The Long-Term Strategy Conference 2026 (LTSP26) is organized around a small number of precisely defined concepts. Each one is introduced not for novelty’s sake but because the existing vocabulary for AI in strategy work is too imprecise to be useful — and imprecision at this level of organizational consequence is its own form of risk.
Trendslop
Trendslop is the term coined in a March 2026 Harvard Business Review study to describe what happens when an LLM is asked for strategic advice. Rather than analyzing the specific organization, the model predicts the most socially desirable response — drawn from the same contemporary management thinking that saturates the internet and therefore the model’s training data. The result is advice that sounds tailored and authoritative while being structurally identical to the advice every other organization is receiving. Differentiate. Augment rather than automate. Think long term. Collaborate over compete. The same five answers, in the same confident register, regardless of industry, size, or situation. Trendslop is not a prompting failure. It is a process failure — and recognizing the difference is where the LTSP26 argument begins.
Deck-Chairing
If Trendslop is the failure of asking AI for too much, Deck-Chairing is the failure of asking it for too little — or more precisely, asking it to improve the wrong things. The term describes the pattern of applying AI to the non-bottleneck stages of a strategy creation process: generating slicker presentations, producing faster literature reviews, automating the parts of the work that were never the constraint. The deck looks better. The process produces the same quality of strategic thinking it always did, because the stage where thinking actually breaks down has not been touched. Trendslop and Deck-Chairing are mirror failure modes. Most organizations caught between them do not realize they are caught between anything at all.
Process improvement/management as the controversial answer
The response LTSP26 proposes to both failure modes is process improvement — and it is a more contentious recommendation than it first appears. Most senior strategy professionals believe their process is already well understood. They can describe its stages, name its outputs, and point to the retreats and reviews that mark its rhythm. What they are describing is a deck, not a process. The distinction matters because AI cannot improve what has not been clearly mapped. Process improvement in the strategy context means developing a stage-by-stage understanding of what is actually happening — where judgment is being exercised, where it is being avoided, and which stages are structurally vulnerable to Trendslop or Deck-Chairing. That level of clarity is rare. It is also the prerequisite for everything else LTSP26 covers.
The EndPoint Method
The EndPoint Method is a six-stage framework for strategy creation developed and refined over two decades of live engagements with organizations across multiple sectors and geographies. Its stages move from current-state diagnosis through horizon selection, scenario generation, commitment and quantification, backcasting from the chosen endpoint, and execution planning. What distinguishes it from comparable frameworks is not its sequence but its treatment of each stage as a distinct sociological and cognitive environment — one that is differently vulnerable to AI intervention, differently served by specific persona types, and differently exposed to the failure modes that Trendslop and Deck-Chairing represent. The EndPoint Method provides the structural spine against which LTSP26’s applied content is organized.
Personas
In the context of LTSP26, a persona is not a user profile or a marketing construct. It is a focused AI instrument built for a single subprocess within strategy creation — option generation, stress testing, current-state diagnosis, stakeholder interview synthesis. The distinction from general LLM use is critical. Asking a large language model what your strategy should be produces Trendslop. Deploying a persona trained on a specific stage of a well-mapped process, with constraints that reflect the sociology of that stage, produces something categorically different. LTSP26 introduces working personas, explains the design logic behind them, and equips attendees to evaluate, adapt, or build equivalent instruments for their own process context.