on evolving revenue models will be crucial for survival and success in this new era of scholarly communications.
In Outsell’s Opinion: Open Access isn’t going away; it’s a question of “when, not if”. Vendors will need to diversify, consolidate, or substantially reduce costs to address the diminished revenue streams. To review strategies and plans for revenue diversification, give us a call.
5. Pressure on Vendors to Innovate & Expand is Propping Up M&A
Against a backdrop of lower M&A volumes across all markets, deal volumes in the markets we cover were also down substantially in 2023. Our free Outsell Doorway headlines service covered 445 M&A deals through the year, down ~40% vs. 2022. Strategic buyers were more active, many of them under pressure to find new avenues of growth without diluting their earnings, and able to leverage their strong balance sheets.
Stand-out examples are Thomson Reuters acquiring Casetext, Wolters Kluwer acquiring Della AI, Publicis acquiring Yieldify, S&P acquiring ChartIQ and TruSight, and IBM acquiring StreamSets and WebMethods. Invariably, the strategic acquirers riff on common themes: addition of AI and analytic capabilities, expansion into workflow tools & platforms, and exposure to higher-growth use cases such as ESG — all of them expected to add value and new growth vectors to slower-growing cores, typically anchored in content.
Private Equity hasn’t been standing completely still — notable deals include Clearlake acquiring Alteryx, Silverlake and CPP acquiring Qualtrics, Francisco acquiring the Weather Company, Blackstone acquiring Cvent, and KKR acquiring S&P’s Engineering Solutions business, now trading as Accuris. Many of these are great businesses that will generate good returns for their investors.
In Outsell’s Opinion: Strategics with strong balance sheets may find some bargains in 2024 if the cost of capital doesn’t fall. There will be more competition for AI assets once there’s more evidence for who has working products and paying clients. Outsell will be covering more of these M&A landscapes and targets through a revised market segmentation — reach out if you’re interested to learn more.
6. ESG Is Morphing Into Compliance and Reporting Obligations
As corporations navigate the evolving landscape shaped by mandates such as the EU’s Corporate Sustainability Reporting Directive (CSRD) and the SEC’s Climate Change Disclosure rules, Environmental, Social, and Governance (ESG) factors are transitioning from optional, marketing-driven elements to compulsory reporting requirements. Formalized in 2023, these directives set the stage for phased reporting obligations over the coming years.
This regulatory shift is reshaping the roles of Finance, Risk, Compliance, and even Tax departments, placing them at the forefront of ESG management and reporting. Consequently, there’s a burgeoning demand for finance- and risk-led technology solutions encompassing data management and analytics, integrated reporting platforms, and ESG-specific compliance tools.
Companies like Munich Re are weaving ESG factors into their risk management frameworks, while collaborations, such as between Wolters Kluwer Enablon and Makersite, are focusing on ESG-centric product development like decarbonizing supply chains. Furthermore, the use of data and analytics for enhanced ESG reporting and compliance is gaining traction, highlighted by innovations like Verisk’s weather enhancements.
This transformation underscores a fundamental shift: ESG is evolving from a focus on social responsibility to becoming a vital, intrinsic component of GRC strategies and statutory reporting obligations. Organizations are overhauling their risk management infrastructures to embed ESG, recognizing its pivotal role in ensuring financial performance and resilience.
In Outsell’s Opinion: There’s still a lot of fragmentation in where and how ESG criteria get tracked, managed, and reported — and that’s unlikely to change. In Outsell’s opinion, ESG compliance is “a feature, not a product”, in the sense that ESG criteria are increasingly tracked as rubrics in existing enterprise tools and solutions. We’ve charted out this landscape — don’t miss where to play in this ecosystem.
Implications
The last year has been a strange one for the Technology, Data & Information Services industry. Traditional industry models that have depended on low interest rates, incremental innovation, long-established moats, and traditional notions of IP ownership are being tested. CEOs are finding themselves adapting to dynamics that don’t align neatly with established playbooks for growth.
While the prospect of lower interest rates in 2024 comes as a relief, end customers and investors will not loosen their purse strings overnight. Confidence in growth, per our 2023 CEO Sentiment Survey, remains subdued, and our quarterly Sales Benchmark shows sustained declines in quota achievement — a trend that began in 2021. Companies entered 2023 with high growth expectations and were forced to temper them as the year progressed.
With growth down and interest rates up, a lot of companies — especially PE- and VC-owned ones — shifted their attention to margin improvement as an easier way to create value for shareholders. We also see it reflected in the client inquiries we field: they are increasingly focused on ‘how-to’ operational questions, a trend that follows tighter economic times.
However, companies cannot cut their way to growth — and as AI pushes more value from content into workflow and automation, the market leaders of today will need to take risks and invest in new capabilities if they want to be the market leaders of tomorrow. And downturns are when the best companies invest for the rebound.
In addition to AI product and feature launches, 2024 will undoubtedly also offer more plot twists on AI, especially as the EU AI Act comes into effect and customers apply more pressure on training data compliance, which we believe will work to the benefit of IP owners.
Essential Actions
We offer some essential actions to take in light of this year’s developments.
✔ Use AI to Provide Answers
AI techniques to answer user questions about proprietary content sets are maturing: retrieval-augmented generation (RAG) is becoming the de-facto pattern for building “talk to data and documents” experiences. Embrace them.
✔ Use AI to Automate Tasks (Internally and for Customers)
Generative AI allows knowledge tasks to be automated, or processes to be reinvented altogether — both in customer workflows, and internally. This is where the next wave of opportunity will play out: in turning knowledge-driven answers into automated actions. Doing this successfully will require a good understanding of how current solutions fit into users’ workflows, and what business problems they are solving in the first place.
✔ Diversify Your Revenue Streams
As content consumption behaviors and user expectations evolve, many of yesterday’s business models start to look increasingly shaky, and might not be around tomorrow. It might be worth thinking about alternative paths.
✔ Protect and Monetize Your Intellectual Property
Licensing content to AI models can look like a quick way to benefit from the AI boom. However, it can also lead to intellectual property loss, cannibalization, and the enablement of new competitors using that content. It’s crucial for rightsholders to carefully evaluate these opportunities. Sharp contractual terms are essential, but insufficient on their own: embracing a “trust but verify” approach will be critical. We have a lot of expertise on this topic and can help evaluate the trade-offs.
✔ Capitalize on ESG
Given generational and regulatory shifts, corporations around the world will need to take ESG reporting more seriously in the coming years. Seize the opportunity — both in terms of products and services that help customers in their ESG endeavors, but also in terms of being seen as a leader by employees and customers.
✔ Leverage Expert Insights and Tailored Strategies
As we navigate the ever-evolving landscape of technology, data, and information services, Outsell remains at the forefront, offering in-depth analyses and tailored solutions. Whether it’s leveraging AI for operational efficiency, exploring licensing opportunities in AI, ensuring compliance with evolving IP and ESG standards, or strategizing for the open access transition in scholarly publishing, we have the expertise and experience to guide you.
Don’t miss the chance to gain a competitive edge in your market. Reach out to us for detailed insights, strategic advice, and to explore the implications of these trends for your specific market. Our team is ready to assist you in harnessing these developments to your advantage.
Contact us today for a consultation and stay ahead in 2024 and beyond.