BIOps Strategy: Preventing BI Sprawl in the Age of AI Analytics
At COLLIDE 2023 in Atlanta, Datalogz CEO Logan Havern said AI copilots in BI tools will make dashboards easier to build and sprawl worse. He outlined a BIOps strategy to find unused, duplicate and risky reports and cut cost, citing his company's client results.
By The AI Enterprise Magazine Team3 min read
Key takeaways
- AI will speed up dashboard creation. Havern said Power BI, Tableau and Qlik were adding AI assistants, making insights easier to build.
- Sprawl follows. More reports mean more duplication, cost, stale data and accuracy risk.
- Governance often stops at the warehouse. Havern said BI is where users touch data most, yet it is often unmonitored.
- Unused content is costly. He cited clients with 70% to 80% of reports unused for six months.
- BIOps is the proposed fix. Use BI metadata and logs to flag and fix duplicates, anomalies and sensitive-data exposure.
At COLLIDE Data Conference 2023 in Atlanta, Datalogz founder and CEO Logan Havern warned that AI assistants in business intelligence tools would make it far easier to create reports, and far easier for BI environments to sprawl. Speaking in 2023, he said the answer is a "BIOps" strategy that monitors cost, accuracy, security and usage across BI platforms. Datalogz sells an automated BIOps platform, and the figures below are Havern's claims.
What was changing in BI tools in 2023?
Havern said almost all BI tools would soon add AI integration that gives business users instant answers. He named Power BI adding Copilot, Tableau adding Tableau GPT within the Einstein suite, and Qlik adding its own AI tool. "Imagine ChatGPT in your BI environment," he said.
That would make companies more data literate, he said, because the skills needed to build insights get much easier. But more reports and dashboards also bring recurring questions: whether they are compliant, duplicated, built on the right data sets and actually used.
What goes wrong when BI environments sprawl?
Havern drew on his earlier work managing BI infrastructure at a large airline. He said the team saw nearly 70 new reports and dashboards each week. Costs rose as the company moved its Power BI premium capacity from P1 to P2 to P3 and then added a separate P1, while performance slowed and users complained about load times.
Accuracy was a bigger worry. One report had been duplicated 12 times, he said, and numbers were reported to the FAA with no checks on whether they were right.
He listed four risk areas: rising spend on BI and compute, security (such as a rogue consultant exporting reports), governance, and misrepresentation of data. On governance, he said BI environments typically run on about 40% stale data and static files such as CSVs joined outside the governance program. He also said 70% of enterprises have three or more BI tools.
What is a BIOps strategy?
Havern said BIOps should catch and fix problems, flag anomalies and save money, with the aim to "end the wild west of reporting." It sits where business users consume data, in tools such as Qlik, Power BI, Cognos, Mode, Looker, Tableau and ThoughtSpot.
BI tools already store metadata and server logs, he said, though they are buried in APIs and infrastructure. A good data practice can move that information into blob storage or an S3 bucket and turn it into insights.
What should teams look for?
Havern described what Datalogz flags once connected to BI environments:
- Unused assets. For large clients, he said, 70% to 80% of reports and dashboards had gone unused for six months, and he cited one report with almost $122,000 a month in Snowflake compute and six months of zero views.
- Duplicate data sets and reports. He said business users had brought in data set exports hundreds of times within one enterprise.
- Stale data. An old report that has not refreshed in three months could still inform a CEO's decision.
- Sensitive data. Most governance programs stop at the warehouse layer, he said, though BI is where users interact with data most.
- Duplicated metric logic. By comparing Power BI DAX measures and Tableau logic, he said, the platform builds "almost a reverse semantics layer" to show where metrics overlap. The business still decides what is right.
What return did Havern claim?
Havern said typical clients see around 30% lower BI-related costs. In one Fortune 100 consumer products example, he said the client spent nearly $16 million a year on BI and had 25 Power BI premium capacities. Through lifecycle management, deduplication, consolidation and archiving, he said, premium capacity costs fell by nearly $4.5 million.
He also cited engineering time saved, with large clients typically seeing about 22,000 BI-related issues, and security benefits. Most chief information security officers, he said, are not thinking about the BI layer even though it is the largest point of access for anyone in an enterprise.
Where to hear more on enterprise analytics
COLLIDE 2027, the largest AI conference in the Southern U.S. with more than 1,000 enterprise data and AI leaders, takes place Wednesday, October 13, 2027, at the Sandy Springs Performing Arts Center in Atlanta. See COLLIDE. AI Connect also runs 24 free live webinars on its 2027 schedule; see webinars.
Frequently asked questions
What is BIOps?
Havern described BIOps as a strategy of monitoring and managing BI environments so reports are accurate, refreshed, not duplicated, cost-effective and low risk.
Why will AI make BI sprawl worse?
Havern said AI copilots in tools like Power BI and Tableau make it easier for anyone to create reports, which increases duplication, cost and accuracy risk.
How much can BIOps save?
Havern said Datalogz clients typically see around 30% lower BI-related cost, with one Fortune 100 client cutting premium capacity cost by nearly $4.5 million. These are vendor claims.
Where does BI metadata come from?
He said BI tools already store metadata and server logs behind their APIs, which can be moved to blob storage or S3 and analyzed.
This article is based on remarks by Logan Havern, CEO and Founder, Datalogz at COLLIDE Data + AI Conference 2023 on October 3, 2023. More coverage: COLLIDE AI Conference.
Filed under: Observability & Data Quality, AI Governance & Trust, AI Strategy & ROI
