One thing is clear about Informatica, six months after it was acquired by Salesforce for $8 billion: The company is a mash-up of cultures learning to work together.
Right now, that means the Informatica team is learning to do things the Salesforce way.
The integration does go both ways — but it's messy. While legacy Informatica leaders have stepped into key slots on the Salesforce data team, many veterans chose to exit when the deal was complete, rather than stick it out with the new regime. Meantime, the new GM of Informatica — Savinay Berry, the former chief product and technology officer at OpenText — had been on the job a scant 72 hours at the start of the company's flagship event, Informatica World.
Post-acquisition turmoil aside, based on what I heard at Informatica World, CIOs will find an Informatica that is actively rebuilding its data platform for an agentic, headless world, while deeply integrating it with the existing Salesforce data stack.
Related:How Anthropic is reordering SaaS — and where CIOs go next
Like others in this marketplace, Informatica is finding that enterprises need mature data processes to participate in the agentic age. This means their AI investments and deep integration are doing great things for the 32% of enterprises that have industrialized their data processes (as documented by polling we did at Dresner Research). And it also means that reaching the 68% whose data is in silos and fortified with bandages will require applying AI to accelerate data maturity.
The good news? There were glimmers at the event of where the data industry needs to go to help the 68%.
This involves using AI to dramatically accelerate the journey from data immaturity to maturity — something that historically has taken years.
Attendees were introduced to a data quality agent that can fix messy data issues using simple instructions written in a standard product requirement document.
Additionally, Informatica showed a configuration agent that will be able to automate most of the manual setup for master data management (MDM) within six months. That's big because creating a single view of customers and suppliers is what autonomous AI agents need to get the context to succeed.
There is also a data governance agent that allows data stewards who typically do not understand data structures to manage the data governance lifecycle with text prompts.
Each agent represents a major advancement over the state of the art from a few years ago and things that this writer hasn't yet seen from Informatica competitors.
Related:Salesforce is disrupting itself -- CIOs can't afford to look away
In a discussion with Informatica customer Theodora Bakker, vice president of data and AI at Hearst, she shared that she has implemented MDM nine times at different organizations.
She said the first time she implemented MDM, it took over nine months. The last time took six weeks with a requirements document in hand.
Informatica took a hard look at the issues that slowed MDM adoption around 2018 and took the hard step of completely rewriting code to make implementations easier. (This helps explain the dramatic reduction in implementation time Bakker experienced.)
Bakker and I agreed, however, that MDM and the entire data stack need to go even faster.
Bakker said she wants MDM systems to work from human-created requirements and rules, then suggest improvements and show the downstream implications for company data.
Rahul Auradkar, president and general manager of data foundations at Salesforce, is a smart, data-savvy leader who understands how Informatica could fit into the Salesforce portfolio in ways that previous acquisitions — including Tableau — did not.
Related:Is SaaS dead -- or just becoming AI?
During his opening remarks at Informatica World, he was candid about what excited him about Informatica.
On stage with Informatica leaders, Auradkar revealed that Salesforce was already conducting due diligence during last year's Informatica World. But he said he was less interested in Informatica's historical market position. It was its forward view that excited him.
"It was their roadmap and the notion of ready-set AI," he said, describing what he sees as a practical path forward for enterprises. "Let's face it: Data management has been hard."
Auradkar also acknowledged the broader uncertainty surrounding the future of enterprise applications. Introducing Berry, he said Informatica's new GM had spoken candidly during the interview process – including with Salesforce CEO Marc Benioff – about the future of the software business.
According to Auradkar, Berry's message was straightforward: No one can be fully certain where enterprise applications are headed. That message has clearly not been lost on financial markets or CIOs.
When I spoke with Auradkar, he did not dismiss the possibility that applications could increasingly become transactional layers sitting on top of data and AI systems.
"In 24 months, LLMs have come a very long way," he said. I do not think anyone would have predicted where they are now, and nobody knows where the software industry is going."
For CIOs, the implication is significant. Approximately one-third of enterprise IT budgets remain tied up in applications, even as the long-term interaction model for software appears to be in transition.
Putting it simply, Auradkar said, "No one is sure how people will interact with software." But he added, "There will always be a data and AI layer. Everything on top of that is going to change."
That idea helps explain Salesforce's current focus on owning the data and orchestration layers beneath enterprise workflows. Auradkar said he believes Salesforce's long-term competitive advantage will come from data tied to workflow dynamics rather than from applications alone.
He argued that fragmentation across systems and data is worsening, and said Informatica's relatively new headless platform — roughly 90 days old — could help enterprises "defragment" their data.
For CIOs, the message is that Salesforce increasingly sees competitive advantage coming from the quality of data, workflow orchestration and contextual intelligence beneath enterprise applications — not just from the applications themselves.
Auradkar — like others within the marketplace — said he sees context as the connective tissue that powers agents, and it requires more than MDM.
When I asked how Salesforce enables context versus simply creating a single customer view, Auradkar said it comes through Salesforce's data foundation, where multidomain relationships are established for agents.
This enables agents to act with more context and fewer hallucinations. Imagine an agent evaluating expense reports. Instead of rigidly applying a standard company expense policy, context allows that agent to notice that a prospect identified in the report is an executive, the company is slated to close this quarter, and the dinner is tied to a multimillion-dollar deal — so the agent approves a $500 dollar bottle of wine.
Auradkar then shifted to the notion of headless. It was interesting to hear the concept applied to data, because it suggests data agents themselves should be directly addressable.
According to Auradkar, this approach will make Informatica both more open and competitive. "It means that you can address us from any point."
He added that he believes it makes collaboration with partners easier, while giving customers flexibility and choice.
In terms of reaching the 68% that are data immature, Auradkar said automation is key. "It will allow Informatica to plug and play in Salesforce."
Auradkar argued the industry must accelerate how quickly people can engage with data, particularly if enterprises expect Agentforce to succeed at scale. With a headless approach, data consumers could ask whether information is correct, without needing to understand its underlying data structures or governance models.
When I asked Auradkar why self-service BI wasn't discussed more prominently, he noted that Tableau had already explored many of those themes. But he agreed the speed of information remains critical — especially in areas like marketing.
At the same time, he argued that self-service needs to include observability as well: Where did the data come from? Are there issues with it? How secure is this data?
It is clear that the software industry is undergoing significant change.
While there has been considerable debate over whether Salesforce's purchase of Informatica will prove accretive given the company's acquisition history, Salesforce appears clear-headed about the value it aims to deliver.
At the same time, Informatica is rebuilding itself for an era in which CIOs no longer have years to industrialize their data environments before deploying AI agents at scale.
For CIOs, the implication is difficult to ignore: If enterprise value is shifting beneath the application layer, then data maturity, governance and context — once relegated to the nerds — become strategic infrastructure, not back-office plumbing.
CIO Analyst and Tech Journalist
Myles Suer is a CIO analyst and tech journalist. Recognized by Leadtail as a top CIO influencer, he is the former leader of #CIOChat, a global community connecting CIOs and senior technology leaders. His insights have been featured in publications such as CMSWire, CIO.com, VKTR, and Cutter Business Technology Journal. Suer frequently reviews books on AI, technology, and business strategy from leading publishers, including Harvard Business Review Press, MIT Press, and Columbia University Press. Additionally, he serves as research director at Dresner Advisory Services.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。