Cameo Doran

Chief Product Officer ⇢ Fractional Prices with Full-Time Impact ◆ Digital Transformations ◆ Advising Startup Founders and Executives Building Innovative Products ⇢ AI / XR / Web3

I’ve always been fascinated (and a little frustrated) by how many great companies can’t seem to ship great software. I’ve sat inside teams where brilliant ideas died in Jira backlogs, and I’ve watched the opposite too: features rushed out fast, only to flop because they were poorly executed or customers didn’t care. Both hurt. Both waste time and burn cash.

Cameo Labs is my answer to that problem. I started it because I couldn’t stand watching founders lose momentum on ideas that should have changed their business and their lives. 

My mission is simple: help founders build the right product, fast enough to matter.

Since starting, I’ve led projects that:

- Shipped an EdTech platform that closed $3M in contracts in under five months.
- Migrated 1,000+ servers with zero downtime, saving $6M.
- Built an AI training platform that scaled to $1.5M ARR in 9 months.

I’m dorky about this stuff. I love the messy early product conversations, poking at assumptions, figuring out where speed really matters, and then helping a team pull off what feels impossible. That’s what gets me up in the morning.

If you’re a founder under pressure to ship something that can’t wait, I’d love to hear about it.

Chief Product Officer / Fractional Product Management

As the owner and founder of Cameo Labs, I am dedicated to helping CEOs and executives of $100M - $500M businesses in the Edtech, Fintech, and SaaS sectors achieve rapid and transformative innovation. 

At Cameo Labs, we understand that staying competitive in today’s market requires more than just innovative ideas—it demands strategic execution and a customer-centric approach.


I founded Cameo Labs to empower our clients by streamlining their product development processes, aligning their business strategies with technological advancements, and fostering a culture of continuous innovation.

We operate differently than most agencies. 

Our holistic approach combines technology, strategic foresight, and customer insights to create cohesive and impactful product visions. Our clients have experienced faster innovation cycles, improved product success rates, and higher customer satisfaction.

With a proven track record of success, we are committed to turning your challenges into opportunities for growth and success. Our goal is to be your trusted partner, helping you achieve your business vision and lead your industry.

Let’s Connect:
Discover how Cameo Labs can transform your business. Visit http://cameolabs.com or reach out to me directly to explore how we can work together to drive your business forward.

This headline cracks me up.

OpenClaw runs on the infrastructure Linux built. Without it, there’s nothing to achieve in 3 weeks or 3 decades.

I don’t know if this is a direct quote, but if it is, it doesn’t make Jensen look smart. It makes him look butthurt. 

Linus has been openly dismissive of NVIDIA for years, and this feels less like an actual data point and more like a score he is trying to settle.

From Vision to Execution | Faster Products, Better Results, Bigger Impact | SaaS Companies | Strategic Software Partner

No alternative text description for this image

IP is the bedrock of modern capitalism. 

It’s the promise that if you build something, you own it and you get paid for it. But that promise is crumbling. 

Of course, LLMs were trained on thousands of copyrighted works without permission. And tools like Claude Code let you screenshot a product and whip up a clone in an afternoon.

And these Ai company’s are so confident they’ll get away with it that they’re not even being subtle anymore. 

Earlier this week, Grammarly got caught promoting the ability to have an literary expert edit your content. Except they never asked the experts if they could use their names, and never had any intention of paying them. 

They didn’t just steal their work, they stole their name and creative identity. 

Grammarly would only do something like this if they believed everyone’s already cavalier about identity, ownership, and the rest of it. 

I don’t believe these are edge cases. The are testing to see what we’ll tolerate. 

If IP protections become unenforceable, we’re not just losing copyright law. We’re dismantling the incentive structure that we built modern business around. 

People talk a lot about jobs being lost, but are we actively destroying the very foundation that makes our economy work?

And if so, what comes next?​​​​​​​​​​​​​​​​

As a side note, Grammerly quickly turned it off after the backlash. But I am sure they will try to do something like it again.

graphical user interface, text, application

Today is the 2026 Utah Tech Week debrief, where hosts and participants will come together to figure out how to make next year even better. 

As I was filling out the pre-debrief survey, I just had to stop and say it:

I love Utah Tech Week.

I remember years ago seeing Trent Mano’s post announcing the whole thing. He’d gone to LA, saw what their tech week looked like, came home, bought the domain, and announced it the next day. From the moment I saw that post, I knew I wanted in.

