AI-Ready Data: Veolia’s Playbook for Data Leaders

When you supply drinking water to 110 million people,¹ your data foundation leaves no room for error. Veolia provides water, waste, and energy services worldwide, with215,000 employees.¹

Lisa Longo, Data Governance Officer at Veolia, manages data at that scale with ruthless prioritization and a zero-ego approach to technology. She joined Veolia after a decade leading data governance at one of the nation's largest financial institutions.² She has also been a design partner for Alation,³ helping shape itsData Quality andCritical Data Manager⁴ products.

AtrevAlation Chicago 2026,⁵ Longo shared her playbook for enterprise-wide data adoption. She also explained why Veolia shelved an approved custom AI build in favor of Alation AIOS. Here's what other data leaders can take from her approach.

The evolving mantra: "All roads lead to AIOS"

For years, Longo's guidance was simple: if you need to know something about your data, go to thecatalog. "I used to say all roads lead to Alation," she said. "No matter where it is—legacy knowledge in people's heads, what you have on docs—all of that belongs in Alation."

With the launch of Alation AIOS in July 2026,⁶ her mantra changed. "Now I've changed that to say AIOS," Longo said, "because everybody's being asked, 'Where do I get this AI everyone's talking about?'"

The two mantras are connected. The institutional knowledge Longo has documented in the catalog is what makes AI trustworthy. As she put it, "That information is what's going to feed your AI." Gartner sees the same pattern at industry scale. It predicts that through 2026, organizations will abandon 60% of AI projects that aren't supported by AI-ready data.⁷

Start with what matters: Critical data and business pain

"Not all data is created equal," Longo said. "People want to get a catalog and ask, 'How many sources can I get in here?' And I say, 'What is wrong with you? Don't do that.'"

Instead, she focuses on the operational data behind the company's key performance metrics.² What goes into the catalog is always the business's call. She starts by asking each domain one question: "Tell me what sucks about your job." Domains with real pain and a real willingness to participate go first, and the rest wait on the backlog.

The takeaway for data leaders: Don't catalog everything. Prioritize the data behind your most important decisions, and start with the teams that are feeling the pain and ready to do the work.

Trust in action: Surfacing billing errors in a week

Longo's team was an early pilot customer for Alation's data quality capabilities.³ During the pilot, a business domain asked whether data quality checks could catch billing discrepancies before accounts reached collections.

Working side by side with Alation's data quality team, Longo's team spent a week defining quality rules where bad data is the cause of payments not processing. In the first week after deployment, the rules caught flawed records across the pipeline, allowing the domain to resolve them and accelerate revenue recovery.²

The takeaway for data leaders: Aim your data quality work at a problem with a dollar sign attached. Quick wins like this, Longo said, "get you recognition to push everything else you need to do. Money talks."

Make the business your team

Longo runs a deliberately lean,federated model. As she puts it, "My team is my business."

Each domain has an executive data owner, because executives make the decisions. It also has a chief data steward, a "unicorn" who blends business and technical skills, who decides what goes into the catalog and leadscuration. "The institutional knowledge about the data belongs in their heads," Longo said. Her job is to support them.

The model paid off when Veolia acquired a water management company, bringing in more uncataloged data.² Because ownership already sat with the business, the domains could harmonize data across ERP systems, including JD Edwards and SAP, and document legacy knowledge in Alation.²

This model has also won over the skeptics. At her previous employer, a senior employee announced on a call with a thousand people that he'd never use the catalog. But then Longo learned that the worst part of his day was answering the same questions by email, hundreds of times. So she made a deal: put the answers in Alation and point people there. Three months later, his email signature sent everyone to the catalog.

His resistance, she realized, came from fear of becoming replaceable. Instead, curation freed him up for even more valuable and rewarding work, and he retired as one of her biggest advocates. "It's not about replacing people," she said. "It's giving them the option and ability to do something more."

The takeaway for data leaders: Put ownership with business executives, empower the stewards who know the data, and make your experts the real heroes.

Winning buy-in: Carrots, sticks, and evangelists

When it comes to incentivization, Longo prefers carrots. One domain leader put the asset utilization metrics that senior leadership reviews into Alation and built a data product on top of them, so leaders can now ask their own questions.² Nobody had to mandate adoption.

