The Return to Simple Times Is Near

Welcome to a new edition of The Board: Distillation Aftershots ^(*).

This is an onbline copy of a newsletter, written by Constellation's Chief Distiller and Board Advisor Esteban Kolsky, that shares curious and interesting insights and data points distilled from enterprise technology to identify what’s notable. If you want to receive this in your inbox, subscribe here.

First, my take.

The transformer architecture was introduced in 2017, and by 2019 several companies, including OpenAI, were already demonstrating language models built on probabilistic token selection. The underlying technology was advancing, but it had not yet become a central enterprise priority. When the pandemic arrived, organizations redirected their attention toward more immediate operational concerns: supporting employees working from home, preserving customer access, extending cloud environments, digitizing interactions, and keeping existing systems functioning under unfamiliar conditions.

Before the pandemic, there was a persistent expectation that enterprises that failed to complete their digital transformations would be unable to survive the next major disruption. That prediction did not materialize. Research showed that many enterprises adapted through a combination of legacy applications, cloud services, spreadsheets, manual intervention, and employees who understood how to bridge the gaps between systems and formal processes.

Enterprises did not need to complete an idealized version of digital transformation to remain viable. They needed enough usable technology, sufficient operating flexibility, and experienced people capable of adapting the environment when conditions changed. The technology estates were imperfect, and the data was often inconsistent, fragmented, and poorly documented, but both remained useful.

By 2022, enterprises had accepted that the pre-pandemic operating model was not returning. Hybrid work, digital channels, distributed operations, cloud infrastructure, and more complex technology environments had become part of the permanent operating model. Organizations began to accept that waiting for perfect data, complete integration, or a fully modernized architecture would prevent them from moving forward.

The public introduction of generative AI coincided with this change in attitude. The technology was accessible, could be demonstrated quickly, and could be applied to existing processes without requiring the enterprise to resolve every underlying technology and data problem first. It provided a common direction at a time when many organizations were beginning to move beyond pandemic response and reconsider their longer-term priorities.

By 2023, conversations with executives and board members indicated that organizations were ready to establish strategies for a post-pandemic environment. Three priorities appeared consistently in those discussions: repeatable, documented growth; greater resilience in preparation for the next disruption; and sustainability in both the use of scarce resources and the creation of systems that could remain viable over time.

The investments required to support those priorities were primarily long-term infrastructure investments. Enterprises needed stronger data foundations, more flexible architecture, improved cybersecurity, reduced technical debt, better integration, and operating models capable of adjusting to changing conditions. These investments were strategically necessary, but their returns were difficult to isolate and often extended beyond the planning horizon used to approve technology spending.

AI helped make those investments easier to support. It provided visible applications, produced early results, and offered a practical reason to fund infrastructure, data, governance, and process redesign. During 2024 and 2025, many existing programs were repositioned around AI. Infrastructure modernization became AI infrastructure. Data programs became AI readiness. Process automation became agentic AI. Platform strategies were described as preparation for AI at scale.

This was useful because it moved delayed investments forward. It also gave AI a larger role in enterprise strategy than the technology could sustain. Many organizations began treating AI as the transformation agenda rather than as one component of the architecture and operating model required to achieve their larger objectives.

That approach began to change toward the end of 2025 and has continued through 2026. The discussion has moved away from demonstrations and isolated pilots and toward the conditions required for enterprise deployment: governance, infrastructure, scaling, economics, talent, and business purpose. These were the areas covered in the CEO’s AI Dashboard series because they determine whether AI can produce durable enterprise value after the initial opportunities for bounded automation have been addressed.

The Business Value installment argued that AI should be connected to measurable changes in process, cost, cycle time, risk, revenue, or margin. The Data Readiness discussion focused on the difference between data that is useful for reporting and data that can be trusted to support automated action. Infrastructure examined the architectural and operating foundations required to deploy AI across the enterprise. Governance addressed accountability when systems can act without waiting for human intervention. Economics considered how to balance short-term productivity, foundational investment, and long-term growth. Talent focused on preserving the experienced judgment required after automation reaches its limits.

Taken together, these areas point toward the next stage of enterprise planning. During the second half of 2026, organizations are beginning to position AI within a larger infrastructure and operating strategy. AI will remain important, but it will increasingly operate as one capability within private platforms, hybrid environments, governed data systems, redesigned workflows, and architectures that support different models and providers.

This shift allows enterprises to return to the priorities that were emerging before AI became the center of the discussion. Growth, resilience, and sustainability can again become the organizing principles for enterprise strategy.

Growth requires repeatable systems that expand revenue capacity, improve process economics, strengthen customer outcomes, and support new products and business models. Resilience requires infrastructure and operating models that continue functioning during disruption and can adapt without another period of emergency improvisation. Sustainability requires systems the enterprise can afford, govern, secure, staff, and operate over the long term, while making better use of capital, energy, data, talent, and other constrained resources.

AI contributes to each of these priorities, but it does not define them. Models will change, vendors will change, and pricing structures will change. The enterprise therefore needs to retain control of its architecture, proprietary context, governance, operating processes, and institutional experience while using AI where it improves the larger system.

The strategies required to do this are only beginning to take shape. The work completed during the past several years provides a useful starting point, but executives now need to place those investments inside a clearer enterprise agenda.

The first step is to reconnect AI initiatives to specific growth, resilience, or sustainability objectives. Business value should be measured through changes in enterprise performance rather than through usage, token consumption, or vendor-defined outcomes. Data and context should be treated as long-term enterprise assets, with continued investment in semantics, quality, access, and controls.

Infrastructure should be designed for reuse across models, vendors, agents, and future workloads rather than optimized for one generation of technology.

Governance also needs to precede expanded autonomy. Enterprises must define what agents may access, which actions they may take, where they must stop, and who remains responsible for the outcome. Investment models should distinguish between short-term productivity, foundational capability, and long-term growth because each has a different purpose and time horizon. Talent strategies must preserve the experienced operators and human linchpins who understand exceptions, consequences, and undocumented processes, and incorporate them into the design and supervision of AI-mediated work.

The return to simple times is a return to a clearer set of priorities. The pandemic displaced growth, resilience, and sustainability while enterprises focused on operational continuity. AI then became the principal organizing idea as organizations searched for a direction beyond recovery. The current transition places AI inside a broader strategy and allows the enterprise to concentrate again on the outcomes it was trying to achieve in the first place.

Growth, resilience, and sustainability provide that direction. AI remains an important part of the supporting infrastructure, alongside data, security, governance, integration, observability, and experienced human judgment. Its value will increasingly depend on how well it contributes to that larger system.

Reading resources: