ActionableOps

ActionableOps is an AI consulting company that helps businesses unlock AI's potential to excel in their operations. We specialize in guiding companies through the AI landscape with services such as AI governance, Responsible AI, AI Policy development, AI Ethics, AI Training, and AI Implementation. Partnering with industry leaders like Textcortex and MindStudio, we ensure that our clients gain the knowledge and tools needed to thrive with AI.

Founder / AI Consulting

Kansas City AI Club

Primary Purpose: Foster awareness, adoption, and hands-on learning of the latest AI technologies by leading discussions, events, and workshops tailored to real-world business applications.

Key Responsibilities:
 - Technology Trends Leadership in AI
 - Curate and present emerging AI technologies (e.g., generative AI, voice agents, AI-driven automation).
 - Deliver accessible, business-focused insights on cutting-edge advancements.
 - Business Use Case Applications
 - Develop and showcase practical AI use cases (e.g., AI in customer service, operations, or marketing).

Workshops & Education
 - Organize hands-on workshops to teach members how to apply AI in real-world business settings.
 - Guide participants through implementation frameworks tailored to their industries.
 - Event Programming Support
 - Host and facilitate demos at Kansas City AI Club meetups.
 - Develop quarterly in-person or virtual workshop series, such as "AI Applications for Business Success."

Board Chair - Emerging Tech

UMKC Henry W. Bloch School of Management

The Regnier Institute Mentor in Residence Program connects seasoned mentors with student entrepreneurs in need of a helpful ear.

Mentor-In-Residence

National Speakers Association-Kansas City Chapter

Speaker

MindStudio

Agency Partner

RunWhen

RunWhen is a platform for troubleshooting complex tech stacks. It helps engineers identify when issues they are chasing lie in their app, their cloud-native platform or their vendors by answering the question: “What do I run and when?”

Advisor

Nobl9

Nobl9 is a service level objective SaaS platform that connects reliability metrics to business goals via a modern engineering workflow. Backed by CRV, Battery, Bonfire Ventures, Harmony Partners Cisco and ServiceNow.

Director of Operations

Code Ninjas

Code Ninjas is the world's largest and fastest growing in person children's franchise startup, that teaches kids how to code computers by building video games. Backed by Eagle Merchant Partners.

Director of Operations / Owner

OMT Brands

OMT Brands created a line of consumable pet products to be marketed and sold online through Amazon FBA and the Amazon, Ebay, Sears, and Walmart marketplaces.

Director of E-commerce Operations

City of Lakewood, Washington

Police Officer

🚀 Founder | Startup Operations Executive | Business Scaling Expert | Data Analytics | P&L Management | Process Optimization | Customer-Centric | Revenue Growth | Operational Efficiency | SaaS | Change Management

I’m Dale Thomas and here’s my story in a nutshell. 

- Authored the best-selling book “The Ultimate A.I. Guide for Small Business Success.” 
- Founded several technology businesses
- 20+ years of operations experience in Start-Ups and Enterprise (Ex - Microsoft) 
- Focused on helping SMB’s make sense of AI.  I specialize in AI agents, AI Consulting, responsible AI, ethical AI, policy creation, governance, and implementation of AI solutions. 

What you’ll get by following me: 
🔥Transparent & actionable AI advice
📆 Tips and Tricks Implementing AI Agents
🔍 Insights on implementing ethical AI practices
📢 Up-to-date information on AI regulations and legislation  

AI Partnerships
* TextCortex AI Partner - get a free account https://textcortex.com?via=ops 

A.I. Education
* MindStudio Certified AI Expert
* The London School of Economics - Ethics of AI 
* University of Pennsylvania - AI For Business
* Vanderbilt University - Prompt Engineering for ChatGPT
* MindStudio Certified AI Developer

Struggling to make sense of AI for your small to medium-sized business? 

I’m here to help - so hit that “Follow” button!  

Dale

Dale Thomas

David Arnoux

community-first @ www.thegenaicircle.com
sharing thoughts @ www.heyarnoux.com
building autonomous ventures @ www.humanoidz.ai

Co-CEO

Growth Tribe is here to empower adults to always be learning. Our solutions ensure people & organisations can acquire the digital skills they need to adapt, stay up-to-date and in the lead!

We’ve built a reputation for fast-paced, impactful & enjoyable learning, offering training & coaching in Data Science, Analytics, Artificial Intelligence, Growth & Marketing, Innovation, UX Design & Product Management. 

With campuses in 6 European countries, we have supported the learning journeys of over 16,000 adults in over 900 organisations. Our 80+ staff are dedicated to creating Europe’s best learning experiences, to empower humans to always be learning.

using AI for 1000+ days...canceled most of my AI subscriptions this month.
now i spend 90% of my time in one tool.
this is probably the best time investment you make this year...

yesterday's post about claude code/agents/skills did better than expected so i'm releasing the FULLER playbook. the tool? claude code. but i don’t code a single line. don't let the "code" in the name fool you… i'm not a developer and it works perfectly. 

here's what happened:
chatgpt, gemini, the tools I used etc forget everything. every chat starts from zero. you re-explain your project, your brand voice, your preferences... over and over. even with customgpts and claude projects it isn't reliable enough..

claude code changed EVERYTHING. you create a simple file called CLAUDE[.]md. that file becomes your AI's permanent memory. write your context once. never repeat yourself again.

but here's where it gets REALLY interesting:
you can build sub-agents. specialists that each have their own "brain folder." one for linkedin content. one for research. one for client work. each remembers everything.

then you connect them to your actual tools via MCP: hubspot, gmail, google analytics, notion, slack. they execute tasks on your behalf. i'd say this is probably the biggest practical shift in knowledge work since the chatgpt moment. and it's simple. it's basically just folders and markdown files. no actual coding required. true story...someone in our community said it best: "my clients think i hired a team."

anyway, i’m sharing and i've used the playbook from Iwo Szapar which covers it in perfect easy-to-follow details, step by step:

→ setting up claude code (even for non-coders) 
→ creating your CLAUDE.md file (your project's permanent memory) 
→ building sub-agents with their own brain folders 
→ connecting to hubspot, gmail, notion via MCP 
→ the mistakes that break the whole system 
→ ready-to-use templates you can copy

this might be the best hours you invest this year.

want it? you know the drill
will send it today if you’re fast (there are some limits on number of DMs per day…first come first serve:
1. 👍 like
2. 💬 comment "Second Brain"
3. 🔄 repost just to be sure

(make 100% sure we're connected)

thank me later for this one...

Helping GTM Leaders & Founders Grow With GTM x AI | Fractional CxO | Building Linkedin Tools @ humanoidz.ai

how I built a second brain that actually works

everyone tells you vibe-coding can't go into production.

they've never done this…

the criticism of vibe-coding is always the same: "cool prototype but can it handle real users?" "authentication?" "good luck deploying that." etc etc

i do fractional gtm with focus on ai and the age old question comes back every single time we identify a high-impact use case worth building out… stick to what we have, buy or build? Before 2022…we used to say "build isn't worth it." the cost of deploying and maintaining a custom tool was too high for most teams. the math completely changed in 2025, the ground underneath us has shifted… the gap between "demo" and "production" that we keep hearing about?? = infrastructure. auth systems. database setup. deployment pipelines. devops headaches. that gap is gone now. the question is more…is this use case worth our time?

let me explain.

you’ll have guessed by now if you follow me on LI… i’m a replit fan boy

replit’s built all of those typical blockers into the platform: 
→ authentication built-in 
→ database (no external config) 
→ integrations via connectors (just login to hubspot, slack, notion, google drive... no api keys, no oauth wiring) 
→ deployment 
→ security environment (default)

the "production gap" was never about code quality. it was about everything around the code. the scaffolding as I like to call it…

1 example of a production app shipped this year (SME/Scale-up):

→ ad asset generation tool for a medical firm strict compliance rules around health claims, disclaimers, and image usage off-the-shelf design tools had no way to enforce these guardrails so we built custom validation that flags non-compliant copy before export/ This is connected to their google drive via connectors so assets save directly to the right folders. legal team (2 freelancer) went from reviewing 100% of assets to spot-checking 10%. still running 6 months later. zero compliance issues since launch.

another example is an ai-powered signal detection tool for a vc fund. scrapes data sources their competitors don't have access to… Shows early traction signals before the usual databases used daily by the investment team

here's the mental model i use for buy vs build, it's simple, it's not totally clear cut, but it'll put you on the right direction and then you'll probably need an executive decision. 3 questions: 
→ is your workflow standard or custom? 
→ do you need it connected to your existing tools? 
→ are you paying for features you don't use?
(then you usually need an executive decision)

let's be honest about the hard parts: 
→ prompting well still takes iteration. expect 3-5 rounds of back and forth before it feels right. 
→ edge cases 
→ use cases that are worth it: platform handles infrastructure but you still need to know what you're building and why.

the skeptics aren't wrong that most prototypes die before production but it was never about the code. it was the infrastructure around it + is this use case actually worth it?

this stack is the fruit of 18 months of tinkering with 400+ AI nerds.

but wait, there's more....

here's what each tool does:
→ granola: takes meeting notes so you can actually be present 
→ wisprflow: talk instead of type (humans weren't meant for typing anyway) 
→ raycast: shortcuts and AI snippets everywhere 
→ viralbrain[.]ai: drop linkedin urls...put your linkedin growth on auto-pilot
→ claude: your thinking partner and writing brain
→ notebooklm: upload docs and have conversations with them 
→ perplexity: research that actually cites sources (but a few mistakes)
→ notion: where everything lives 
→ gamma: branded presentations in minutes 
→ n8n: automations that actually do stuff 
→ replit: build apps without the setup headache
→ a few more...

and at the center? claude with sub-agents, skills and claude code running like a bunch of extra brains.

anyway, i put together a full playbook with a 90-day challenge to actually build this into your workflow + some extra tools/goodies in there

want the full version? you know the drill: 
1. 👍 like 
2. 💬 comment "Augmented Humans" 
3. 🔄 repost just to be sure
(make 100% sure we're connected)

will send it within 1hr. DM me if you have questions. Also offering a 1-hr personal session to get you up and running

this is the closest i've ever felt to 'work' AGI...
I'd even say this is potentially the best investment of your time
when asked what's the 20/80 signal vs noise to focus on atm, I answer this:

claude code + sub-agents + skills... hear me out:

most people use claude like a chatbot. ask a question, get an answer, repeat.
i've been using it like a powerful team, that gets me...

here's what changed: claude now has skills and sub-agents. skills are basically recipe cards you create once, then claude follows them perfectly every time. no more rewriting the same prompts. no more inconsistent outputs.

sub-agents take it further. instead of one claude doing everything, you create specialists. one for linkedin content. one for data analysis. one for email drafting. each one remembers your preferences forever.

(someone in our community said "this replaced everything. i barely use any other tools anymore.")

the real magic? you can run 3-4 agents simultaneously. while one analyzes a spreadsheet, another writes a post about those numbers.

anyway, i put together a full playbook covering: 
→ setting up skills and enabling code execution in claude 
→ creating your first SKILL.md file (claude literally writes it for you) 
→ the debug loop that makes skills sharper over time 
→ building sub-agents with their own "brain" folders 
→ connecting agents to google drive, slack, hubspot via MCPs 
→ the packaging mistake that breaks everything

it's 20 pages of step-by-step stuff that took months to figure out.

want it? you know the drill: 
1. 👍 like 
2. 💬 comment "augmented" 
3. 🔄 repost just to be sure
will send it later today.. first come first serve as there are limits.

(make 100% sure we're connected)

DM me if you have questions.

Helping Founders & GTM Leaders Grow With GTM x AI | Fractional CxO | Building Linkedin Tools @ humanoidz.ai

Claude_Code_Claude_sub-agents_and_Claude_skills

this playbook should be illegal.

Decks are the worst part of my job. so i automated branded deck creation.

here's the problem: decks take forever. and keeping them on-brand is basically impossible when you've got multiple people making them. 
couple weeks ago i was introduced to claude code during one of our online community braintrusts. then gamma released their api. that created magic. this playbook connects Claude Code to Gamma's API. you feed Claude your brand voice doc once. then it writes all the content in your tone. Gamma handles the design automatically. 
no more starting from scratch. no more fixing fonts and colors for an hour. just consistent, on-brand decks every time. 

anyway, it's 11 step by step pages covering: 
→ setting up Claude Code and Gamma API access 
→ teaching Claude your brand voice so it actually sounds like you 
→ generating slide content as structured JSON 
→ calling Gamma's API to auto-design the full deck 
→ the common mistakes that break the whole thing

the whole system basically turns a 5-hour deck into a 15-minute review.

want the full version? you know the drill: 
👍 like 
💬 comment "branded decks" 
🔄 repost just to be sure
(make 100% sure we're connected)

will send it right away. DM me if you have questions.

