Why Data Fitness Is The Foundation For AI Success
We're entering a new era of enterprise automation, one where intelligent agents can analyze, decide and act independently. But while the industry's imagination races forward, infrastructure still lags behind.
Ninety-three percent of enterprise IT leaders have implemented or plan to implement AI agents in the next two years, according to a recent survey by MuleSoft and Deloitte Digital cited by ZDNET.
However, the survey also found that 95% of the IT leaders said they are struggling to integrate data across systems. Meanwhile, Gartner recently noted that, through 2025, 'poor data quality will persist as one of the most frequently mentioned challenges prohibiting advanced analytics (e.g., AI) deployment.'
AI agents can only be as good as the data that feeds them. And most enterprise data environments are far from ready. They're fragmented. Opaque. Siloed by system, department, geography or format. Putting AI to work in those environments won't be transformative. That's automation built on sand.
Enterprises are at a crossroads. Many are in the initial stages of deploying next-gen agents to boost productivity, streamline decision making and lower costs. But without a trusted, accessible and well-governed data foundation, those ambitions rest on shaky ground.
AI agents don't just need data. They need data that is structured, contextualized, traceable and aligned to the organization's goals.
Organizations must overcome a false sense of readiness.
Think about what AI agents do: They generate signals, draw inferences and act.
If they're trained on or use inconsistent or incomplete inputs, the decisions they make will reflect those flaws. While agents may move faster than humans, if they're working off half-truths or hidden assumptions, the consequences can quickly multiply.
Thankfully, this isn't a technology gap. It's a data-readiness gap. And it needs executive-level attention.
Over the past 20 years, my company has worked with some of the most data-intensive organizations in the world, from global banks and national defense systems to airlines, telecoms and critical infrastructure providers. Across these environments, four themes consistently separate those who scale well from those who stall and will be essential to an agentic world:
1. Build a unified, queryable data catalog.
Most enterprises don't have a clear inventory of what data they have, where it lives or how it relates. A living, searchable catalog makes it possible for both humans and machines to understand and access what is available and use it responsibly.
Organizations need to catalog both structured data and the rapidly growing pool of unstructured data, which makes up 80% to 90% of data today. Unstructured data is everything from videos to sales presentations to emails to social media posts that provide context for decisions.
While this data will still likely live in different silos, a catalog is foundational to creating the data products that AI agents need to get a full picture of the business challenge they are addressing.
2. Operationalize data governance.
Governance isn't about limiting access; it's about enabling trusted use.
Yes, AI agents must know what data is relevant. But agents also must know how it was processed and what rules apply to its use. Think lineage, version control and explainability as baseline requirements. Every agent needs to know where the data came from, when it was changed, if it was changed, by whom and for what reason.
For example, data may be scrubbed so voraciously that a person's middle initial gets removed in one dataset but not others. That can confuse an AI agent when, for example, gathering intelligence on whether a certain loan product would be good for a particular person.
3. Apply intelligent access controls.
The data democratization that will drive agents doesn't mean letting every AI workload touch every dataset. It means giving the right agents access to the right data, under the right guardrails, for the right purposes.
This requires policies that adapt to roles, risk levels and regulations.
4. Design pipelines for business relevance and scale.
A smart pipeline isn't just fast; it's aligned. Different agents will need different datasets, formats, and levels of latency depending on the task. For instance, AI agents working on procurement will want and need access to different datasets than AI agents working on marketing.
Once agents are working, monitor their success—or lack thereof—and make necessary changes in the data, the data pipelines and the access and control policies. Build data flows that can evolve with business needs and AI capabilities.
Readiness is an ongoing practice, not a one-time project.
Being 'ready for AI' isn't a one-time certification. It's a commitment to continuously refining how your organization collects, manages and mobilizes information.
The companies that win with AI won't be the ones who adopt the most tools. They'll be the ones who do the most with what they already know and ensure their intelligence systems are built not just to automate, but to align.
