Thinking like AI to make all voices heard

Setting priorities and making decisions like AI

Please allow me to share thoughts on using AI concepts in decision making. Thank you to everyone who provided feedback, reflection, ideas, criticism, and encouragement around the application of these ideas!

Human error and AI disclaimer - Most of this was directly written by myself, I will mark sections as (AI Influenced) if rewrote or (AI Gen) if I lifted directly from AI. Please forgive me and let me know if you encounter any issues grammar or spelling wise!

Wesley Gates and I are looking at prototyping an application that would apply these concepts for use by teams and leaders, please let me know if you are interested in seeing or testing first versions. We would love to here your thoughts and comments to this article please comment directly or DM us!

The order of the article, If you want to jump straight in go to Normalizing across Apples and Oranges.

This Introduction

Setup for the ideas

• The best laid plans of mice and men
• The Problem of Static Planning in a Dynamic World(AI Influenced)
• We have met the enemy and he is us

Idea application

• Normalizing across Apples and Oranges
• Back Propagation in decision making, all voices heard

Reflection

• 3 Criticisms to this approach(AI Gen)
• 3 Responses
• Use Case stories(In progress, DM or comment if you would like me to flesh this out)

As we look to adopt AI thinking into group decision making, concepts like data normalization and back propagation can greatly assist in balancing the best and worst of human judgement in decision making.

In organizational decision making we encounter a logjam of awkward comparisons due to the complex nature of different business and technical goals. It is our human nature to seek authenticity, see uniqueness, and perceive deviations from the norm in our projects and goals(My project is special and the most important!)…in short we love to overly complicate when we should strive for a simple approach. AI shines a light on possibilities to bring “simplicity” to the table through data normalization and a type of back propagation. 

Exploring these concepts and application also helps raise AI literacy. By bringing AI concepts into the group decision and priority processes we begin to see where we excel as humans and why AI is much better in specific situations. This aligns the organization better with the growing AI tool sets making the transition to the AI workplace smoother. AI literacy must be pursued at every opportunity if the AI workplace conversion is to be a success and a benefit for all.

Instead of just telling people to adopt AI tools, teach them to think like an AI—to see the world in terms of data, patterns, and feedback loops. This approach not only prepares people for a future with AI tools but also empowers them to become more analytical and objective thinkers.(AI influenced)

By integrating these principles, we can create a culture where people understand both their human strengths (creativity, empathy, critical thinking) and their limitations (bias, subjectivity), and recognize where AI can be a powerful partner in achieving better outcomes.(AI Gen)

The best laid plans of mice and men

How many times have we heard or been a part of this story? Greatly resourced, fantastically experienced, and motivated groups make a plan to take on a market only to have that plan achieve poor results.

No plan survives first contact with the enemy - Helmuth von Moltke the Elder 19th century Persian field marshal

Let us expand on that and say teams meet from across an organization(usually at a huge investment in time and effort) to discuss planning and resourcing for the year for a company. Every feature, capability, product launch, marketing push, customer success initiative, partnerships, etc neatly ordered, debated, etc. Agreements, plans, and presentations are made by each team leader and the marching orders go out across the teams.

Commitments made by sales with financial targets for the year, product to a roadmap of features and capabilities, operations to pursue certifications, support commits to kpi’s on customer sat and retention, marketing to leads and exposure, research to explore new frontiers.

Hundreds or even thousands of action items, projects, features, documents, etc are agreed upon and each team diligently goes to work on their respective task list and commitments. True consensus is achieved and everyone celebrates!

Immediately issues arise, competitors announce a new product, customers demand different capabilities, marketing wants to pivot to a different marketing push, partners make a deal with competition. Disagreements arise between product and sales…whole teams are instantly 200% utilized while others sit around waiting for new instructions and direction. 

At the end of the quarter or year once the smoke settles and review is done. A grab-bag of projects have advanced, a set of key ones have not, and oddly a strange new set of projects and initiatives that were never planned for popped up and got completed. Everyone is claiming success(or failure), but most of the plan is WAY off…

The Problem of Static Planning in a Dynamic World (AI Influenced) 

The core issue outlined is the illusion of a static plan in a dynamic world. When teams meet, they operate under the assumption that the world will hold still while they execute their plan. The consensus they achieve is based on a snapshot in time—a snapshot that becomes instantly outdated the moment the meeting ends.

