LLMs Are Not a Default Execution Engine

The current state of AI adoption is eerily reminiscent of the 2010s’ “mobile-first” craze, where companies felt pressured to mobilize their entire product suite, often without questioning whether it added value. Today, the trend is to incorporate Large Language Models (LLMs) into every workflow, regardless of the consequences. This mirrors what happened to Blackberry in 2010, when the company prioritized adding features over refining its core product, ultimately leading to its downfall. The latest cautionary tale comes in the form of a psychological horror film, Obsession, which serves as a metaphor for the dangers of prioritizing AI adoption over critical thinking.

The film’s protagonist, Bear, becomes so enamored with a magical shortcut that he stops questioning its consequences, much like teams today who are so focused on adding AI that they forget to ask whether it’s truly needed. AI maturity is not about building more AI, but rather about deliberately introducing AI and confidently choosing not to use it when it doesn’t add value. The real tragedy is not the initial decision to adopt AI, but the subsequent drift towards prioritizing AI adoption over critical thinking.

This drift is particularly concerning in the context of LLMs, which are incredibly good at granting wishes but won’t ask whether those wishes create meaningful value. Governance exists to protect the quality of decision-making and create space for teams to question the consequences of their actions. Before optimizing prompts, teams should first optimize the decision that introduced the prompt in the first place. Using AI wisely requires knowing when saying no creates more value than another workflow ever could.

The Decision Logic and Mechanics of AI Adoption

While companies like Unmeshed offer tools to combine model calls with deterministic rules, decision tables, and human approvals, the decision-making logic behind AI adoption is often flawed. Teams are incentivized to prioritize AI adoption over critical thinking, often due to pressure from investors or the desire to appear innovative. This leads to a focus on measuring success by adoption rather than value creation. Mature teams, on the other hand, measure AI maturity by how deliberately AI is introduced and how confidently teams choose not to use it when it doesn’t add value.

The operational mechanics of AI adoption are often opaque, with companies prioritizing the development of AI-powered workflows over the optimization of those workflows. Poorly designed prompts, for example, can drive up LLM costs, while the lack of governance can lead to unintended consequences. Companies must prioritize the development of governed workflows with deterministic rules and token-aware execution to ensure that AI is used wisely.

The cost of prioritizing AI adoption over critical thinking can be significant. Companies may waste resources on unnecessary AI-powered workflows, while also exposing themselves to unintended consequences. The lack of governance can lead to a drift towards prioritizing AI adoption over value creation, ultimately harming the company’s bottom line.

Winners, Losers, and Disrupted Parties

The trend towards prioritizing AI adoption over critical thinking will have significant consequences for various parties. Companies that prioritize value creation over AI adoption will be winners, while those that prioritize AI adoption over value creation will be losers. The development of governed workflows with deterministic rules and token-aware execution will create new opportunities for companies like Unmeshed, while also disrupting traditional software development methodologies.

The impact of this trend will be felt across various industries, from software development to customer service. Companies that prioritize AI adoption over value creation will struggle to compete with those that prioritize value creation over AI adoption. The lack of governance will lead to unintended consequences, ultimately harming the company’s bottom line.

The development of governed workflows with deterministic rules and token-aware execution will create new opportunities for companies to optimize their workflows and reduce costs. Companies that prioritize value creation over AI adoption will be well-positioned to take advantage of these opportunities, while those that prioritize AI adoption over value creation will struggle to keep up.

The Skeptical Case

One could argue that the trend towards prioritizing AI adoption over critical thinking is inevitable, given the rapid progress of AI technology. However, this argument relies on the assumption that AI adoption is always beneficial, which is not necessarily the case. The lack of governance and critical thinking can lead to unintended consequences, ultimately harming the company’s bottom line.

A more skeptical view would suggest that companies are prioritizing AI adoption over critical thinking due to pressure from investors or the desire to appear innovative. This leads to a focus on measuring success by adoption rather than value creation, ultimately harming the company’s bottom line. The development of governed workflows with deterministic rules and token-aware execution is necessary to ensure that AI is used wisely, but it is not a silver bullet.

The Signal to Watch Next

The next verifiable event that will confirm or disprove the thesis of this article is the release of quarterly earnings reports from companies that have prioritized AI adoption over critical thinking. If these companies report significant increases in costs and decreases in value creation, it will confirm the thesis that prioritizing AI adoption over critical thinking is a recipe for disaster.

Alternatively, if companies that have prioritized value creation over AI adoption report significant increases in value creation and decreases in costs, it will disprove the thesis. The release of these earnings reports will provide a concrete reason to return to this topic in 30-90 days and re-evaluate the thesis.

Bookmark this one — it will matter to your business decisions this week.

By Priya Nair, AI & Startup Reporter at TrendFlashy