Mineral, part of Mitratech

●	Led a $25MM+ strategic growth initiative, developing a next-gen HR compliance SaaS platform, enabling a shift to a product-led growth model and unlocking new revenue channels.
●	Led discovery, assessment, and launch of an AI-powered HR agent enabling SMBs to self-serve regulated HR guidance, laying the foundation for scalable, enterprise-wide AI adoption.
●	Reduced vendor renewal costs by 30% through formalizing vendor management processes, enhancing service quality, and embedding continuous improvement practices.
●	Drove strategic alignment across the executive team by defining and operationalizing corporate strategy through OKRs, increasing cross-functional cohesion and value delivery.
●	Established and scaled a high-performing PMO to oversee critical transformation initiatives, ensuring governance, agile execution, financial discipline, and long-term operational excellence.

Vice President Strategic Transformation

Tyeideas, LLC

● Entrepreneur / Inventor / Owner @ Tyeideas, LLC B2C eCommerce retailer
● Invented and patented the DogTye® hands-free dog leash harness (patent granted Oct 2018), leading the full lifecycle from concept through market & competitive intelligence market validation through design, prototyping, and production.
● Built and operated a B2C e-commerce business on Amazon, managing product sourcing, listings, FBA fulfillment, and customer support.
● Drove go-to-market execution, including customer and competitive research, pricing, branding, and online launch, using reviews and performance data to refine the product and increase adoption.
● Owned P&L, inventory planning, and cash-flow management while coordinating with overseas manufacturers and logistics partners to maintain margins, quality, and in-stock performance.

Entrepreneur / Inventor / Owner / B2C eCommerce retailer

Slalom Consulting

●	Mineral Inc. - Led a multi-year transformation program to develop a new SaaS platform, consolidated two legacy systems, and increased annual recurring revenue (ARR) by 244%, with improved user experience, operational efficiency, and in-platform sales.
●	XPO Logistics Inc. - Led HRIS integrations using Boomi during strategic acquisitions of 150+ companies. Optimized HRIT workflow with ServiceNow to reduce manual work by 40%.
●	Portland General Electric (PGE) – As primary project lead for CSOC, led critical infrastructure OT SIEM security logging program to provide real-time threat detection and security monitoring to reduce cyber/operational risk by ~30%.

Senior Management Consultant

North Highland

●	Nike Inc. - Increased e-commerce efficiency by 60% in 3 months by optimizing Nike’s first-article digital imagery program, setting new benchmarks for digital marketing and boosting global online sales by 23%.
●	Lattice Semiconductor Inc. - Enabled scalable growth by leading a comprehensive current-state evaluation for Oracle ERP implementation, improving business performance and operational readiness.

Management Consultant

Advisicon

●	Led the consulting, design, & implementation of EPM systems for Fortune 500 clients, improving project delivery efficiency by 15–20%, enabling data-driven decision-making across business units, and equipping executives with real-time portfolio insights for cross-functional alignment.
●	Built and scaled PMO and portfolio governance frameworks that connected corporate strategy to project selection and resource allocation, improving on-time delivery and enabling leadership to manage tradeoffs across competing strategic initiatives, with advanced MS Project practices (dependencies/critical path/resource plans).

Director of Professional Services

Metro

●	Increased on-time project delivery by 30% by building and leading a centralized PMO, establishing project management governance, and improving cross-business unit alignment.
●	Enhanced organizational transparency by establishing consistent capital project reporting frameworks, improving stakeholder visibility, and accelerating decision-making and delivery.

Manager, Project Management Office

KBR, Inc.

●	Ran the regional QA program supporting the Department of State and U.S. military operations across six remote Iraq sites with a 55+ person team of SMEs, driving continuous process improvement efforts through 1,250+ monthly contractual inspections to ISO 9001 (Quality), 45001 (OSHA), DCMA (Defense Contract Management Agency) contractual standards.
●	Increased Quality Audit monthly on-time reporting by over 34.2% within a seven-month period using performance metrics and organic competitive incentives.
●	Built metrics and KPI dashboards to track nonconformances across construction, engineering, basic life support (food, fuel, water), logistics, fire services, medical, laundry, MWR, power gen, housing, waste disposal, airfield operations, property accountability, subcontractor compliance, operations and management (O&M) and contract compliance.

Senior Quality Management Leader

United States Army

Honorably Discharged. (Veteran)
Lead 87+ Soldiers

Officer

I turn strategy into outcomes by building the systems, teams, and operating rhythms that connect executive intent to measurable results. 
Former VP of Strategic Transformation with 15+ years across SaaS and consulting, I align C-suite priorities, product, and operations so organizations move faster and execute better. I have led a $25MM+ next-gen SaaS platform build, scaled ARR from $24MM to $86MM, launched an AI HR agent with real adoption and ROI, reduced vendor spend by 30%, and built high performing PMOs that drive cross-functional delivery.

𝗦𝗽𝗲𝗰𝗶𝗮𝗹𝘁𝗶𝗲𝘀:
•	Corporate & Business Strategy
•	Strategy Execution & OKRs
•	Enterprise & Digital Transformation
•	AI strategy & Adoption
•	Portfolio & Program Management
•	PMO Leadership & Governance
•	Operational Excellence & Process Improvement
•	Change Management & Adoption
•	Roadmap & Investment Prioritization
•	KPI & Dashboard Design
•	Benefits Realization & Value Tracking
•	Risk & Dependency Management
•	Executive & Board Communications
•	Cross Functional Stakeholder Alignment
•	Vendor & Partner Management
•	Team Leadership, Coaching, & Development
•	AI Program & Use Case Delivery

This Thanksgiving, I’m grateful for the conversations that pushed past small talk. People in my network took the time to ask better questions, challenge my thinking, and remind me to give back. That kind of support has given me real momentum during this career transition.

One thing I’ve noticed: a lot of professionals are quietly doing the same work. Whether they’re employed or between roles, they’re testing assumptions, sharing what they’ve learned, and helping others move forward. That kind of generosity compounds.

I won’t pretend it’s been effortless. Reaching out can feel vulnerable, but it builds resilience. For me, that’s looked like hosting short “strategy office hours,” swapping playbooks on OKRs, AI fluency, and transformation, and sharing lessons I wish I’d learned sooner.

I’m heading into the next chapter with gratitude and a clear intention: keep learning in public, keep serving, and keep connecting with purpose. If you’re on a similar path and want to compare notes, I’m always open to the conversation.

And to everyone who showed up for me this year... thank you.

~Tye

Tye Trepanier

No alternative text description for this image

Continuing my learning journey to deepen my understanding of AI’s impact on strategy and transformation in the business world.

I joined a couple of my network buddies (Graeme Rudd & KC O'Carroll, fellow military veterans), and Spenser Meeks at AI Portland last night. What a blast!

I also ran into two former colleagues who are on their own AI learning journeys and giving back in their domains (Suzame Tong & Aimee Pedretti,).

We heard from two fantastic speakers from DevSwarm, CEO Mike Biglan, & CTO Trevor Dilley. They walked through high velocity engineering (HiVE) using their AI-assisted development platform and showed how they’re leveraging Toolhive.dev for secure MCP integrations to rapidly share a live AI development use case for MCP (Model Context Protocol).
 
I'm grateful to keep learning alongside smart people who are building, experimenting, and sharing what works.
 
❓If you’re experimenting with AI in your domain or transformation work, drop a note, I’d love to hear what you’re trying.

Executive alignment is not a meeting. It is a system.
LEADING EXECUTIVE ALIGNMENT: 5 Steps That Make All The Difference

Alignment at the top is not about a nice slide deck or checking a box in a strategy session. I have learned the hard way that misalignment, even among smart leaders, can stall initiatives before they leave the boardroom. Getting executives truly on the same page requires more than meetings. 
 Here is how to make it stick:

 1. AGREE ON THE OUTCOME, NOT THE PROCESS.
 Leaders often argue over “how” instead of “why.” Start with the end in mind, what success looks like for the organization. Once the outcome is clear, process debates are easier to resolve. Just as important, once your teams understand the “why,” they can better endure the “how.”

2. ESTABLISH A CLEAR, SHARED LANGUAGE.
Terms like “agility,” “innovation,” or “value creation” mean different things to different leaders. Define the language upfront so conversations are about substance, not semantics. This often requires a simple “terms” dictionary to keep teams aligned.

3. BRIDGE THE STRATEGIC-OPERATIONAL GAP.
Executives think in big moves, teams think in tasks. Regularly translate strategic priorities into concrete, operational objectives that everyone can act on, then feed the results back up. One way to do this is by agreeing on three simple metrics for each major bet, Original, Target, and Actual. Track them honestly so the discussion centers on “what did we learn” rather than finger pointing and excuses.

4. CREATE VISIBLE ACCOUNTABILITY
 Assign clear owners for each critical initiative. Public dashboards, weekly check ins, and simple accountability charts prevent “alignment by assumption” and keep execution on track. Socialize the actual owners of the deliverables, and give them room to grow by having them report the learnings directly to the executive oversight group. This builds both ownership and capability.

5. TREAT ALIGNMENT AS A JOURNEY, NOT A ONE-AND-DONE TASK.
Alignment is not permanent. Revisit it regularly, adjust for new insights, and reinforce shared goals through action and communication. Leadership cohesion is maintained, not assumed. Setting clear metrics up front, by defining the Original, agreeing on the Target, and openly sharing the Actual, creates real alignment and shifts the conversation from blame to what was learned.

If your exec team is “aligned” on paper but not in practice, this is the work I help lead.
I have been there... I have done that…

"Action Favors the Bold"

𝗖𝗵𝗮𝗻𝗴𝗲 𝗱𝗼𝗲𝘀 𝗻𝗼𝘁 𝘀𝘁𝗮𝗿𝘁 𝗼𝗻 𝗮 𝗿𝗼𝗮𝗱𝗺𝗮𝗽. 𝗜𝘁 𝘀𝘁𝗮𝗿𝘁𝘀 𝗯𝗲𝘁𝘄𝗲𝗲𝗻 𝘆𝗼𝘂𝗿 𝗲𝗮𝗿𝘀.

Transformation is not magic. It is a series of uncomfortable, intentional choices backed by a clear picture of the future. I share a simple way to visualize success and lead by example so your team moves with you, not after you.
 
 If you lead people, you lead emotions, stories, and choices. This video shares how to align all three so change sticks, not just starts. Save it for your next kickoff.

Drop one change you’re committing to this week in the comments so others can borrow it.


"Action Favors the Bold"

I’ve learned something about myself over the years: feeling completely out of my depth isn’t an “early career” memory for me. It’s a pattern. It shows up every time I jump into something new.

And honestly, that’s been true across my entire career.

Whenever I’m in that space, the skill that saves me every time is becoming a subject matter expert as fast as possible. Not pretending to know it all, but closing the gap quickly enough that I can add real value.

For me, that looks like digging deep into the research. Reading everything I can find—articles, books, case studies. Asking the people who have been there and done it. Taking a ton of notes so I can make sense of what I’m learning and connect it back to things I have done.

Consulting teaches you this the hard way. You get put in situations where you say, “Yeah, we can help with that,” and then immediately have to figure out how. And you do, because you have to! And because the team you’re supporting needs to feel like you’re invested enough to learn their world quickly. 

It’s a mix of keeping a smile on, figuring it out fast, and being honest about what you bring to the table.

And I’ve realized something important: Being out of your depth isn’t a sign you don’t belong. It’s a sign you’re stretching into the next version of your capability.

"Action Favors the Bold"

You don’t need better AI. You need to stop pretending your data is fine.

Scaling AI is brutal because it exposes every crack you've been hoping no one notices.

AI feels magical.

Until you try to scale it.

Bain & Company makes the point bluntly: "Many companies have yet to invest; or are just beginning to invest; in critical enablers for AI value realization, including end-to-end process redesign, disciplined AI governance, solid change management, executive commitment, and an effective data strategy."

And this isn’t new. Bad data has been the kryptonite of digital initiatives long before AI showed up.

The companies that actually win with AI aren’t just chasing the newest model; they’re investing in the data behind it. They set clear ownership, build governance into everyday workflows, and treat data like a real strategic asset.

The bottom line is this: AI will amplify what you already have. Weak data? Weak results. Solid data? Transformational results.

"𝗔𝗰𝘁𝗶𝗼𝗻 𝗳𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗯𝗼𝗹𝗱"

Open article: Why AI Stumbles Without a Solid Data Strategy by bain.com

Most AI pilots stall before they scale. Data strategy and governance are not new, but they’re even more critical in the AI age.

Why AI Stumbles Without a Solid Data Strategy

One of the myths I wish more executives would let go of is the idea that AI adoption is all about efficiency or cutting costs.

Sure, AI can automate tasks and reduce time spent on repetitive work... but the real value comes from improving decision-making and empowering employees to do higher-value work. 

If leaders focus narrowly on efficiency, they risk sidelining the people who actually make AI successful: critical thinkers, problem-solvers, and those who understand the customer experience.

History is full of lessons here. Remember the outsourcing wave in the ’90s? Companies cut costs by moving IT and help desk roles overseas... only to bring them back when they realized some roles can’t be automated without damaging customer experience. 

AI is different, but the principle is the same: tools amplify humans. They don’t replace them.

"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

If your team doesn’t understand your strategy, you don’t have a strategy... you have a memo.

I’ve seen it often. Teams get handed the “perfect plan” and promptly check out because the plan doesn’t tell them why it matters. It doesn’t connect the dots between their day-to-day work and the bigger picture. That’s where stories come in.

A good story does three things: it shows the problem, it lays out the path forward, and it makes each person’s role clear. 

When you do that, people don’t just comply... they actually care. And they start thinking creatively about how to make the change work in real life.

The teams I’ve witnessed embrace change fastest are the ones who hear the story and say, 𝘰𝘬𝘢𝘺, 𝘐 𝘨𝘦𝘵 𝘪𝘵. 𝘏𝘦𝘳𝘦’𝘴 𝘩𝘰𝘸 𝘐 𝘮𝘢𝘬𝘦 𝘢 𝘥𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘤𝘦. They see the impact on customers, on the organization, and even on themselves.

"𝗔𝗰𝘁𝗶𝗼𝗻 𝗳𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗯𝗼𝗹𝗱"

AI, digital transformation, strategy… we all chase the next shiny tool thinking it’ll fix everything. But the truth is, a lot of leadership problems existed long before AI showed up.

Bill Flynn hits it on the head: “Many leaders who excel early in their careers eventually become the very reason their organizations stall.”

Think about it. Early on, being the hero by solving problems personally, making snap decisions, and pushing hard can work. You get results. You rise. You feel unstoppable.

Then your organization grows. Complexity spikes. And suddenly those same “hero” behaviors are holding you back. 

I’ve seen it firsthand. Leaders default to "doing" when "building" would serve better. But what really scales is building systems that empower others to make strategic decisions without waiting for your stamp of approval.

"Action Favors the Bold"

Open article: Great Leaders Empower Strategic Decision-Making Across the Organization by hbr.org

Great Leaders Empower Strategic Decision-Making Across the Organization

Resistance isn’t the enemy, it’s a signal.

I’ve seen it countless times during transformations: employees push back, question new tools, or hesitate to adopt changes. Too many leaders see this as a roadblock. But it's really an opportunity.

Resistance tells you where your strategy or support isn’t landing. It highlights the friction points that can create hidden inefficiencies. Ignoring it is a mistake; leaning into it is where you find the real insights.

The first step is to 𝗹𝗶𝘀𝘁𝗲𝗻 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗷𝘂𝗱𝗴𝗺𝗲𝗻𝘁. Create space for honest feedback, and make it clear that dissenting voices are valued, not punished. 

Next, 𝗮𝘀𝗸 𝗰𝗹𝗮𝗿𝗶𝗳𝘆𝗶𝗻𝗴 𝗾𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀. What specifically isn’t working, and what would make this easier or better? 

Finally, 𝗰𝗹𝗼𝘀𝗲 𝘁𝗵𝗲 𝗹𝗼𝗼𝗽: show that the feedback is being incorporated, or explain why a suggestion can’t be implemented. Transparency builds trust.

"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

𝗜𝗳 𝘆𝗼𝘂 𝘄𝗮𝗻𝘁 𝘁𝗼 𝗲𝗺𝗽𝗼𝘄𝗲𝗿 𝗼𝘁𝗵𝗲𝗿𝘀, 𝘆𝗼𝘂 𝗵𝗮𝘃𝗲 𝗴𝗼𝘁 𝘁𝗼 𝗹𝗲𝗮𝗿𝗻 𝘁𝗼 𝗟𝗲𝘁 𝗚𝗼!

The hardest leadership lesson I’ve learned about empowering others is this: 𝗹𝗲𝘁𝘁𝗶𝗻𝗴 𝗴𝗼 𝗶𝘀 𝗵𝗮𝗿𝗱𝗲𝗿 𝘁𝗵𝗮𝗻 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗸, 𝗯𝘂𝘁 𝗶𝘁’𝘀 𝘁𝗵𝗲 𝗼𝗻𝗹𝘆 𝘄𝗮𝘆 𝗿𝗲𝗮𝗹 𝗶𝗺𝗽𝗮𝗰𝘁 𝗵𝗮𝗽𝗽𝗲𝗻𝘀.

Early in my career, I believed empowerment meant giving people freedom while staying close enough to control outcomes. I wanted the best for my team and feared the consequences of failure. But what I saw was predictable: teams held back, decisions slowed, and creativity was stifled.

𝗧𝗿𝘂𝗲 𝗲𝗺𝗽𝗼𝘄𝗲𝗿𝗺𝗲𝗻𝘁 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝘀 𝘁𝗿𝘂𝘀𝘁 𝘁𝗵𝗮𝘁 𝗽𝗲𝗼𝗽𝗹𝗲 𝘄𝗶𝗹𝗹 𝘁𝗮𝗸𝗲 𝗿𝗶𝘀𝗸𝘀 𝗮𝗻𝗱 𝘀𝗼𝗺𝗲𝘁𝗶𝗺𝗲𝘀 𝗳𝗮𝗶𝗹. It means providing clarity on the vision, the boundaries, and the outcomes you care about, then stepping aside. The results are rarely perfect, but they’re almost always better than what I could have done alone.

The other part of the lesson? 𝗚𝘂𝗶𝗱𝗮𝗻𝗰𝗲 𝗱𝗼𝗲𝘀𝗻’𝘁 𝗺𝗲𝗮𝗻 𝗰𝗼𝗻𝘁𝗿𝗼𝗹. Coaching and mentoring allow teams to stretch without breaking the organization. They build confidence, and you discover capabilities in your team you never knew existed.

𝗘𝗺𝗽𝗼𝘄𝗲𝗿𝗶𝗻𝗴 𝗼𝘁𝗵𝗲𝗿𝘀 𝗶𝘀𝗻’𝘁 𝗮 𝗼𝗻𝗲-𝘁𝗶𝗺𝗲 𝗮𝗰𝘁; 𝗶𝘁'𝘀 𝗮 𝗱𝗶𝘀𝗰𝗶𝗽𝗹𝗶𝗻𝗲. 𝗜𝘁’𝘀 𝗺𝗲𝘀𝘀𝘆, 𝘂𝗻𝗰𝗼𝗺𝗳𝗼𝗿𝘁𝗮𝗯𝗹𝗲... 𝗮𝗻𝗱 𝗱𝗲𝗲𝗽𝗹𝘆 𝘄𝗼𝗿𝘁𝗵 𝗶𝘁.