So that first year I co-hosted two events: 
A Web3 in 2023 panel with Krystee Miller, Derek Wride, and Nicki Sanders. 

and an Assumptions and Risk workshop with Connor McLeod and Jacob Crockett

The next year, I co-hosted three events: 
AI acceleration - using AI to map assumptions with Lucy Harrison, 
a PMF panel with Jennifer Moore, 
and a strategic foresight working session with Christopher Federer (which somehow I have 0 pics of)

The next year I didn’t run anything of my own. I just showed up as a volunteer and spent the week helping where I could. 

And this year, I co-hosted one event. Vibe Mania with Krystee Miller and it was a huge success.

4 years and honestly it has been life changing. 

But the best part has never been the events. It’s been the people. Meeting new people that turned into true friends. Collaborators who turned into business partners. Individuals who became customers. 

So may people who have changed the direction of my life and I met them because of Utah Tech Week. 

So thank you Trent Mano for imagining what could be possible, and thanks to Kyle Spencer and the rest of the team for making it happen every year.

undefined undefined

cmeo-labs

voyantai

productinnovationlab

theory-of-anything-podcast

wow3

Utah Tech Week

Pre-seed investor/GP @ Convoi Ventures

Head of Product | UX Strategist | Designer | All the things

Founder & CEO @ CPAI - Crypto Tax Intelligence

Blockchain & Crypto Tech/Strategy Leader and Consultant | Engineering Leadership at Anchorage Digital

AI, GTM, and Product at SageCreek.ai and SageCreek Partners

fixing the hiring bottleneck for home services || ask me for my cookie recipe

CEO at Voyant AI | Driving Enterprise AI Adoption | From Readiness to Deployment

Chief Product Officer | Board President | Keynote Speaker | Community Builder

I share tips and tricks for strategists working in Experiential Learning industries.

Helping others buy or try campervans Operations | Strategy | Events

Most of the crap in your backlog isn't worth building. 
And AI is about to make that problem a whole lot worse.

I know that sounds harsh but it's true.

According to the product analytics company Pendo, close to 80% of features in the average software product are rarely or never used. 

Now hand your team AI coding tools and watch what happens. They will build faster. If they are smart and have good fundementals, it could be way faster. 

Oh YAY, now you are building more things your customers dont want. 

So instead of using AI to continually think of new things to add to your product, figure out how to eliminate more out of your backlog. 

Then you can truly focus on things that bring impact and value. 

The teams winning with AI aren't the ones generating the most code. They're the ones who figured out what not to build, and then used AI to build the right things really well.


~~~~~~~~~~~~~~~~~~~~

Hi 👋 I am Cameo
▶ Helping startups fast-track product success! 💡

All these data centers will be worth it, right?

Claude code can’t fix everything!

That was my takeaway from yesterdays The AI Collective Utah Claude Code event (sponsored generously by Anthropic) 

Josh Dennis did a fantastic job demonstrating how his company is using Claude code to build their customers custom marketing solutions. 

But then I got home and noticed I had this enormous chocolate blob on my shirt, which means I was talking to people and networking with people all night with this enormous chocolate blob on my shirt, and Claude didn’t do one damn thing to fix it. Seriously what’s even the point of AI?

So as much as Claude has improved and changed my life, it hasn’t figured out how to eliminate my remarkable ability to embarrass myself. 🤦‍♀️

Thanks to Leslie Koch and the rest of the amazing AI Collective on a great event. The chocolate cookies were delicious even if I’m not sure if they were worth it.

SheTech Explorer Day is a full day experience designed to expose teenage girls to the possibilities of working in technology.

This year that meant 3,500 girls from across Utah coming together in one place. And let me tell, that is a lot of teenage girls.

I was grateful for the opportunity to host an AI workshop with Leslie Koch and The AI Collective. 

Even though research shows that women are adopting AI at a slower pace than men, I was pleasantly surprised to see how engaged and thoughtful these young women were about AI and how they are using it in their day to day. 

Thank you to Cydni Rogers Tetro, Kristin Wright and the whole team at Women Tech Council for creating such a critical annual event for our girls. 

I believe it doesn’t only change the trajectory for the girls who attend, but it changes the perspective for any adult who gets to be there as a mentor or volunteer.

Growth Marketing & AI Strategist | B2B, Creator and Social Marketing | Ex-TelevisaUnivision, CNN, HBO

The AI Collective

Growth Leader| AI Women Power List | 3× CEO & Co-Founder | Board Director | Investor | xSwire Coca-Cola | xDisney

Programs Director at Women Tech Council

Women Tech Council

$14 billion can buy a lot of things. But apparently it can't buy your way into AI leadership.

In this week's Build Signal, I break down how Mark Zuckerberg's record-breaking investment in Scale AI and its founder, Alexandr Wang, has unraveled in less than a year.

I think this matters for anyone building in AI right now because the core mistake is one I see constantly: Confusing expertise in one area with the ability to lead the whole thing.

Read it, and I'd love to hear your thoughts.

There's a specific type of mistake that executives keep making when hiring or promoting AI leadership. They find someone who is clearly brilliant, cle

Open article: The Most Expensive Hiring Mistake in AI History by Cameo Doran • 4 min read

The Most Expensive Hiring Mistake in AI History

Say it with me:
It’s dumb to build without talking to customers.
It’s dumb to build without talking to customers.
It’s dumb to build without talking to customers.

This year you are going to see this argument a lot: “Software is so easy now, so the constraint moved to marketing.” Or sales. Or distribution.

Nope!
If you’ve been building for months and haven’t talked to a single customer, the constraint didn’t move anywhere. You’re just doing it wrong.

CB Insights analyzed 431 VC-backed startups that shut down since 2023. The number one root cause? 43% failed due to Poor product-market fit, and two-thirds of those were early-stage companies that never found a market at all. 

These weren’t teams that had a marketing problem. They had a “We spent a boatload of VC money and we still built the wrong thing” problem.

Vibe coding and agentic development can be a trap if you’re not careful. It feels so easy and productive while you are building that you never stop to ask whether anyone cares. 

You can ship in days features that used to take weeks and mistake speed and velocity for validation.

A product that’s easy to market doesn’t happen because you figured out distribution after the fact. 
It happens because you solved a problem that actually matters to real people, in a way that resonates with how they think about it. 

That requires thoughtful discovery and customer work. 

Knowing what to build is still the whole game. The speed at which you can generate code doesn’t change that.

~~~~~~~~~~~~~~~~~~
Hi 👋 I am Cameo Doran 
I help your org move faster by identifying the right thing to build the first time.

It sure is easy to start things 

And vibe coding is like magic.
New project? Ten minutes and you've got a working prototype. You get the dopamine hit and start to think you a genius.

Now all you have to do it polish it, publish it, and promote it, right?

Except you said that about the last four projects too. And the six before that. And you've never done anything with them either. lol

Starting has always been the easiest thing. No matter how many times you see hyperbole that says the first step is the hardest, it's actually continuing to take step after step that's really hard.

LLMs make it easier than ever to start.
You better get damn good at finishing.

”Who is a woman who's made an impact on your career? Celebrate her with a post or video.”

I’m sure you’ve seen this prompt from LinkedIn for International Women’s Day. 

Yeah, That’s nice. Go ahead and do that.
But don’t pretend it’s anything other than being performative. It’s weak sauce.

Women don’t need more celebration or more empowerment. We don’t even need more mentorship.

We need to get paid. 

We need sponsorship. We need partnership.
We need more promotion. And we need more investment. 
We need to get funded and we need to get elected to office. 

Anything else is just lip service.

Anthropic wants you afraid. 

They released this chart last week.
And if you are in the blue zone, your job is disrupt-able with AI. 

But look at it carefully. The blue is what AI could theoretically do. The red is how people are actually using AI. 

That gap is massive. For industries supposedly facing replacement within 12 months, real adoption sits at 10-20% of potential. And that is across every category.

This matches what I see with every client. There is a high desire to adopt AI. But organizations don’t have the capability to execute on it.

I think this chart tells more truth than Anthropic meant. We are nowhere near the disruption that they are selling. The organizations aren’t ready.

The savviest companies aren’t just buying AI tools. They’re doing the change management work to make sure people actually use them.

I’m partnering with Voyant AI on executive training built around exactly that. Drop a comment or message me if you want details.

Hey Anthropic, where's the data?

I'm sure by now you've seen these charts. 
The first is the radar graph Anthropic released with their labor market report last week. 
The second is a bar chart version that Peter Walker made by extrapolating values off that radar graph, because that was literally the only way to get the numbers.

I went and read the full report. And guess what?
Other than computer/math and a passing mention of office & admin, they don't publish the actual percentages for any of the occupational categories. 
Not anywhere in the report in the report, not in the appendix and not in a supplementary table.

Interesting... The only place that info appears is in that unlabeled spider chart. 👀 

The entire point of this paper is introducing a new metric called "observed exposure" that they want researchers and policymakers to adopt as a framework for tracking AI's labor market impact. 

They go out of their way to describe their methodology. They make a big deal of how they came up with their unemployment analysis. So it's super baffling that the central finding of the paper doesn't provide any actual numbers. It's almost like it's all made up.

The report still tells an important story though. There iss a massive gap between what AI can theoretically do and what the vast majority of organizations are actually doing with it. 

If you're an executive trying to figure out how to close it, send me a DM.

Your developers aren't the problem. Your decisions are.

I sat in a meeting last month where a CEO realized his CTO and head of product had been building toward a different vision from his for five months. He thought they were close to release, but now he's looking at something that is likely to miss the mark.

This kind of thing happens all the time.
In this week's newsletter I talk about the 6 decision that would have caught and prevented this problem early.

The single biggest factor in whether a project succeeds or fails isn't methodology, team size, or technology. It's how fast leadership makes decisions

70% of Software Projects Fail. Here Are the 6 Decisions That Beat the Odds

Have you heard that Spotify's engineers haven't written a single line of code in months?

A lot of CEOs are seeing that headline and now they are mad at their teams. 
Why can't their team do that? 

They want their developers using AI that effectively. Hell, they want their whole company using AI that effectively.

They want to achieve what Spotify has, and they're looking around wondering who to blame. And they should want that. But most of them aren't being honest about why they don't have it.

Spotify didn't just hand their engineers AI tools one day and watch magic happen.  This isn't a bunch of schmucks who got lucky with AI. They worked for years on this foundation first. 

- clean data 
- clear processes 
- reliable architecture
- consistant standards

So when AI showed up, it had something to work with.

But at most companies, it's the opposite. Their data is a mess, and their processes are fragmented. Nobody's following the same playbook because there isn't one. And now they've dropped AI in the middle, and they're wondering why it's not performing miracles

So before you get mad at your team for not being Spotify, ask yourself what foundation they are working with.

How important is AI accuracy?

It really depends on what you are using it for. 
If you're generating marketing copy or brainstorming product names, close enough might be fine. Nobody dies if your tagline is a little off.

But 40 million people are using ChatGPT Health every day to decide whether they need to go to the ER.

And according to a new study in Nature Medicine, the results are abysmal. 

ChatGPT Health failed to recommend emergency care in over half of actual emergencies. Among a host of other issues, It told patients with impending respiratory failure to wait 24-48 hours instead of going to the emergency room.

and the suicide crisis safeguards didnt perform any better. In some cases, when patients described specific plans for self-harm, the crisis banner didn't appear at all.

We keep hearing that speed and scale are the metrics that matter. But speed without accuracy is worthless when the situation is life or death. 

And remember this is the same systems that the US government is going to let make autonomous decisions...

What are you willing to sacrifice to make a good impression? 

If you're like me, you get daily reminders from LinkedIn about how important it is to have a verification badge. 

Oh, and how exciting! It's free! 

I have spoken out several times about LinkedIn’s verification service, and the dark patterns they use in promoting it and how unnecessarily invasive it is, but I honestly didn’t realize the full horror of what they are doing 

This article is a pretty shocking breakdown of what they are asking you to give up to get that blue badge, and "make a stronger Impression"

I highly recommend you read it. A lot of the info in the article is specific to the EU and rules like GDPR, but if you're American, you don't even get that protection, so it is even more horrifying.
https://lnkd.in/gguNZ53v

Is it worth it to you? Its a hell no for me. 
Also LinkedIn, I get plenty of profile views, even without verification, thank you very much.

Nano Banana 2 dropped today and it is so impressive! 

But damn my productivity took a hit today.

I am a big fan of leaders being willing to take risks and make mistakes.

But also, this seems egregious for someone in this role.

The Director of AI Safety at Meta just let an AI agent delete her entire inbox. 
She told it “confirm before acting.” It did not confirm. It just deleted everything. 

She had to physically run to her computer to disable it.
She called it a “rookie mistake.” 

But she is the person Meta hired specifically to make sure AI doesn’t do things it’s not supposed to do. That’s the whole job.

I don’t want to pile on her personally. She has gotten plenty of that, including from Elon Musk.

But the lesson here is that if the head of AI alignment can lose control of her agents, what does that mean for the rest of us handing AI access to our business systems and customer data?

The answer isn’t to stop using the tools. But you best make damn sure you are taking the guardrails a lot more seriously than most people currently are.​​​​​​​​​​​​​​​​

~~~~~~~~~~~~~~

Hi 👋 I’m Cameo. I turn AI pilots into production software for companies that can’t afford to get it wrong.

Expert catalyst in driving Innovation from 0 to 1, excelling in building innovation teams, launching new products, embracing emerging tech, and a passionate couture design enthusiast

When leaders face their most complex, mission-critical challenges, the ones that keep them up at night, they call me.

I'm a strategic advisor and execution partner who transforms emerging technologies into measurable business results. For over two decades, I've been the trusted right-hand to leaders across global enterprises and fast-growing ventures alike navigating their most complicated transformations, from AI adoption to digital overhauls to innovation at scale — unlocking growth, reducing costs, and building lasting competitive advantage.

I don't just give advice, I roll up my sleeves and build alongside you. Whether it's leading AI adoption strategies, driving digital transformations across multiple brands, or tackling the high-stakes, multi-stakeholder initiatives where failure isn't an option and collaboration is key. With deep technical expertise and business acumen, I bridge the gap between strategy and action.

My work spans financial services, consumer goods, technology, and more. I'm recognized as a thought leader and regularly advise industry organizations and universities on emerging technology trends and implementation strategies. 

I'm passionate about helping leaders navigate what's next. So, let's talk about what keeps you up at night.

Darlene Newman

The most valuable person in your AI initiative right now isn't an AI engineer, it isn’t a prompt engineer, and it isn’t a data analyst.

It's the person who can answer two questions with enough precision to be useful:

How does this work actually happen?

Not the process diagram on the wall. The real one… the decision rules, the exceptions, the escalation paths, the moment where someone just knows what to do next.

What does the system need to know to do it?

That second question is where most organizations have the biggest challenge, especially when moving from simple chatbots on your policy documents.

AI agents don't hallucinate because the model is bad… they fail because the model will always generate a response, whether or not the structured context it needs actually exists, generically.

No one did the hard work of making enterprise knowledge AI-legible before deployment.

--> The entities weren't defined.
--> The relationships weren't mapped.
--> The terminology meant three different things across three different teams.
--> The criteria for "successful" completion lived in someone's head for 15 years.

What you actually need is a knowledge architect — someone who can sit with your domain experts, extract the real process logic at an atomic level, understand what structured information an agent requires to reason over that logic, and build the bridge between the two.

This isn't prompt engineering. It's not data engineering. It's a distinct discipline… part knowledge management, part systems and data thinking, part translation.

Most organizations don't have this person. Many don't even know they need one.

But if you want to scale beyond pilots… where agents need access to systems of record and must operate across many domains, data sources, and teams… this role becomes the difference between a proof of concept and a production system.

AI doesn't create that knowledge as it performs. It depends on it to complete anything successfully.

Strategic partner for leaders’ most complex challenges | AI + Innovation + Digital Transformation | From strategy through execution

The AI ROI measurement problem isn't just your problem. It's everyone's problem, even those that claim to be experts in AI.

Accenture tells clients… measure outcomes, not AI usage.

Yet, they just rolled out an internal policy tying senior promotions to weekly AI login frequency.

Take a second with that. Then keep reading… because this isn't about hypocrisy. It's about how hard transformation measurement actually is.

This is the same Accenture that co-published an AI maturity framework with Carnegie Mellon just two months ago… explicitly built around moving organizations beyond hype-driven experimentation toward measurable outcomes.

Two months of telling clients how to measure ROI… yet, their measurement of performance is related to how much a senior leader logs in to an AI tool. Not whether those logins were worth the cost of the tool.

A login tracker doesn't prove usage, doesn't prove innovation, and certainly doesn't prove the work got better.