(thx a bunch to Iwo Szapar for motivating me to take up claude code…)

Auto-Generate Branded Decks with Claude Code and Gamma API

I run a community with some of the world's best ai marketers (and more)

this is a playbook that works $$$. shared secretly. sharing it with you today.

distribution is everything nowadays. outreach still works, but only when done right. what doesn’t work: start outbound with cold lists. buying random contacts, blasting emails, hoping something sticks.

but you already have warm prospects sitting in your CRM. website visitors. ad engagers. people who've actually shown interest.

this playbook flips the whole approach. you run a linkedin "sensor" campaign first to see who actually engages. tag them as hot, warm, or low. then only email the people who showed signals.

way better than spray and pray.

anyway, it's 18 pages covering:
→ splitting your CRM into core vs GTM workspaces
→ setting up a safe outbound subdomain
→ enriching data in Clay
→ building matched audiences across linkedin, google, meta
→ the sensor campaign trick
→ personalized sequences that don't feel like spam

the whole warm-first system basically.
want it?
👍 like
💬 comment "warm pipeline"
🔄 repost just to be sure 

(make 100% sure we're connected)

will send it right away. DM me if you have questions

(thx a bunch to Shaun Thresher for his beautiful playbook)

Founder @ humanoidz | Fractional CxO (AI/GTM)

Build a Warm-First Outbound Pipeline 

6 months, 3 months ago i'd never thought I'd be posting this

But i did something last week i never thought i'd do.

canceled my $200/month ChatGPT Pro account.

3-6 months ago you couldn't have paid me to switch. i was all in on OpenAI. but here's the thing now...i haven't missed it once.

it's all Claude and Gemini now. Claude for the deep thinking work, building stuff with Claude Code, actual system design. Gemini for the speed when i need quick answers + plugged it onto most of our apps/websites. and honestly? the combo works better than ChatGPT Pro ever did. it feels like AI is out of "beta" now.

(and yeah, i know Sam Altman is apparently in code red mode trying to figure this out. makes sense, a lot of people I talk to are making the same switch)

the interesting part is this isn't just me. saw Walid Boulanouar's post asking if people would jump ship if Gemini adds full MCP support to custom GPTs. the comments are basically everyone saying they already left or they're on the edge. Damian Nomura mentioned ChatGPT has been "gathering dust on his virtual shelf" for a while. some pointed out Gemini CLI already has full MCP support. that's the pattern i'm seeing everywhere. people aren't waiting around anymore. anyway, saved $200/month and got better results. sometimes the obvious move takes a while to make...

will this post age like milk in a couple weeks/months?

just got this from my accountant.

she mentioned...

my subscriptions might be getting out of hand

(note: in ~6 months time it'll be 100% google/gemini anyway)

entrepreneurship might actually be safer than a corporate job now.

So i've been looking at the data on entry-level jobs lately and something weird jumped out.

tech layoffs hit 264,000 workers in 2023. another 150000+ in 2024. and 65% of companies use "last in, first out" policies. meaning if you're entry-level, you're first to go. doesn't matter how good you are. here's a public example: emily pitcher at Meta got "exceeds expectations" on all her performance markers. two months later? laid off. made forbes 30 under 30 and still couldn't land another corporate job. entry-level job postings dropped 35% since january 2023. median tenure for ages 20-24 is just 1.4 years.

meanwhile the "90% of businesses fail" stat everyone quotes? completely false.
actual data: 80% survive year one. 50% still operating at year five. side hustles? only 3% report failing. compare that to a 65% chance of being laid off first in any downturn. stanford just found that employment for workers 22-25 in ai-exposed roles dropped 13% since 2022. older workers in those same roles? up 6-9%. we know now. ai eats entry-level work first. it's the routine stuff. exactly what juniors do. goldman considered cutting junior hires by two-thirds. microsoft writes 30% of code with ai now. salesforce cut support from 9000 to 5000 using agents and startup costs collapsed. 43% of businesses now start with under $10k. what used to need $100k and a tech co-founder now takes a laptop and some evenings.

The math changed somewhere in the last few years. corporate entry-level: 35% job posting decline, 2.7-year tenure, 3-year wait for promotion, first out in layoffs. entrepreneurship: 80% survival rate, 3-6 month profitability, skills that compound, multiple income streams.

not saying quit your job tomorrow.

but if you're early career and thinking about your next decade?

the "safe" path isn't what it used to be.

thanks to Sebastien Diaz de Corcuera for triggering this reflection.

we were prmised agi before year end

and what we got will surprise you…

a slight delay in timeline

graphical user interface, text, application, chat or text message

gemini 3 is insane

I'm building something new and exceptional...