Make space for the agents of change, but first, build the conditions they need to succeed. AI is moving. Fast. The quality of your data will determine whether you're leading or watching from behind.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
CNN
3 minutes ago
- CNN
Trump administration targets Harvard's patents
The Trump administration is opening a new front in its battle with Harvard University on the issue of patents, marking a new escalation with the elite school that could result in the loss of additional federal funding or intellectual property rights. In a Friday letter to Harvard President Alan Garber obtained by CNN, Commerce Secretary Howard Lutnick writes that the Trump administration believes the university is 'in breach of the statutory, regulatory, and contractual requirements tied to Harvard's federal funded research programs and intellectual property arising therefrom, including patents.' The Commerce Department, Lutnick said, is issuing an 'immediate comprehensive review' of Harvard's federally funded research programs. The secretary said the administration was also initiating the 'march-in' process under a law called the Bayh-Dole Act that lets universities patent research and inventions. That means that if Harvard has failed to disclose or patent its inventions, the federal government could take ownership of the patents or grant third-party licenses. The letter, first reported by Reuters, marks the latest action by the Trump administration to exert pressure on the school. The Trump administration has frozen billions of dollars in federal funding for research and has targeted the school's ability to host international students. Harvard and the Trump administration are currently embroiled in a pair of lawsuits. Still, officials remain optimistic about the prospects of a deal with Harvard to restore funding to the school and drop lawsuits and investigations. The Trump administration has recently struck multimillion-dollar agreements with Columbia and Brown universities. 'While there's a lawsuit pending with Harvard, and I'm sure that lawsuit will play out, I do hope that Harvard will continue to come to the table with negotiations. Those talks are continuing, and we'd like to have a resolution there, outside of the courts,' Education Secretary Linda McMahon said in a phone interview with CNN last month. Harvard has sent some signals it is willing to work with the Trump administration, including last month when The Harvard Crimson reported that websites for Harvard College centers serving minority and LGBTQ students and women disappeared. The White House welcomed that development, viewing it as a goodwill gesture that one official described as 'good news.' McMahon last month also pointed to the departure of the heads of the university's Middle Eastern Studies center as a positive step.
Forbes
3 minutes ago
- Forbes
Obama's Flirtation With Supply-Side Economics
In his first term as president, Barack Obama extended the reduction in the top rate of the income tax to 35 percent through 2012, two years past the 2010 expiration date that his predecessor, President George W. Bush, had set. Obama presided over the lowest estate tax rate since Herbert Hoover's time, 35 percent in 2010 and 2011. (In one year, 2010, the estate tax rate was zero for those who elected to take it.) And Obama temporarily cut the payroll tax rate by about fifteen percent. In our new book Free Money: Bitcoin and the American Monetary Tradition, we ask why gold peaked after a phenomenal rise, as Obama got going, and why Bitcoin, though founded in 2009, took into Obama's second term to sport extreme price appreciation. We ask why the Great Recession bottomed early in 2009 and never came back (though the recovery was slow). A big reason for these things is that Obama was coquettish, 2009-12, toward supply-side economics. The marginal rate of the income tax, the top estate tax rate, and the payroll tax are three classic targets of supply-side economic policy. These tax rates are to be cut as the top priorities of supply-side economics. The theory is that each of these tax rates distinctly discourages the production and the seizing of initiative in the economy; therefore, cutting them enhances economic activity to an uncommon degree. The primus inter pares of supply-side economics is the marginal rate of the income tax. In a graduated tax system, the marginal rate is that which hits only earners of highest income. Cutting this rate encourages economic activity in two distinct ways. First, a cut in the top rate is the most powerful among all possible rate cuts in a graduated scale, on a simple percentage basis. A cut of 4.6 points from 39.6 to 35 percent (that of the W. years), for example, increases marginal after-tax 'take-home' income from 60.1 cents to 65 cents on the dollar—an increase of 8.2 percent. In comparison, a cut in the bottom rate of 10 percent (that of the W. years) by 4.6 percentage points to 5.4 percent increases marginal take-home income from 90 to 94.6 cents on the dollar—an increase of 5.1 percent. Given progressive income taxes, equal rate cuts mean more at the top than at the bottom. Second, those who are subject to the highest graduated rates—the highest earners—by definition have the most ability and desire to avoid, legally, those rates. High earners do not even need the money. They can decline to earn, change the way they earn (taking advantage of lower rates elsewhere in the tax code), the timing, shelter the stuff, whatever. The highest earners are most adept when it comes to making money. They can slip the top rate because they have the savvy and inclination to do so, and because the tax code gives them ample opportunity to represent income beyond declaring it ordinary. (Forget about closing these loopholes without lowering rates—an inevitable lesson of tax history.) Obama maintained a cut in the marginal rate of the income tax through the entirety of his first term in office. Undoubtedly, this was a central component of this president's strategy to get re-elected. When the Republican opponent in the 2012 election, Mitt Romney, made his gaffe about 47 percent of the electorate's not having to pay any income tax, Obama must