Gartner Giveth Guidance on Data & Analytics, And We Taketh

(Piotr-Swat/Shutterstock)

Gartner arguably is the most respected IT analyst firm on the planet, so when its analysts and VPs share what they’re thinking, as they did during the company’s Data & Analytics Summit this week, it’s worth taking notice.

What moves the needle for enterprise–in the field of bit data and analytics, or any realm for that matter–isn’t necessarily what everyone is talking about. Hype permeates our society like never before, but billion-dollar-companies tend to play their cards close to the vest. Instead of jumping headfirst into the latest thing, they prefer due diligence.

With its close enterprise partnerships, Gartner tends to be the voice of rationality when it comes to IT investments. Its famous hype curve reflects the fact that new technologies often flame out before delivering the goods, while others take years to mature. It’s a meat-and-potatoes approach that doesn’t always yield big, bold headlines, but does gain the ear of the folks who wear the suits and control the purse strings.

So, with that said, what do the Gartner folks see happening in the world of data and analytics? What new technologies or techniques does it think companies should invest in? Are generative AI and AI agents legitimate advances, or will they flame out too? Gartner shared its views on these topics.

(Rymden/Shutterstock)

For starters, let’s look at Gartner VP Analyst Gareth Herschel’s list of the top nine trends in the data and analytics space:

1. Highly Consumable Data Products
2. Metadata Management Solutions
3. Multimodal Data Fabric
4. Synthetic Data
5. Agentic Analytics
6. AI Agents
7. Small Language Models
8. Composite AI
9. Decision Intelligence Platforms

The list includes some hype-driven tech here, namely agentic analytics, AI agents, and small language models. There is definitely potential in these areas, as we have written about in the pages of BigDATAwire (for instance, look at what Alation and Immuta are doing with agentic AI in the fields of data management and data governance, respectively).

But the rest of Schlegal’s list is fairly anodyne, from a hype perspective. Data products, metadata management, and data fabrics aren’t necessarily ends in their own rights, but rather foundational components that D&A teams would do well to establish before trying to build higher order analytics and AI products. The same can be said for composite AI and decision intelligence platforms, which are the opposites of the “Let’s ChatGPT everything” trend that has taken over some parts of the analytics and AI space in the past two years.

Every enterprise environment is different–and organizations in the scientific and technical computing arenas are dealing with different data and have different requirements. But there’s enough commonality across enterprises for a CTO at one company to see how another company’s success in building solid D&A foundations might translate into their own D&A success, which is part and parcel of the Gartner method.

Dealing with D&A Adversity

We’re all prone to the “shiny object syndrome,” and GenAI definitely is the latest shiny object to steal all our attention. (Which is ironic considering the GenAI boom can be traced back to a Google paper titled “Attention is All You Need.” Or maybe it’s not ironic at all. We’ll get back to you on that.)

Kurt Schlegel, Gartner VP Analyst

In any case, implementing AI and analytics isn’t easy, and how you respond to challenges says a lot about whether you will ultimately succeed or fail. Once again, Gartner VP Analyst Kurt Schlegel provided some sage advice that’s light on hype and heavy on substance.

Challenge No. 1: Establish trust: “Provide a heads-up of industry and technology trends to key stakeholders — focus on impact, not hype,” Schlegel says.

Challenge No. 2: Demonstrate benefits: “Tie data pain points and opportunities to organizational goals by pinpointing what is inhibiting data-driven decision making and determining its downstream impact on business outcomes,” he says.

Challenge No. 3: Establish a solutions-first approach: “A modern data and analytics strategy architecture fosters data quality and data governance as a source for real-time insights and actionable response across functions,” Schlegel continues.

Challenge No. 4: Focus on more than just the technology: “A solutions-first approach requires a deep understanding of the problem and what it’s causing. Once the problem is understood, identify or create a solution to address it. Technology changes quickly, so stay open to new possibilities,” he says.

Challenge No. 5: Determine responsibilities between business and IT: “Set up a hybrid multi-tiered organizational model and determine where to position the global hub and CDAO. Balance traditional and emerging roles and actively engage with domain roles,” Schlegel concludes.

GenAI and Agents

Gartner has a protective force field against hype, which generally shields its analysts from succumbing to the “Let’s ChatGPT everything!” trend in D&A today. But the folks at Gartner aren’t dumb, and they recognize that GenAI holds real potential to increase the efficiency of a range of D&A tasks.

The AI agents are lining up (IM Imagery/Shutterstock)

Large language models (LLMs) dominate the GenAI conversation, but the future may see a proliferation of small language models (SLM), according to Sumit Agarwal, a VP Analyst at Gartner.

“Since the introduction of the transformer architecture in 2017, the most significant advancements in natural language processing have been driven by scaling model sizes and training datasets from millions to trillions, resulting in exponential growth in capability,” Agarwal says, according to a Gartner press release.

However, that trend may not continue. Specifically, SLMs may provide advantages in on-prem or private cloud scenarios where private information is being handled. SLMs also hold advantages in the customizability of the model, which leads to better accuracy, robustness, and reliability, Agarwal says. Finally, enterprises can further boost their GenAI fortunes by embedding their “static organizational knowledge” directly into SLMs, which can reduce cost and boost efficiency, he says.

Agentic AI has emerged as the latest AI hotspot generating excitement in the data and analytics community, particularly as it relates to automating manual data management and governance tasks, as Alation and Immuta are doing. Ben Yan, a director analyst at Gartner, provided some insight on how enterprises can integrate AI agents into their environments.

Rita Sallam, Gartner Distinguished VP Analyst

Yan encourages companies to prepare for agentic AI by first identifying the applications where agents can make a big difference. “Prepare software engineering teams for disruptive practice where AI agents make sense,” he says, according to a press release.

He also suggests that enterprises double down on AI literacy, considering that the deployment of AI agents “implies a deeper understanding of composite AI techniques,” which leverage multiple AI techniques, such as traditional data science, machine learning, knowledge graphs, and optimization techniques. Finally, people should prepare for the next stage of AI agents by familiarizing themselves with “software simulation environments,” Yan says.

Turbo-charging the traditional analytic workflows is one area that GenAI could provide a productivity boost. Rita Sallam, a distinguished VP analyst at Gartner, shared her thoughts on the impact that GenAI will have on analytics.

For starters, GenAI will accelerate the pace of doing business, provide for a more connected ecosystem, and set the stage for continuous learning and improvement, according to Sallam. The challenges are using AI in a way that benefits the business while dealing with AI risks around expandability and ethics.

“Perceptive analytics uses LLM-powered reasoning and AI agents in order to achieve proactive, contextual, outcome-driven decision-making,” Sallam adds. “By 2027, augmented analytics capabilities will evolve into autonomous analytics platforms that fully manage and execute 20% of business processes.”

Related Items:

The Future of AI Agents is Event-Drive

Will GenAI Modernize Data Engineering?

Three Ways Data Products Empower Internal Users