When she needs a stick, she points to risk. In her previous position at a bank, she learned that 27% of its data professionals had left in a single year.² So she showed the board a slide of people walking out the door and told them, "You just lost 27% of your institutional knowledge." Support for data governance quickly followed.

Longo brings that same clarity to privacy teams: they protect the data, but the business ultimately owns it. Similarly, within Alation, policies are tied to the data they govern, and automated checks flag sensitive data so the domain owners can decide who sees what.² It'sagentic data governance that scales without adding headcount. With this new model, the company's privacy lead went from her toughest critic to a close ally.

Her rollout advice is to find the people who raise their hands, get quick wins, and let those people tell the story.

Build vs. buy: The no, the build, and the second look

About a year ago, Longo's Alation account team showed her an early version of Alation's data products. "I said, 'This isn't for me,'" she recalled. She could write DAX herself, but her business-side stewards needed something easier to use every day.

So her team planned its own build. "Everyone has ChatGPT, Claude, Copilot," she said. "But you have your technical debt, and you also have the employees you need to run that." She secured approval for "a pretty large sum." Then acquisition integration work put the project on hold while her team built "a foundation set for AI."

When the company was ready to accelerate, her Alation team asked her to take another look. Five minutes into the new demo, she stopped it to ask how soon she could show her Chief Data Officer. After seeing it, he called it a game changer.²

"No ego; we scrapped everything I was going to do," Longo said. "It costs us less money to use Alation versus our homegrown build."² Her experience matches early research. A preliminary 2025 MIT NANDA study found that AI tools bought from specialized vendors succeeded about twice as often as internal builds.⁸

The takeaway for data leaders: Count technical debt and staffing in the total cost of a custom, in-house build. Judge any tool by whether your business users can use it every day. And be willing to revisit a "no" when the product innovates.

Data products: Multiplying the impact of every data leader

Executive questions used to mean tickets, waiting, and follow-ups that restarted the cycle. Now domain leaders build trusteddata products for the questions executives ask most, such as asset utilization and win/loss ratio.² Executives then ask their follow-ups in plain language, like which region performed best or how one region compares with another. They getgoverned answers in real time.

"Instead of looking at that dashboard and saying, 'I have more questions,' just ask your questions," Longo said. After seeing a demo, one senior commercial executive messaged her: "How soon can I have chat with my data?"²

Today, each data product keeps answering executive questions without anyone building the next report. That's how a one-person governance office extends its influence across the company.

Longo's playbook for AI-ready data leadership

Longo's approach comes down to a few principles any data leader can borrow, whether you're standing up a catalog for the first time or preparing your organization for enterprise AI. Here are six lessons from her work at Veolia:

• Curate the data that matters. Govern the critical data behind your biggest decisions, and treat its quality as a prerequisite for AI.

Curate the data that matters. Govern the critical data behind your biggest decisions, and treat its quality as a prerequisite for AI.

• Let business pain and engagement set priorities. Start with the domains that are hurting and ready to participate.

Let business pain and engagement set priorities. Start with the domains that are hurting and ready to participate.

• Win with money. Early projects that make or save dollars fund everything else.

Win with money. Early projects that make or save dollars fund everything else.

• Make the business your team, and your experts the heroes. Address the fear of becoming replaceable head-on.

Make the business your team, and your experts the heroes. Address the fear of becoming replaceable head-on.

• Judge technology by business usability, and keep ego out of build-vs-buy. The right tool is one your business users will actually use, at a cost you can sustain.

Judge technology by business usability, and keep ego out of build-vs-buy. The right tool is one your business users will actually use, at a cost you can sustain.

• Use data products to multiply your impact. Answer executives' most common questions once, and let them ask the follow-ups themselves.

Use data products to multiply your impact. Answer executives' most common questions once, and let them ask the follow-ups themselves.

With zero ego, business-first priorities, and a strong foundation, Longo is positioning Veolia to scale AI on data its people trust.

Sources & notes

Every external claim on this page is independently verifiable. The public sources are listed here.