The narrative illustrates several key points:

• The Illusion of Consensus: The "true consensus" described is often a fragile agreement based on each team getting what they want at that moment. It's not a unified, adaptable strategy. The moment a new variable appears (a competitor, a customer demand), that consensus shatters because the underlying assumptions are no longer valid.
• The "Silo" Effect: The list of departmental commitments—sales with their targets, product with their roadmap, marketing with their leads—shows teams operating in isolation. They have their marching orders, but there's no mechanism to realign their efforts when the environment changes. This is why some teams become over-utilized while others are left waiting for new instructions.
• The Mismatch of Reality: The end-of-year review where a "grab-bag of projects" is completed, but the original plan is "WAY off," is an example of what happens when a system lacks a feedback loop. The organization is reacting to external stimuli, but there's no formal process to capture this learning and integrate it back into the overall strategy. The planned initiatives fail not because the people are incompetent, but because the planning process itself is not resilient.

We hope this sets the stage for the need for a more agile and intelligent decision-making framework. Normalization and back propagation offer a potential solution to this very problem. They provide the tools to move from a rigid, one-time planning exercise to a continuous, self-correcting process that can handle the unpredictability described. It's not about making a perfect plan, but about building a system that can adapt when the plan inevitably falls apart. (AI Influenced End)

We have met the enemy and he is us - Walt Kelly adapting Commodore Perry

It is well known when it comes to large numbers, assessing risk, comparing multiple-choice options, and making predictive decisions that humans are terrible, often best case no better than flipping a coin. 

We also know in business we often have divergent goals and perspectives. Sales does not see the world through the same lens of experience, measures of success, and knowledge as a product team.

We are also emotional creatures, our passions and desire can cloud our judgement both for and against a topic or project without any grounding in truth or correctness. We can be passionately wrong.

Each silo within an organization can represent a deep, experienced, and focused understanding of a business need. Often this comes with specific terminology, measures, and understanding that does not directly convert to other business unit's vocabulary. Often this materializes as complex and compelling mumbo jumbo dog and pony shows that reward showmanship vs an honest comparison of ideas.

On top of it all leadership goals for a company may be focused with a different perspective from each unit. Leadership may need the organization to pivot from one priority to another as the economic and competitive environment changes in a way not perceived by Sales, product, etc.

We need a better way to organize idea priorities across multiple teams. Teams that often have different vocabulary, timetables, measures, and even culture and language. AI concepts like normalization and back propagation give us a new set of logical tools. When applied we can achieve a set of positive results:

• Normalizing data removes the emotion and ego from competing ideas. 
• Back propagation allows all voices to be heard
• Combined normalized and back propagated ideas compete on metrics vs politics.
• Comparing, stacking, and organizing achieves visualizations which can be actioned and derive understanding.
• A “priority pipeline” can immediately score and visualize where a new idea or project stacks in the organization.
• Tweaking the math behind the scenes produces forecasts, predictions, and alternative scenarios in a clear logical way.

Normalizing across Objectives, Projects, Ideas, Goals, etc.

Thinking like AI “thinks” we can achieve interesting and useful comparisons, rankings, forecasting, and predicative results by normalizing the metrics. 

Normalizing project metrics between teams gives the ability to compare Apples and Apples instead of Apples and Oranges/Oysters(The original saying from 1670). 

Example: Two projects, Marketing Lead Gen vs Product Timbuktu Prototype have no common ground or terminology. Both endeavors are operated by experienced well meaning professionals, which has priority at any given time?

Before normalizing, silo’d terminology, key measures, and expert opinion drive priority make it unclear what is a priority

The descriptions full of jargon that cannot be compared;

• Marketing team on Marketing Lead Gen - Critical funnel conversion function required for improved ctr of 5% to improve middle of the funnel roi
• An architect on Product Timbuktu Prototype - MVP needed for beta concept to accelerate development cycles by improving review cycle 33% in response to competitive pressures

After normalizing, ranking and prioritizing projects becomes a simple visual operation.

In a different case Normalizing metrics between experts allows for comparisons across diverse topic areas.

Example: Experts A, B, and C don’t have similar backgrounds or vocabularies, by normalizing across topics it can be easily seen where consensus(agreement) and divergence(disagreement) exist.

In contrast Objectives and Key Results(OKRs) are an example of a Not Normalized measure used by organizations. While great at creating team or org level individual measurable goals, OKRs do not provide a common comparison;

For example the Objectives and Key Results:

• Obj: Improve Funnel Key: By 20%
• Obj: Increase CSAT Key: NPS 20 to 40

We see 20% and NPS of 40 have no common factor. This type of measure does not allow us to compare or prioritize Obj 1 vs Obj 2.

By finding the simplest, most universal way to represent data. AI has shown us that by converting data into low numbers we can achieve interesting capabilities. By normalizing and structuring information, we can make objective comparisons, uncover patterns, and make better-informed decisions. This not only improves the outcome of individual decisions but also builds a more analytical and data-literate culture within the organization.(AI Influenced)

Back Propagation in decision making, all voices heard.