I'm not dunking on one firm. I'm pointing out what every organization does under pressure… we revert to what's visible. Logins are visible. Adoption dashboards look good in a deck. Activity feels like progress.

Even the best of the best argue for alternatives to checkins:

🔹 McKinsey says winning companies embed AI into workflows and track business KPIs, not deployment counts.
🔹 Gartner called activity metrics out directly… they don't translate to boardroom value.
🔹 Berkeley researchers put it plainly… we're not in an AI failure era. We're in a measurement failure era.

The problem isn't that Accenture, or any other organization is lying about progress in AI. It's that we genuinely don't know how to measure the middle… the space between "tool was accessed" and "business improved."

So we take the easy route.

But there are ways... and it's how innovators always measure... incrementally.

1️⃣ Prove it changes the work

Before anything else, is the task itself different? Time to output. Error rate. Rework cycles. Quality against a baseline. This is your proof of concept at the individual task level. Not "did they log in" but "did output positively change."

2️⃣ Prove it changes the workflow

One person using AI quietly doesn't transform a process. Look at end-to-end cycle time. Handoff friction. Exception rates. Throughput. This is where you see whether AI has actually changed how a team operates, or whether the process around them is still broken.

3️⃣ Prove it changes the business

Now you have causality. Connect it up the chain. Revenue influenced. Cost per transaction reduced. Margin impact. Risk reduction. This is what earns leaders the seat at the table.


Each step builds evidence. Each step tells you whether to keep going. 

If you measure activity, you will optimize for activity, not whether that activity made any positive change. 

This isn't easy by all means. But you need the to measure what's hard to see, not just what's easy to track.

Links to sources in comments.

If your people are replaceable by AI, then your business might be too.

There's a growing narrative from tech leaders (e.g. Mustafa Suleyman at Microsoft, Dario Amodei at Anthropic, just to name a few), predicting that most white-collar work will be automated, possibly even eliminated, within the next 12–18 months.

Maybe. But I highly doubt it. And you probably don’t want it to.

AI can absolutely automate tasks, sometimes even better than humans.

But enterprises don't compete on tasks. And if they do, that's less of a “bold strategy” and more a confession about how thin their competitive advantage really is.

There’s a significant difference between “tasks”… discrete, automatable outputs and “work”… the complex, relational, judgment-driven ecosystem where value is actually created.

Enterprises compete on judgment under uncertainty, context carried across decisions, institutional memory, relational trust, and to know escalation logic when things break.

--> The senior partner who spots the deal-killing clause from memory.
--> The project manager who knows when to give space.
--> The marketer whose best idea came from a hallway conversation.

None of that lives in a dataset. It comes from human experience, trust, and social fabric… and it's exactly what holds organizations together when things get tough.  

When every firm in your industry runs on the same foundation models, the same copilots, the same automation stacks… capabilities, outputs… everything converges.

The only remaining lever is price. And price competition is how industries destroy their own economics.

Strip out the human root structure in the name of efficiency, and you may not notice the damage until the drought hits.

A regulatory shift. A model failure. A market inflection. Without judgment embedded in the system, there's nothing left to stabilize it.

🔹 Where does judgment materially change outcomes? If you can't point to them, you don't have an advantage.
🔹 What context walks out the door when your experts and senior leaders leave? If the answer is "a lot," your moat isn't embedded.
🔹 Does this AI investment increase differentiation, or just efficiency? If everyone can buy the same capability, it's not strategy.
🔹 Are we amplifying human judgment, or removing it? AI should reduce friction around your highest-value thinking, not eliminate it.

AI should make your organization more differentiated. Not more interchangeable.

Article in comments

Why is everyone is talking about "context graphs” when most organizations haven't even defined their own controlled vocabulary.

Gartner's latest report argues that traditional databases are structurally unfit for agents… and that leaders have roughly two years to close the gap by building what they call context graphs.

Foundation Capital makes a similar case… the next platform shift won't come from bolting AI onto systems of record. It will come from capturing what those systems never stored… decision traces, exceptions, authority, the reasoning behind how work actually gets done.

They're not wrong.

Agents can't reason without understanding structure. RAG retrieval isn't enough. Vector search doesn't encode authority. Copilots don't understand the process.