"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

I’ve always thought the principles of building AI or driving transformation would translate neatly from one industry to another. Turns out…not even close.

Reading Sol’s reflections on pharma reminded me of the first time I stepped into a highly regulated, science-driven business. My instincts told me to start where I’d been most effective before: in commercial operations, customer experience, and revenue impact. 

I quickly learned that wasn’t where the leverage was.

It’s a humbling lesson I’ve seen replayed countless times across industries: the playbook that works in one context can be entirely the wrong tool in another. 

Your leverage points (the places where change actually sticks) are defined by the mechanics of the business, not by your assumptions.

"𝗔𝗰𝘁𝗶𝗼𝗻 𝗳𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗯𝗼𝗹𝗱"

I thought building AI in one industry would be almost the same in others…phew. 😅

Pharma was a completely different beast.

- In consumer goods, the product drives everything. 

- In entertainment, it's the creative output.

- In travel, it's the experience.

But pharma? Pharma is different.

Here's what I learned:

In pharma, urgency is dictated by patents and molecules. And it starts with R&D. Let me elaborate.

Because if a pharmaceutical company owns the patent for a molecule, they control a drug or vaccine for 10, 15, 20 years. 

That gives them control of an entire therapeutic area and dictates market functions for a decade or more. If someone needs the drug and you're the only one who makes it, you already won.

This insight changed where I focused my energy as a change agent.

Here are the 3 lessons this experience taught me.

1. Don't assume one playbook works everywhere.

2. Every industry has its own center of gravity. In pharma, it's the science to start. Follow the money and the power structure.  My effectiveness depended on understanding how they made their margin and profit. Not just understanding the business, understanding the actual drivers of profit.

3. Pick your battles based on leverage, not just opportunity.  There was plenty to do in commercial, , plenty of ways to add value. But those big leaps? They were harder to sell internally.

If I could do it over, I'd have focused my energy on R&D and manufacturing from day one based on the circumstances. 

Which industry surprised you the most when you got inside the engine room?

I'd love to hear your experiences in the comments below.

Measuring Real Competitive Advantage from AI: 5 Steps That Make All the Difference

I’ve seen organizations celebrate reduced headcount or faster reports, only to realize months later that their competitors had pulled ahead with entirely new business models powered by AI. Measuring real advantage means going beyond efficiency and asking if AI is actually shifting the game! 

Here’s how to do it:

1. Track impact on revenue and market share, not just cost. Look at how AI drives new sales, expands into untapped markets, or improves customer retention. If the bottom line moves only because of lower expenses, you’re optimizing for survival only. 

2. Measure time-to-market advantage. Competitive edge comes from speed. Track whether AI reduces the time from idea to launch, enables faster decision-making, or accelerates innovation cycles. A tool that simply automates a report doesn’t count; one that shortens your go-to-market by weeks does. 

3. Evaluate customer experience lifecycle gains. AI should enable more meaningful interactions. Measure engagement lift, churn reduction, or adoption rates of new AI-driven features. Customer impact often signals real strategic value faster than internal metrics.

4.  Link AI outputs to business objectives. Raw accuracy or model performance is meaningless if it doesn’t move the needle. Create explicit KPIs for strategic outcomes: revenue growth, market expansion, or customer loyalty.

5. Review and adjust with frequent loops. Regularly assess whether AI insights are still translating into better decisions than your competitors. If not, iterate!

"Action Favors the Bold"

Ever wonder how your custom AI agent actually decides what to say?

I had to share this visual from Tom Yeh on the unseen math behind vector databases.

In simple terms:
When you ask a question, your AI is not “searching documents” the old way. It is:

* Turning your question and all the source documents into vectors (numbers).
* Storing those vectors in a special database.
* Using math to find the closest match so it can pull back the most relevant chunk of knowledge.

That pipeline is called Retrieval Augmented Generation (RAG), and it is what powers things like:
* A customer service agent that can answer from your help center
* An internal chatbot that knows your policies, playbooks, or SOPs

I worked on this back in late 2023 when wiring up vector databases for real use cases was much harder than it is today. Now we have services that handle most of the heavy lifting so teams can focus on the strategy of what knowledge to load and how it should be used.

If you are experimenting with AI in your org, this is the mental model to keep in your back pocket.

Credit: Original post and visual by Tom Yeh

14. Can you calculate a vector database by hand? ✍️
[Like] if you can follow the calculation.

Vector databases are revolutionizing how we search and analyze complex data. They have become the backbone of Retrieval Augmented Generation (RAG).

How do vector databases work?