• Veolia supplied drinking water to 110 million people and had 215,000 employees in 2025, across its water, waste, and energy businesses. — Veolia, "Veolia in brief" ↗ https://www.veolia.com/en/veolia-group/veolia-in-brief

Veolia supplied drinking water to 110 million people and had 215,000 employees in 2025, across its water, waste, and energy businesses. — Veolia, "Veolia in brief" ↗ https://www.veolia.com/en/veolia-group/veolia-in-brief

• Customer-reported: Lisa Longo, remarks at revAlation Chicago, September 17, 2026, per the session transcript. This covers her background, the critical-data focus, the invoice result, the federated domain model, the acquisition and ERP harmonization work, the 27% attrition figure, policy automation, the build approval and cost comparison, and colleagues' reactions (paraphrased, with names withheld). [Pending Veolia approval. Longo to confirm her prior-employer background and choose the invoice wording.]

Customer-reported: Lisa Longo, remarks at revAlation Chicago, September 17, 2026, per the session transcript. This covers her background, the critical-data focus, the invoice result, the federated domain model, the acquisition and ERP harmonization work, the 27% attrition figure, policy automation, the build approval and cost comparison, and colleagues' reactions (paraphrased, with names withheld). [Pending Veolia approval. Longo to confirm her prior-employer background and choose the invoice wording.]

• Veolia's role as an Alation design partner and early Data Quality pilot customer. — Alation product team, internal record. [Confirm with the product owner before publishing. No public source exists.]

Veolia's role as an Alation design partner and early Data Quality pilot customer. — Alation product team, internal record. [Confirm with the product owner before publishing. No public source exists.]

• Alation Critical Data Manager, announced as Alation CDE Manager on November 12, 2025. — Alation ↗ https://www.alation.com/aios/critical-data-manager/

Alation Critical Data Manager, announced as Alation CDE Manager on November 12, 2025. — Alation ↗ https://www.alation.com/aios/critical-data-manager/

• revAlation 2026 Chicago took place September 17, 2026, with Veolia among presenting companies. — Alation press release, September 9, 2026 ↗ https://www.alation.com/news-and-press/leaders-from-ally-financial-bbc-electronic-arts-flutter-and-veolia-to-present-at-alations-revalation-2026-conference-series/

revAlation 2026 Chicago took place September 17, 2026, with Veolia among presenting companies. — Alation press release, September 9, 2026 ↗ https://www.alation.com/news-and-press/leaders-from-ally-financial-bbc-electronic-arts-flutter-and-veolia-to-present-at-alations-revalation-2026-conference-series/

• Alation announced AIOS on July 14, 2026. — Alation via GlobeNewswire ↗ https://www.globenewswire.com/news-release/2026/07/14/3326909/0/en/Alation-Launches-AIOS-All-New-Intelligence-Operating-System-for-Enterprise-AI.html

Alation announced AIOS on July 14, 2026. — Alation via GlobeNewswire ↗ https://www.globenewswire.com/news-release/2026/07/14/3326909/0/en/Alation-Launches-AIOS-All-New-Intelligence-Operating-System-for-Enterprise-AI.html

• Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data. — Gartner, Q&A with Roxane Edjlali, February 26, 2025 ↗ https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk

Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data. — Gartner, Q&A with Roxane Edjlali, February 26, 2025 ↗ https://www.gartner.com/en/newsroom/press-releases/2025-02-26-lack-of-ai-ready-data-puts-ai-projects-at-risk

• Provisional: In MIT NANDA's preliminary July 2025 report The GenAI Divide, AI tools purchased from specialized vendors succeeded about 67% of the time, versus about one-third as often for internal builds. Reported by Fortune, August 18, 2025. The primary report still needs locating, and the study's sample is small and self-reported. ↗ https://www.fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo

Provisional: In MIT NANDA's preliminary July 2025 report The GenAI Divide, AI tools purchased from specialized vendors succeeded about 67% of the time, versus about one-third as often for internal builds. Reported by Fortune, August 18, 2025. The primary report still needs locating, and the study's sample is small and self-reported. ↗ https://www.fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo

Analyst attributions & disclaimers

Gartner, "Lack of AI-Ready Data Puts AI Projects at Risk," Q&A with Roxane Edjlali, 26 February 2025.

Gartner does not endorse any vendor, product or service depicted in its research publications and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner's research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose. GARTNER and Magic Quadrant are registered trademarks and service marks of Gartner, Inc. and/or its affiliates in the U.S. and internationally and are used herein with permission. All rights reserved.