By Jeremy Burton, CEO of Observe

2024 was the year of observability, with enterprises understanding that they needed to think beyond the traditional framework and adopt a holistic approach to their observability solutions.  Now, headed into 2025, Observe CEO Jeremy Burton shares his top predictions on what enterprise leaders can expect in the coming year.

Read on for seven predictions on agentic AI momentum, innovation in observability, "on the fly-UI," and more:

End Of The Model Wars - The past two years have seen big players and some well-funded startups compete to bring ever-larger models to market. But funding is a finite resource. When asked earlier this year about the cost of training foundation models, OpenAI CEO Sam Altman admitted it is more than $100 million. Such costs aren't sustainable for anyone other than the hyperscalers and infrastructure giants, particularly since there is no proven way to monetize them. The end of the model wars will ultimately be a good thing. Less choice means third-party ecosystem players will spend fewer dollars placing multiple bets and can concentrate their investments on fewer alternatives.

The Ai Investment Frenzy Comes Down To Earth - This was a banner year for AI investment, with 35% of startup dollars going to AI companies compared to 15% in 2021. The party won't end in 2025, but the cover charge will get a lot higher. With less funding being thrown at foundation models, more will target use cases that yield demonstrable value on top of those models that drive rapid growth in annual recurring revenue. Venture capitalists are already becoming more demanding; the median time lapse between Series A and Series B rounds in 2024 was 28 months, the longest in over a decade. While we aren't looking at a dot-com bubble-like collapse just yet, we can assume that business models with solid unit economics will get the lion's share of venture dollars next year.

The Rise Of AI Agents In The Enterprise - Much has been made of AI's potential game-changing impact to consumer search, but perhaps some of the biggest short-term opportunities in AI may lie in addressing the woeful inefficiency of many day-to-day office tasks. ERP and CRM systems may be the backbone of the enterprise, but most of the work still gets done in emails and spreadsheets.

Agentic AI promises to do what robotic process automation didn't: scour through that mess, eliminate ROT (redundant, obsolete and trivial) data and make what's left part of actionable workflows. That isn't sexy stuff, but it will give birth to many more successful businesses over the next five years than training ever larger LLMs.

Say Hi To The "On-The-Fly" UI - Enterprise software users have long had to deal with the complexities of a feature-rich but confusing user interface.  Even the revered "Business Dashboard" is often a dizzying array of numbers and charts. A new breed of AI-native interfaces will drive a completely new interaction model between human and software.  User interface will be constructed "on-demand", in response to the users inputs - showing just enough information, but no more. Every user will be comfortable with AI-native interfaces immediately, gone are the days of training courses and mountains of documentation.

Exploding Costs Drive The Great Observability Shakeout - Incumbent vendors have built their offerings on a mish-mash of proprietary agents, proprietary databases and query languages. The result? As data volumes have sky-rocketed so too have their bills and - worse - their customers are locked in. 

OpenTelemetry, Object Stores (such as AWS S3) and Apache Iceberg have provided a new breed of vendors with a standard, low cost, way to collect and store data. These products - Observe among them - are based on a completely new architecture built on open standards and priced in a way that doesn't result in the usual overage bill each month. Legacy vendors will inevitably try to make it difficult for customers to switch, but those barriers will fall. Customers should plan now for the coming migration opportunities.

Bloom Comes Off The LLM Rose - Despite the excitement around AI, many will begin to realize  the limitations of large language models. While impressive at summarizing, translating and regurgitating well-known information, these models are clearly not the foundation for artificial general intelligence (AGI). In fact, LLMs have arguably siphoned investment dollars away from approaches that may have had a greater chance of success.

It's not that LLMs are useless - quite the opposite - they are a better way for humans to interact with all kinds of software, devices and systems and will fundamentally change many industries. But the AI super-intelligence - the kind that interacts and reasons about its knowledge, surroundings and people in the same way humans - that's still many, many years away.

Knowledge Work Will Be Disrupted - If your job involves digging through mounds of information and synthesizing it into concise summaries, or maybe writing in a way that only professionals of a certain level in a particular industry would understand - then now is a good time to look into another line of work. LLMs are clearly better than humans at distilling information - they can absorb more, retain more and synthesize more than any human. They do not get bamboozled by technical terms, legalese or foreign languages; they can read and generate images, voice as well as text; and they never get tired so they can work 24x7x365. 

That potentially makes them a threat to everyone from  law clerks, management consultants, translators right the way through to interior designers and architects. However, one human trait that is very difficult to ‘generate' is creativity and original thought - there will always be a need for the spark of innovation, a thesis and logical reasoning.

As observability steps firmly into its own consolidated and expansive market in 2025, I am excited to see how the industry shakes out.

##

ABOUT THE AUTHOR

Jeremy Burton is the chief executive officer of Observe, Inc. Prior to Observe, Jeremy was Executive Vice President, Marketing & Corporate Development of Dell Technologies, and served as President of Products, overseeing EMC's $15 billion business. Jeremy joined EMC from Serena Software, where he was President and CEO. Previous to Serena, he led Symantec's Enterprise Security product line as Group President of Security and Data Management. Jeremy also served as Veritas' Executive Vice President of Data Management Group and Chief Marketing Officer. Earlier in his career, he spent nearly a decade at Oracle as Senior Vice President of Product and Services Marketing. Jeremy has been a member of the board of directors at Snowflake since 2015, and maintains a seat on the advisory board at McLaren Racing.