Deliberate decisions and enterprise AI value

AI progress is moving quickly. Expectations are rising, experimentation is spreading and organizations are under pressure to show meaningful returns. But the decisions that matter most are rarely about whether AI is possible. They are about where to commit, what to prioritize and how to move forward when the evidence is still incomplete.

That is the focus of Deliberate Decisions: How Grounded Leadership Is Advancing the Path to Enterprise AI Value, a report produced by DataIQ in partnership with Databricks. The report brings together perspectives from senior data and AI leaders who are making these choices in real organizations.

Why deliberate decisions matter now

The report draws on open peer discussions at DataIQ Guild 2026, an invite-only think tank for senior data and AI leaders. Across industries, leaders described a similar challenge: they are expected to maintain momentum while technology, operating models, workforce impact and business cases continue to evolve.

Early AI success can raise expectations faster than an organization can absorb. Pilots and proofs of concept can multiply before teams have the capacity to assess, govern and support them. Ownership can become unclear as AI crosses functions, while roles and workflows change faster than organizations can define them.

The answer is not to slow down until everything is certain. Nor is it to commit to every promising use case. The report points toward a more grounded form of leadership: making deliberate, evidence-based choices while preserving the ability to adapt.

Report snapshot:

• Two days of peer discussion at DataIQ Guild 2026 underpin the report.
• Five recurring pressures are shaping the current data and AI environment.
• Five leadership disciplines emerged to help leaders respond deliberately.
• Four stages describe the AI adoption journey: concerned, curious, comfortable and competent.

The five disciplines behind deliberate decisions

Five leadership disciplines emerged from the discussions at DataIQ Guild 2026:

1. Confidence grounded in reality: test AI progress against evidence, readiness and the realities of production.
2. Enterprise choices, not demand-led reactions: prioritize business outcomes over isolated enthusiasm or use-case volume.
3. Leadership through influence, not control: create clarity, ownership and proportionate governance across the enterprise.
4. Human judgment as a strategic capability: preserve the authority, context and skills people need to challenge AI outputs.
5. Stability without rigidity: anchor purpose, decision rights and standards while keeping roles and technology choices flexible.

Four foundations for confident, flexible AI leadership

To lead data and AI today is to navigate a constant tension between commitment and flexibility. Every investment, architecture and operating model decision requires leaders to balance confidence in what they know today with openness to what may change tomorrow.

The Databricks perspective in the report identifies four foundations for navigating that tension:

• Context grounds AI in trusted, governed data.
• Control gives leaders confidence that innovation can progress responsibly and securely.
• Cost keeps sustainable adoption aligned with ambition and efficiency.
• Choice preserves the ability to evolve as technologies, models and business priorities change.

These foundations are reflected in Databricks capabilities such as Lakebase, a governed transactional foundation for building and running data and AI applications and agents, and Genie, which helps people and AI agents access organizational context responsibly and efficiently.

Together, they reinforce the report’s central message: progress does not require waiting for certainty. It requires making informed decisions, knowing where to commit and preserving choice.

What will you find in the report?

How to distinguish progress from readiness

A successful pilot is not the same as an AI capability that is ready for production. The report explores the difference between early results and sustainable adoption, including the operational support, governance, skills, monitoring and cost model that become visible once AI enters a live workflow. As Thomas Gebetsroither, Data & Analytics Operations Lead, Axpo Trading & Sales, mentioned:

“The cost is not just the model or the tool. It is the security, the support, the adoption, the governance and the operating model around it. If those costs are invisible, the business case is incomplete.”

Leaders will find a practical way to test whether an AI initiative can remain useful, reliable and affordable beyond the conditions of a bounded experiment.

How to make enterprise choices

AI demand can come from the board, business functions, employees and vendors. But demand alone cannot determine the agenda. The report looks at how leaders start with a well-defined business problem, make trade-offs visible and prioritize work with enterprise-scale potential.

It also examines why AI investment may be better managed as a portfolio. Organizations can make a range of initial bets, increase investment as evidence strengthens and stop or redirect work that does not justify further commitment.

“Not every investment needs to pay back in the same way. Some will deliver direct value, some will generate learning and some will create capabilities that can be reused across the organisation.”- Gavin Goodland, Chief Data Officer, National Grid

How to clarify ownership across the organization

AI adoption is not owned by one team. Business outcomes, workflow change, risk, adoption and ongoing costs are distributed across the organization. You will learn why clear decision rights are not enough unless a business owner is also prepared to own the outcome and its consequences.

This matters because an approved use case without committed ownership can create a new problem rather than solve an existing one.

“We want the business to be accountable. Any of the use cases that we get prioritized need to have an impact on the long-range plan.”- Bharathi Rajan, VP – Data, Digital & AI, Swire Coca-Cola

How to protect human judgment

As AI handles more execution, people spend more time framing problems, interpreting results and deciding what happens next. The report examines why human authority must be explicit, how organizations can build role-appropriate AI fluency and why automating away opportunities to develop expertise can create long-term risk.

The central question is not whether humans remain somewhere in the process. It is whether they still have the authority, context and time to challenge an output when it matters.

How to build stability without rigidity

Organizations need enough continuity to know where accountability sits and which principles apply. At the same time, technology choices, roles and operating models need room to evolve. As the report delineates, the challenge for leaders is not to protect people from automation. It is to avoid automating away capabilities the organisation cannot replace.

The report offers a way to identify the fixed points that matter — purpose, decision rights, standards, human authority and business ownership — while keeping other decisions flexible and, where possible, reversible.

What makes the report useful for leaders?

Deliberate Decisions does not predict which technology will win. It focuses on the judgment calls leaders are making now: what to commit to, where to push back, how to move forward responsibly and how to preserve choice as conditions change.

For leaders moving from experimentation toward enterprise AI value, the report provides a practical framework and shared language for making trade-offs visible. Rather than presenting a use case as an isolated technical project, connect it to the business outcome it should change, the evidence supporting the next investment and the ownership required to sustain it.

That creates more useful conversations with executive sponsors and business teams. It also makes it easier to assess readiness, stop, redirect, or redesign work when the evidence changes, treating course correction as disciplined leadership rather than failure.

Three questions to take into your next AI review

Before approving the next use case, ask:

1. What evidence shows that this capability is ready for the conditions of the live enterprise?
2. Who owns the outcome, including adoption, operating consequences and ongoing cost?
3. Which parts of this decision should remain flexible if technology, roles or business priorities change?

These questions will not remove uncertainty. They can improve the quality of the decisions made in its presence.

Turn deliberate decisions into action

The full report develops these ideas through five leadership disciplines and perspectives from the DataIQ Guild community. It is designed to help leaders compare their own decisions with the choices other organizations are making as AI becomes more consequential.

Frequently asked questions

How can leaders tell if AI is ready to scale?

Look beyond pilot results. AI is ready to scale when the organization can support, govern, adopt and afford it in a live workflow.

How should organizations prioritize AI use cases?

Start with a clearly defined business problem and enterprise value. Manage initiatives as a portfolio, increasing investment as evidence strengthens.

Who should own an AI initiative?

A business owner should be accountable for the outcome, adoption, operating consequences and ongoing costs, with clear decision rights across teams.

How can leaders protect human judgment as AI scales?

Define which decisions remain human-owned and give people the authority, context and time to challenge, override or stop an AI system.

How can organizations stay flexible as AI evolves?

Set fixed points for purpose, standards, ownership and human authority, while keeping roles, operating models and technology choices open to review.