have smiled. Obama had ensured that by keeping the top earners' tax rate reduced, top earners paid an outsized share of income taxes. Low top tax rates, high top-earner tax revenue—he knew the verity would hold. Let us be clear: keeping top tax rates down got Obama to a second term. The estate tax is another classic supply-side target. Work and earn your whole life, have the government take it away: a major disincentive to acquire. A reduction in the estate tax prompts, once again, precisely those who are capable of succeeding greatly at enterprise to do just that. Lots of people succeeding at enterprise spells a good economy. Obama took the estate tax to zero. If one took the zero rate, heirs did not get the step-up basis in capital gains. If one did not take the zero, again Obama's rate (of 35 percent) was the lowest since 1932. Supply-side essence, from President Obama. Obama cut the employee portion of the social security tax. For decades, supply-siders have identified the social security tax as one of the best illustrations of the problem facing modern tax-heavy economies. Social security taxes, paid by employer and employee, are a 'wedge' that interposes itself at the place where employee and employer would normally meet to contract labor. Cutting the rate leads to greater employment, and greater returns to both parties, employer and employee. The Obama cut would have been more purely supply-side if it had included the employer side as well (and been permanent), but a rate cut is a rate cut. More people contracted to work because their take-home pay was greater because of the policy. And employers could settle at slightly lower wage rates because their employees were taking more home after-tax. Barack Obama giving a clinic on supply-side economics! One can say that tax cuts are Keynesian. It is true that every tax cut makes the beneficiary spend more than before. But the effect is the absolute least at the marginal rate, and the least in general when the cuts are in rates of a progressive tax system. The JFK tax cut of 1964 that reduced progressive tax rates remains an exemplar of Keynesianism—actually it doesn't, thanks to Kudlow and Domitrovic, JFK and the Reagan Revolution—because of misinterpretation. A cut in progressive rates disproportionately has supply-side, not demand-side effects. Obama maintained cuts in progressive tax rates. Obama did much of this without Republicans forcing his hand. The Tea Party sweep of 2010 brought in a new Congress in 2011, after Obama had settled on most of his accommodations of supply-side economics. Politicians in foxholes—which is to say facing re-election—may talk a Keynesian game (it soothes the chattering classes). But when they act, they take supply-side economics into their confidence.
CBS News
3 minutes ago
- CBS News
Cherry Creek North residents want Denver neighborhood removed from improvement district conversations
In July, CBS Colorado reported on the potential for a Cherry Creek General Improvement District, and received several emails from concerned Cherry Creek North residents saying they want their Denver neighborhood removed from the conversation. "Neither of these postcards say on here, we're going to tax you for this, anything that you want on this, you're going to be paying for," said Dana Busch as she showed mailers regarding the GID that she felt lacked important information about the project. Residents invited CBS Colorado over for a conversation. What was originally supposed to be with three people quickly turned into a dozen. "We didn't even publicize this, so I can only imagine how many people would have shown up had we publicized this," said Busch. The neighbors coming together with a common goal. "Our group is advocating to have our residential area pulled out," said Busch. When asked if the group knows if any of their neighbors in Cherry Creek North are in support of the potential GID, Busch responded "zero." Residents Wayne New and Greg Jenkins added that they haven't met anyone, either. As it stands, the proposed improvement district would include Cherry Creek Shopping Center, Cherry Creek North, Cherry Creek East and a triangle just south of the area. "Each area should have the right to decide whether they want to be part of this larger project or not," said Busch. "We do not need this, and yet, when you look at this map, we're making up about two thirds of the dollar of the residential cost." The group believes Cherry Creek North doesn't need or want the GID and shouldn't be looped in with other nearby neighborhoods. They even created a website called to help their cause. A greater concern they say, is that they could be footing the bill for projects outside their area. "We're just really concerned, not only that our residents don't have a strong voice about what's happening with their tax money, how is it really going to make an improvement that makes a difference for our quality of life?" said Wayne New, who previously represented the area as a Denver City Councilman. City Councilwoman Amanda Sawyer represents the area and hired a consultant to look into the concept. They surveyed neighbors and found 65% were not interested in exploring a GID for Cherry Creek. Despite those results, the plan is moving forward. "They want to spend more of our taxpayer money on something that we've already said no to," said Greg Jenkins. The next phase puts residents into working groups to brainstorm ideas about what a GID could entail. Nearly all dozen of the residents present for the conversation said they signed up to lead a working group, while only one said she heard back. "We've always had a voice in our neighborhood," said New. "All of a sudden, now an advisory board, almost like a bureaucratic organization, is being developed over all of us, and whether we'll have a voice again, as strongly we had in the past, we don't know. So, it's a big, big concern to all of us." When CBS Colorado spoke with Councilwoman Sawyer about the project, she was adamant that if residents ultimately don't want this GID, then they won't move forward with a petition or work to get it on the ballot. Four community meetings are planned in September to talk about the potential improvement district.