But before we encode "institutional memory," into a context graph, we first need institutional meaning.

Unless you're AWS or Google, most organizations are starting from zero... no knowledge graph, and in many cases, no consistent definitions to build one from.

Ask a simple question inside almost any firm:

What exactly is an "incident"? What makes a contract "high risk"? Is it a client, an account, or a customer… and does that distinction change the number? What's the value of that contract across amendments?

Depending on who you ask, you'll get different answers, different ways of identifying, different ways of aggregating.

And now we need to automate decision-making on top of that?

Context graphs assume shared meaning, not just in standard terms, but in decision ownership and authority. That's like jumping from I'm learning to full maturity...

So before we jump to context graphs, let's start somewhere more practical… and more revealing.

🔹 What are the key terms in this process?
🔹 Does everyone call them the same thing?
🔹 Do you agree on where each concept begins and ends?
🔹 Who owns each definition?
🔹 When departments disagree, which version wins?
🔹 Is any of that actually documented?

It's not glamorous work. It doesn't demo. You can't wrap it into a 12-week sprint.

But without it, agents wil never accurately reason. They amplify semantic drift.

The maturity path isn't... LLM → Context Graph

It's... LLM → RAG → Controlled Vocabulary → Ontology → Knowledge Graph → Decision Authority Layer

The breakthrough isn't "context graphs." It's admitting we've never formally agreed on what our business actually means.

Until we do, autonomy is theater.

Gartner and Foundation Capital reports in the comments

What will it take to eliminate the Silicon Valley mindset?

If you haven’t seen it, the AI-generated video of Tom Cruise fighting Brad Pitt is impressive. Still, slightly fake-looking… but it’s pretty darn good. And, apparently it was created from a two-sentence prompt.

AI is getting better. Fast. So fast many of us practioners can’t keep up.

What wasn’t impressive though was the response from ByteDance’s CEO when asked about copyright concerns, which was framed as something that could just be addressed… after the fact.

I get it. It’s “just copyright”

Receive a cease and desist letter
Pull it down
Update the model
Pay the fee
Move on

Meanwhile, the video did exactly what it was meant to do… drive traffic, drive usage, drive revenue, drive market share.

And the cleanup simply becomes a cost of doing business.

But this isn’t just about fan creations… it’s about the mindset we’ve normalized.

Move fast and break things
Ask for forgiveness, not permission

Whatever phrase you choose, what is really being articulated is…

Succeed at all costs
Clean it up later

That logic may work when the damage is reversible. But what about when it isn’t?

We are already seeing what happens when scale outruns safeguards.

Social media companies are in court over algorithmic amplification and addictive design, particularly its impact on young users.

AI systems are being scrutinized for how they respond to human vulnerabilities when safeguards fail or lacking.

These aren’t abstract ethics debates. These systems we’re building don’t just entertain.

They shape attention
They shape behavior
They shape mental health

Some of that cannot be reversed easily.
Some of that does not come with a takedown button with a penalty payment.

And yet we still act as if consequences are simply an operating expense.

At what point does doing the right thing supersede quarterly growth?
At what point does leadership mean accountability… not just velocity?

Responsible deployment isn’t anti-innovation.
It’s adult leadership.

Speed is a strategy when harm is reversible.
Speed is negligence when it isn’t.

So while this video is impressive, what does it tell us about our current leadership mindset?

No, you’re not behind. You’re just early and that takes a different mindset.

HBR released its agentic AI report. Here's what 623 executives told them:

🔹 90% expect most organizations in their industry will use agentic AI
🔹 84% believe it will transform their business

And yet…

🔹 Only 26% are effective at using AI for positive outcomes
🔹 Only 13% say their data architecture is very well equipped
🔹 Only 5% have well-defined success metrics

That's not a failure of technology. That's what early-stage innovation looks like.
The report segments organizations into leaders (34%), followers (40%), and laggards (26%). 

The more honest segment? Most have already started and are quietly underwhelmed. Here's why:

1️⃣ "We hired a vendor and what we got wasn't worth the price" 
There are no ten-year agentic AI veterans. Most firms selling "proven methodology" are packaging 12–18 months of experimentation. You're co-innovating. If can’t assume an ERP-style implementation right now.

2️⃣ "We paid for embedded AI and it's underwhelming" 
That tool is optimized for the vendor's median customer… not your processes, compliance posture, or institutional nuance. Without that definitional layer, agents infer. And inference is not the same as institutional knowledge.