[1] Given
↳ A dataset of three sentences, each has 3 words (or tokens)
↳ In practice, a dataset may contain millions or billions of sentences. The max number of tokens may be tens of thousands (e.g., 32,768 mistral-7b).

Process "how are you"

[2] 🟨 Word Embeddings
↳ For each word, look up corresponding word embedding vector from a table of 22 vectors, where 22 is the vocabulary size.
↳ In practice, the vocabulary size can be tens of thousands. The word embedding dimensions are in the thousands (e.g., 1024, 4096)

[3] 🟩 Encoding
↳ Feed the sequence of word embeddings to an encoder to obtain a sequence of feature vectors, one per word.
↳ Here, the encoder is a simple one layer perceptron (linear layer + ReLU)
↳ In practice, the encoder is a transformer or one of its many variants.

[4] 🟩 Mean Pooling
↳ Merge the sequence of feature vectors into a single vector using "mean pooling" which is to average across the columns.
↳ The result is a single vector. We often call it "text embeddings" or "sentence embeddings."
↳ Other pooling techniques are possible, such as CLS. But mean pooling is the most common.

[5] 🟦 Indexing
↳ Reduce the dimensions of the text embedding vector by a projection matrix. The reduction rate is 50% (4->2).
↳ In practice, the values in this projection matrix is much more random.
↳ The purpose is similar to that of hashing, which is to obtain a short representation to allow faster comparison and retrieval.
↳ The resulting dimension-reduced index vector is saved in the vector storage.

[6] Process "who are you"
↳ Repeat [2]-[5]

[7] Process "who am I"
↳ Repeat [2]-[5]

Now we have indexed our dataset in the vector database.

[8] 🟥 Query: "am I you"
↳ Repeat [2]-[5]
↳ The result is a 2-d query vector.

[9] 🟥 Dot Products
↳ Take dot product between the query vector and database vectors. They are all 2-d.
↳ The purpose is to use dot product to estimate similarity. 
↳  By transposing the query vector, this step becomes a matrix multiplication.

[10] 🟥 Nearest Neighbor
↳ Find the largest dot product by linear scan.
↳ The sentence with the highest dot product is "who am I"
↳ In practice, because scanning billions of vectors is slow, we use an Approximate Nearest Neighbor (ANN) algorithm like the Hierarchical Navigable Small Worlds (HNSW).

Daniel Svonava, co-founder of Superlinked, generously lent his immense knowledge about vector databases to me by reviewing a draft of this post and providing valuable feedback. Follow him to learn more about vector databases.

#vectordatabase #rag  #deeplearning #llms #database #embeddings #aibyhand

Repost this ♻️  to help people in your network understand vector databases.

BUILDING PEOPLE: How I helped someone grow

Most leaders say they develop their people. Here’s what it looked like in real life.

I’ve noticed that our desire to learn comes in cycles.
We go from “head down, just surviving life” to short bursts of curiosity, to real excitement about going deep on a new skill.

One of my team members helped oversee a key program I managed. He was good at his job, but I knew formal certifications would boost his credibility, confidence, and long-term marketability.

I started by asking where he wanted to grow in the role. He wasn’t sure.
So I shared my own pattern: seasons where I binge on learning and certification, followed by seasons where my brain just needs to coast for a bit. I asked him to sit with that idea.

A few weeks later, I asked again. This time I came with options: specific classes and certifications that would strengthen the muscles he needed and make a bigger impact on the organization. I also made it clear the company would pay for it.

Then I backed off and gave him space.

On our next 1:1, his energy was completely different. He had picked a course, was excited to start, and finished the week-long class quickly. He immediately applied what he learned and leveled up the program in ways we hadn’t considered before.

The real point here isn’t the certification.
It’s recognizing the natural cycles of learning, catching people in that curiosity window, and guiding them toward opportunities that benefit both their long-term career and the organization.

As leaders, our job is to help people catch that next learning wave and leave them with skills that open doors long after they’ve left our team.


"Action Favors the Bold"

If your AI project isn’t blowing up your ROI in month one, that doesn’t mean it’s failing... but it might mean you're measuring it wrong. 

It’s funny, so often we walk into AI projects expecting immediate ROI fireworks. 

But in my experience, if you’re chasing a million-dollar lift in the first six months, you’re setting yourself up for a lot of awkward conversations with the board. Been there. Felt that. 

What I’ve noticed works better is pausing and asking: 
- What are we actually changing? 
- Are we preventing things we couldn’t before? 
- Are we performing better on the stuff that actually matters day-to-day? 
- Are our forecasts more reliable than last year's?

Big thanks to Sol for starting this dialogue. 

"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

I give CEOs 3 new ways to measure AI success and none of them are traditional ROI.

Not because ROI doesn't matter. Because by the time you see it, you've either already given up or your board has lost patience.

Here's the pattern I see:

Leadership teams bring me in after their AI strategy is developed. The roadmap looks perfect. The business case projects impressive returns.

