Faster Farther

calendar May 19, 2026 · 5 min read · ai engineering best-practices  ·
Share on: 🦋 linkedin copy

I’m back on the word hunt, or maybe just durable taxonomy. I’ve read some pieces lately that cover how companies adopt intelligence tools. These have mostly focused on internal automation & creation pipelines. The basic premise of these pieces seems to span two stage-based maturity models:

Consultancy Based

  1. Awareness - (I think we’re all past this one)
  2. Experimentation - pockets of an org trying to use AI to automate or build
  3. Foundational Deployment - something is in production either for the pipeline itself or driving business value
  4. Functional - multiple teams are using AI, and multiple applications are productionized
  5. Embedded - AI is now part of how a company approaches every opportunity or process

OR

Engineering Based

  1. Auto Complete - tab to build (HIL*)
  2. Task Focused - simple prompt narrow scope (HIL)
  3. Spec Driven - larger feature scopes (HIL)
  4. Semi-Autonomous Short Workflow - Agents running in controlled loop, heavily gated (HIL)
  5. Fully Autonomous Workflows - Agents running fully independent, triggering alerts and flagging anomalies (Semi-HIL)
  6. Interwoven Autonomous Workflows - Agents running independently, with additional agent workflows managing anomalies (Non-HIL)

HIL = Human in the Loop

The Consultancy Based maturity models tend to focus on the larger org, while the Engineering Based models tend to focus on the automation of digital product creation. There’s definitely overlap, even if the language is different. In both cases they represent the evolution from introducing intelligence tools into a functioning system to a system that functions because of intelligence tools. This MIT Sloan excerpt simplifies it down to 4 stages, but it’s basically getting at the same concepts.

What I don’t see is a ton of focus on what real value this transition creates. If you poke around the white papers from tech vendors and consultancies in this space, they tend to be very focused on helping you understand your level of maturity. They communicate the urgency with which you must evolve, while being obsequious at best on ROI. AI WILL accelerate your business to the moon and $$?

From what I can tell, there are two main pathways a high level of AI maturity might impact corporate fiscal outlooks:

  • Operational Efficiency - Automation is crucial to operational efficiency. If I can now have a team of 1 or 2 complete the work previously managed by 10 - that’s going to push the bottom line down - even if the cost of these tools rises significantly.
  • Time to Market/Agility - If it now takes days or weeks to get to market compared to months or years, that’s a huge expansion of the value capture window

There are some challenges inherent to both avenues that I believe organizations will need to address. These are risks that I believe teams will need to actively manage in order to unearth durable value:

Operational Efficiency

  • Groupthink: fewer people in the process means fewer divergent opinions; there’s an elevated risk of accelerating on the wrong things
  • Burnout: Super-ICs are great and may well be able to sustain long durations of extreme output, but everyone has limits and the cognitive load is going to catch up eventually
  • Refrigerator Biology - Who’s checking the back of the fridge for that yogurt from 1/13/2026 before you added the automated expiration validation AI workflow?
  • Homogeneity - If everyone can be lean using similar mixes, what makes you unique?

Time to Market/Agility

  • Cognitive Load - Even in the most automated scenario, the end product still likely has a human touch point, and there’s a finite limit to how much “new” a human can consume
  • Fragmentation - Sci-fi visions of custom-to-you software or processes built in real time are fascinating, but controlling a brand or consumer journey gets really interesting if no one encounters the same experiences.
  • Dilution - As much as being first to market can be an advantage, bringing the wrong idea, or even a less than ideal incarnation to market brings the risk of value dilution.
  • Pop-up revolts - Nobody misses the world of pop-up windows. Inundation foments irritation, if you can race to market so can everyone else

It’s not that I think there isn’t value to extract, rather I’m looking for more dialogue focused on the why. It’s never comfortable to talk about reducing a workforce, and it’s flashy to say AI ==> $$Profit$$, but there are legitimate ways to address the why in maturity assessments. For me this has been a fantastic time to flog the post-vital equestrian embodiment - Utility vs Strategic Dichotomy. It’s possible to have the conversation around this transition and center it on the ability for an organization to optimize for Strategic delivery while leaning and hardening its Utility functions. Fast and Cheap is neat, but what’s really amazing is what you can now do with your unlocked resource pool.

The word or phrase I’m seeking today is one designed to capture the intent of purposeful optimization. Somewhere between pigheaded and pragmatic lives a phrase that conveys the urgency, the gravity, and the reasoned nature that really describes an organization’s AI maturity. For now I’ll stick with outcome-centric, using the modeling linguistic tool “In order to __ I will __.” Maturity is a measure of how clearly you can tie the optimization to the outcomes you really want to achieve. The reason I like this framing is that you might only need tab-complete as an unluck. Outcomes are the goal, not the automation of production.