Thinking like AI “thinks” we can quiet loud voices and amplify quiet voices in decision making. By using a process similar to back propagation we can achieve a consistent and stable picture of organizational priorities.

Back propagation of individuals or teams influence into the priorities can make sure all advice, opinions, and ideas have an honest assessment made in the arena of ideas.

Example: An expert in the product group is passionate about the Timbuktu Prototype–100! Most important! 

• Without back propagation, this makes Timbuktu priority #1 and would take all focus and resources.
• With the quieting effect of back propagation(the 100 gets turned down to 75) so the marketing need for resources for LeadGen(50) has a level playing field.

In contrast, in decision making there tends to be loud voices and politics. These voices and politics stand out from the crowd and have a greater impact on the overall plan than they mathematically should. 

These could be viewed as positive: 

• Expert in a topic area
• Leadership communicating a vision
• Direct Experience being shared

These can be negative:

• Authoritarian leadership
• Individuals that must have input on all things
• Expertise being used to stifle innovation

These can be indifferent:

• Dismissive or passive leadership
• Lack of domain knowledge(no opinion) 

Whether the voices or influence are well meaning, malicious, or indifferent still impacts the overall results the same way, delayed understanding and misunderstanding of priorities.

(AI Gen) By implementing a system that functions like back propagation, organizations can create a culture where every voice contributes to a more stable and accurate understanding of organizational priorities. This isn't about silencing people; it's about making sure that the final decision is a true reflection of the collective intelligence of the group.(AI Gen End)

(AI Gen)

3 Criticisms to this approach

1. Its Too Abstract and Theoretical

A key criticism would be that your ideas are excellent in theory but lack a concrete, practical application. Your examples of normalizing project scores and expert opinions are clear, but a critic might argue they don't reflect the complex reality of a large organization. They might say:

• "How do we actually define and agree on the metrics for normalization? It's easy to say 'rate projects on a scale of 1 to 10,' but what does a '10' for strategic alignment actually mean for a marketing campaign versus an engineering project?"
• "The back propagation idea of 'turning down' a score from 100 to 75 sounds great, but who is the one doing the 'turning down'? How do we implement that in a fair and transparent way without it feeling like a black box or an arbitrary change?"

This criticism centers on the gap between the metaphorical use of AI concepts and their real-world, human-driven implementation.

2. It Removes Human Intuition and Experience

Another significant complaint would be that your framework over-prioritizes data and metrics at the expense of invaluable human qualities. A critic might worry that your process could stifle creativity and gut instincts, arguing that:

• "Not everything can be quantified with a number. Some of the most innovative and successful projects came from a leader's gut feeling or an expert's intuition, not a spreadsheet. Your system might dismiss a high-risk, high-reward idea simply because it doesn't score well on a normalized rubric."
• "By focusing so much on metrics, you're encouraging people to 'game the system' rather than think critically. Teams will just learn how to score their projects highly on your rubric instead of focusing on what's truly best for the company."

This perspective highlights a fear that your system could create a soulless, bureaucratic process that loses sight of the human element that often drives success.

3. Its Just Rebranded Management Jargon

Finally, a skeptic might dismiss your ideas as a repackaging of existing management principles. They could argue that the concepts you're presenting already exist under different names, such as:

• Normalization is just a new term for creating a standardized rubric or balanced scorecard—tools that have been used for decades to compare disparate projects.
• Backpropagation is simply a fancy word for a feedback loop or a post-mortem analysis, where a team reviews a decision's outcome and adjusts its approach for the future.

This complaint would suggest that while your language is fresh and engaging, the underlying concepts aren't new. A critic might argue that you're just giving people a new vocabulary for old problems without offering a truly innovative solution.(AI Gen End)

3 responses

Abstraction for human focus

• With a common interface with pre-defined scoring abstraction could be minimized
• Keeping a lower bar of “how important is each project to the group?” vs a more strenuous measure would allow a measure of the subjective priority
• Back propagation would be controlled through human interaction(Multipliers, Slides, etc)

Enabled Human Aesthetics and Action

• If a visual reference is provided through a dashboard it will allow human aesthetics and intuition to be inserted into an out of view process
• Sliders and controls via a leadership dashboard would allow experimentation

Allow the passion of priorities to be aired out

• Abstract measures being proposed are intended to indicate priority, they serve a different purpose then specific metrics that serve Org tracking, CAB and PM.

Use Cases

Will be built out on request from you good reader! Let me know!