But the expectations? Those need recalibrating.

AI deployment isn't just about design and deployment. 

It's about operationalization, critical mass adoption, and getting people to actually change how they work.

And that takes time. A year, maybe two before you see the numbers everyone promised.

So when I walk into these conversations, I do something different.

I give them 3 new ways to measure success, what I call the “Three P's of AI Value” (© 2025 Sol Rashidi. All rights reserved):

🔹𝗣𝗿𝗲𝘃𝗲𝗻𝘁𝗶𝗼𝗻: What are we stopping from happening now that we couldn't prevent last year?

🔹𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲: What are we executing better with AI this year that we struggled with 12 months ago?

🔹𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻: What can we forecast with more accuracy now than 12 months ago?

These metrics might feel softer than a hard ROI calculation. But here's why they matter:

If you invest $2.5 million and expect $5 million back in year one, you're setting yourself up for disappointment.

- You'll see incremental lifts. 

- Small improvements. 

- Numbers that are genuinely hard to quantify financially.

But when you reframe success around Prevention, Performance, and Prediction, something shifts.

Your ticker symbols start moving up. 

Your team sees progress. 

Your board understands the journey.

And most importantly, you set realistic expectations while staying optimistic about what's possible.

I'm positive about AI's potential. I'm just realistic about the timeline.

How are you measuring AI success in your organization? Traditional ROI or something else?

Hey my peeps, check out Raja dispel the AI productivity myths.
His message aligns well with the AI content I have shared here on LinkedIn

Thanks Raja.

Fractional Chief AI Officer | Creator of AWARE™ Driving AI ROI, Adoption & Revenue Per Employee | Ex-Deloitte

Myths are for bedtime stories.
Let’s dispel the AI productivity ones.

𝐌𝐲𝐭𝐡 𝟏 - AI Should NOT be Used just for Productivity Hacks 

I once led an ERP implementation where the sponsor kept saying: 
“We’ll start with the low-hanging fruit.”

But AI actually does have low-hanging fruit. And it’s immediate.

They buy back time. 
Time → becomes problem solving. 
Problem solving → becomes value.

With AI, the low-hanging fruit is the meal.

Your time saved is 𝑓𝑢𝑒𝑙 𝑓𝑜𝑟 𝑡ℎ𝑒 𝑖𝑛𝑛𝑜𝑣𝑎𝑡𝑖𝑜𝑛 you seek.

(ERP implementations have no low hanging fruit, the sponsor got let go)

𝐌𝐲𝐭𝐡 𝟐 - You Need 47 AI Tools to See Results
Move over Fast & Furious, there are more AI tools being released than Vin Diesel sequels.

Here’s the truth: You don’t need a tool zoo. You need one or two tools - and fluency.

Your Copilot alone - used well - can unlock millions in value. 
(See post from Yesterday in comments)

𝐌𝐲𝐭𝐡 𝟑 (𝐦𝐲 𝐅𝐚𝐯𝐨𝐫𝐢𝐭𝐞) - “AI is a Fad, We Can Wait and See”

AI adoption is not a tech race. It’s a behavioral race.

Your competition knows, believes that and is scaling accordingly.

Those who start learning now will think differently, work differently, and lead differently - long before the “wait and see” crowd even realizes the gap has formed.

By the time most leaders notice the shift, they’ll already be on the wrong side of it.

𝐷𝑜𝑛’𝑡 𝑡ℎ𝑖𝑛𝑘 𝑙𝑖𝑘𝑒 𝐵𝑙𝑎𝑐𝑘𝑏𝑒𝑟𝑟𝑦 𝑤ℎ𝑒𝑛 𝑡ℎ𝑒 𝑖𝑃ℎ𝑜𝑛𝑒 𝑤𝑎𝑠 𝑙𝑎𝑢𝑛𝑐ℎ𝑒𝑑.

Cheers - Raja Sampathi
Executive Coach | AI Leadership | AI Adoption

PS: Part 3: Mastery drops tomorrow

𝐏𝐏𝐒 - Which AI myth riles you the most?

Great post from George Paras. 

I also keep seeing the “𝗮 𝗽𝗹𝗮𝗻 𝗶𝘀 𝗻𝗼𝘁 𝗮 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆” mantra floating around LinkedIn, and it makes me pause. Not because I disagree, it’s absolutely true, but because so many posts stop at the catchphrase. Their one-pagers look like cool eye-candy, but miss the core elements of strategy.

It reminds me of times early in my career when I thought I had a strategy simply because I had a document full of bullet points and timelines. Looking back, that “strategy” didn’t guide tough choices. 

It didn’t force me (or my team) to make hard calls about where to play or how to win as defined by Roger Martin. It was a plan, yes, but it wasn’t strategic thinking in action. True strategic action is how we "Transform" our organizations.