3️⃣ "Our teams aren't getting more efficient" 
Only 47% of organizations say employees understand the value of agentic AI… dropping to 24% among laggards. It's not a training problem. It's an enablement problem. Training teaches usage. Enablement builds capability over time.

Among organizations already using agentic AI; 36% report greater productivity, 35% better decisions, 33% real cost savings. What separates the 34% of leaders?

-> Better data readiness
-> Stronger governance
-> Employees who understand what they’re working toward

Not bigger budgets. Fundamentals. Here's what moves the needle. Not by going bigger. By going narrower.

1️⃣ Smaller bets with a big vision. Enterprise ambition with task-level execution. Pick one process, one decision, one hand-off… and prove it there first. Autonomy is earned through evidence, not defined in a flow chart.

2️⃣ Clearer definitions. Write down what "correct" looks like before you build anything. Define the processes, decision logic, and institutional vocabulary agents need to do your work… not generic work. This is your knowledge layer. You need to build.

3️⃣ Instrumented learning. Where did it fail? What required human override? What improved… and why? What does that tell you about your data quality, your definitions, your process design? Measurement isn't for reporting. It's for knowing what to fix next.

4️⃣ Innovation discipline. Hypothesis. Tight scope. Guardrails. Iteration. Honest diagnostics. Governance built into the loop, not bolted on at the end. This isn't a deployment. It's a learning system. Treat it like one.

Transformation doesn't come from one big deployment. 
It comes from hundreds of instrumented learning cycles that compound.

Agentic AI, Expectations, Readiness, Results (HBR)

Six consultants. Five sprints. Six figures. A month past deadline and it still wasn’t live.

A leader pulled me aside last week and asked:… “Should we have gone long and narrow rather than short and deep?”

When I looked under the hood, what they’d built wasn’t really agentic AI. It was a traditional digital interface layered with rule-based logic… sprinkled with LLM. The complexity made it nearly impossible to see where things were breaking down.

The answer was yes. 

But the more important question is why smart teams keep making the same mistake.

They’re executing agentic AI initiatives the way they would a traditional digital transformation. 

Big team, fixed scope, predefined outcome. 

They hired a traditional implementation consultant. And the execution? A rinse and repeat of everything I've seen before, before AI.

But agentic AI isn't traditional digital transformation. It's innovation at scale.
And that requires a different execution muscle.

The recent MIT Sloan + BCG "Agentic Enterprise" report names why the old playbook keeps failing:

🔹 Scalability vs. Adaptability. Projects optimize for standardization. Agents handle ambiguity. Building for that ambiguity isn't sprint work, it's tuned over time from usage.

🔹 Supervision vs. Autonomy. Autonomy boundaries aren't known upfront. You don't start with it in full force. You iterate toward it.

🔹 Retrofit vs. Reengineer. Bolting AI onto old workflows accelerates poor design, and redesigning everything at once stalls. Redesign decision by decision. Task by task.

🔹 Expediency vs. Capability Building. Projects fund outputs. Agentic systems compound through learning. Optimize for quarterly ROI and you reset progress every time.

So, what does "long and narrow" actually mean?

1️⃣ Select for execution readiness, not excitement. Choose use cases with clearly defined processes, understood data, and decisions that can be isolated.

2️⃣ Decompose to the atomic task level. Not the workflow. Not the department. The smallest discrete decision: one set of inputs, one defined output.

3️⃣ Keep the team small and sharp. A true SME, a product architect, and someone hands-on who can build and iterate.

4️⃣ Build and test in real conditions. Observe how people actually use it. Real behavior reveals design flaws faster than workshops ever will.

5️⃣ Calibrate autonomy deliberately. Define guardrails through use. Identify where human judgment truly belongs. Autonomy is discovered through iteration... not declared upfront.

6️⃣ Iterate decision by decision. Refine. Adjust. Move to the next atomic task. Resist the urge to redesign everything at once.

7️⃣ Let redesign emerge. As tasks stack, patterns surface. The workflow begins to reveal what it wants to become.

You can’t scope a future that is unknown. That comes through disciplined iteration.

Big teams accelerate delivery.
Small teams accelerate understanding.

Agentic AI doesn’t reward speed of deployment.
It rewards speed of learning.

taplio

The Battle

Cameo Doran

Darlene Newman

Followers

Posts

Engagements

Likes

Performance ScoreCameo Doran's LinkedIn Performance

Performance ScoreDarlene Newman's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Cameo Doran

Darlene Newman

Followers

Posts

Engagements

Likes

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);}Cameo Doran's LinkedIn Performance

Performance ScoreDarlene Newman's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreCameo Doran's LinkedIn Performance