What fascinates me is why so many leaders conflate the two. My experience tells me it often comes down to comfort. 𝗣𝗹𝗮𝗻𝘀 𝗮𝗿𝗲 𝘁𝗮𝗻𝗴𝗶𝗯𝗹𝗲. 𝗕𝘂𝘁 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆 𝗶𝘀 𝗺𝗲𝘀𝘀𝗶𝗲𝗿. It's about imagining futures that don’t exist yet and committing to them even when uncertainty is high.


"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

Everyone is saying a plan is not a strategy.

But, do you know who said it first?

The answer is B) Roger Martin.

Lately, I've seen an influx of templated 'a plan is not a strategy' posts in my feed gaining suspiciously high engagement, some over 3,000 likes, and some of which has prompted Roger to comment with a polite reminder to give some credit to his thinking. 

But beside the issue of lack of giving credit, the increasing cut + paste 'a plan is not a strategy' posts on here are also pretty unhelpful. Why? Because they do not get into the core of why execs often mistake a plan for a strategy, and the "advice" is flawed and muddled. 

More helpful advice? If you really want to know why a plan is not a strategy, then go to YouTube, search for "a plan is not a strategy" and hit play on the original viral video from Roger Martin that has over 6 million views. 

Have you seen the video? What did you learn? Why do people confuse a plan and a strategy? Share your views and answers below. 

-----------------
👍 React to support 
💬 Comment with your thoughts
♻️ Please repost this to also help others
➕ Follow me, George Paras for strategy
P.S. A link to the video is in my first comment.
------------------

𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝘆 𝗶𝘀 𝘁𝗵𝗲 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗯𝗲𝘁𝘄𝗲𝗲𝗻 𝗯𝗲𝗶𝗻𝗴 𝘁𝗵𝗲 𝗻𝗲𝘅𝘁 𝗔𝗺𝗮𝘇𝗼𝗻... 𝗼𝗿 𝘁𝗵𝗲 𝗻𝗲𝘅𝘁 𝗕𝗹𝗼𝗰𝗸𝗯𝘂𝘀𝘁𝗲𝗿.

In this article, Roger Martin makes that clear. 

I really like this quote from him: “Simply, what has changed most in strategy is the diminished efficacy of throwing spaghetti at walls and playing-to-play.”

We’ve been trained to chase multiple initiatives, hedge bets, and hope something sticks. But today, with customers discovering value instantly and fixed costs dominating investment, half-hearted plays get exposed fast. Winners now win bigger, and losers get crushed quicker than ever.

We also can't overlook AI as a powerful new accelerant in strategy today. But no algorithm is going to pick your Where-to-Play (WTP) or How-to-Win (HTW) for you. That’s the human part. 

𝗦𝗼 𝗵𝗲𝗿𝗲’𝘀 𝗺𝘆 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆: focus ruthlessly, invest decisively, and use AI to speed execution... not to make up for lack of clarity.


"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

Open article: https://www.fastcompany.com/91434102/whats-changed-most-strategy-44-years by fastcompany.com

https://www.fastcompany.com/91434102/whats-changed-most-strategy-44-years

AI is moving faster than most strategies can keep up with. But here’s the thing: AI won’t rescue a weak strategy. 

Shadow adoption is already happening, the barriers to entry are nonexistent, and the benefits are obvious. Employees are using AI tools every day. Some organizations are formalizing it with structured tool adoption. And then there are the rare few using AI to completely rethink how they operate.

The nuance? Incremental gains from tools feel good. They create momentum. 

But without anchoring AI adoption to a bigger strategic vision, to not only to improve efficiency, but to reimagine your organization, all that momentum can just be spinning wheels. 
AI is a lever. It can amplify your strategy, or accelerate failure.


"𝗔𝗰𝘁𝗶𝗼𝗻 𝗙𝗮𝘃𝗼𝗿𝘀 𝘁𝗵𝗲 𝗕𝗼𝗹𝗱"

Open article: Does Your Organization Have An AI Strategy? by forbes.com

Your organization needs an AI strategy. The real question is: how do you develop a strategy when the disruption itself is still unfolding?

Does Your Organization Have An AI Strategy?

taplio

Tye Trepanier's Linkedin Analytics

Tye Trepanier

Performance ScoreTye Trepanier's LinkedIn Performance

Who is engaging with Tye

Tye Trepanier's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?

taplio

.css-6ytdur{background-image:linear-gradient(284deg, rgba(39, 43, 72, 1) 0%, rgba(33, 47, 183, 1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Tye Trepanier's Linkedin Analytics

Tye Trepanier

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Tye Trepanier's LinkedIn Performance

Who is engaging with Tye

Tye Trepanier's Best Posts .css-1emoo84{font-size:var(--chakra-fontSizes-md);font-weight:400;color:var(--chakra-colors-gray-600);}(last 30 days)

Ready to grow your LinkedIn presence?

Tye Trepanier's Linkedin Analytics

Performance ScoreTye Trepanier's LinkedIn Performance

Tye Trepanier's Best Posts (last 30 days)