I'm building a new shed/house for our two donkeys

This is 1 shot...all the steps and measurements are spot on.

wow.

apparently your forties suck

only issue I have with this meta analysis is mixing up happiness and life satisfaction (those are 2 distinct things)

happiness declines due to career and financial pressures peaking, sandwich generation stress caring for children and aging parents, unmet youthful expectations creating disappointment, relationship strains, health concerns emerging, existential questioning about life choices + the gap between aspirations and reality being at its widest​​​​​​​​​​​​​​​​

apart from that 

enjoy the day!

a secret weapon i've used for years: Analogies.

the reason they're so powerful and a prompt that I use...

your brain already has mental models built. analogies just borrow them. instead of building something from scratch in someone's head, when you use an analogy, you're remixing something that already exists there. analogies are basically just saying "this new thing is like this thing you already understand." that's it. you're borrowing someone's existing mental model instead of building a new one from scratch. This has been around for ages. there's an ancient rhetorical exercise called Chreia that was taught for over a thousand years in greco-roman schools. basically, it was how they trained the greatest orators, philosophers and writers. and one of the core elements? arguing by analogy. Today it's used by the best writers, communicators, influencers. Pick up any 'airport' book and it usually starts with one big Chreia or analogy.

so use them... in your pitches, in your prompts, in your decks, in all your comms.

two cool examples that rewire our brain and that we get in seconds:
- the 'domino theory': the idea that if one country fell to communism, others would follow like dominos...this literally shaped american foreign policy for decades. (it was wrong, but that's the power of a sticky analogy. they can move armies.)
- "it's like uber but for X" became a meme for a reason. it's lazy now, sure. but it works because in one sentence you transfer an entire business model, user behavior and market opportunity into someone's brain.

Here's a prompt you can use (I use a variation of this)


[ I'm explaining [topic]. give me two analogies from history or science that make this concept click - one that's familiar and obvious, one that's lesser-known but surprisingly accurate. avoid exaggerated comparisons. ground them in real events or facts people can verify. ]


in an ai world, we're all becoming editors in chief, curators. the raw material is everywhere. what matters is how you curate, connect, and frame ideas.
analogies are your secret weapon for this.

killer robots on the ukrainian front + claude code can train an open llm.

two stories that will define 2026, but NOT for the reasons you think...

historians have this concept called "the hinge of history"... these are basically moments where everything pivots but nobody living through them realizes it. they're (we are) too close. it just feels like another tuesday. I think we might be in one right now. then again it feels like that every week. it's not about the tech being scary (although the killer robot thing is getting way too real and china's not even involved yet). it's about the speed. the gap between "we have a few years before this happens" and "oh that happened last tuesday" has basically collapsed. 

so here's what actually happened. december 10th, a ukrainian ground robot destroyed a russian armored vehicle full of troops. first confirmed kill by a combat robot in history. another one held an infantry position completely alone for six weeks = no soldiers, just a tracked machine with a .50 cal doing daily patrols and suppressing russian advances. ukraine's deploying a bunch of these... russia has them too. robot-on-robot combat is here...

same week, julien chaumond (cto of Hugging Face) casually posts that they got Claude Code to train an open llm. not just write the training script, actually submit jobs to cloud gpus, monitor progress in real-time, and push finished models to the hub. you say "fine-tune Qwen3-0.6B on this dataset" and it handles everything else. ai training ai. autonomously.

both of these stories came and went. killer robots got maybe a day of coverage. autonomous model training got some linkedin engagement. then we all scrolled on. we're so saturated with "insane ai news" that the actual arrival of the future just blends into the feed. i think 2026 is the year this catches up with us. not because anything dramatically new will happen... but because the stuff that's already here will start compounding. agents that build agents. robots that fight robots. and a whole lot of industries still pretending they have time to figure it out.

anyway, something to think about with your morning coffee 😳☕️

erik brynjolfsson ran a crazy study in 2019

you won’t guess what he found...

he found that the real value of google in 2019 was $17500 a year. he didn’t ask how much they’d pay for Google… he flipped the question. what would we have to pay you to give it up entirely? in 2019, turns out the answer was around $17500 a year. for something that’s free. the method is called measuring “consumer surplus” which basically reveals how much value we get from free digital services that never shows up in gdp or traditional economics (joni mitchell was right about not “knowing what you’ve got till it’s gone”). his whole point was that we massively undercount the value of the digital economy because so much of it costs nothing.

I’ve been thinking about this with LLMs (Claude, Gemini, chatGPT). I pay maybe 600 euro a month for all on the max plans (excluding my use of APIs). but if you asked what you’d have to pay me to stop using them completely? €2000, €3000 a month in real value. maybe more.

anyway, genuine question… what would someone have to pay you to give up your LLM access entirely? gut feeling, first answer that comes to mind

ps: yes yes, what happens if we can’t keep subsidizing these models and the whole thing crumbles? funny thing is, even the people I see spewing hate at LLMs, the ai bubble etc on linkedin are clearly using them to build their images and write their posts. the irony isn’t lost on me.

pps: thanks to Pascal Malengrez for his beautiful photography and to Gemini for the makeover

letting agents run wild without controls isn't innovation... it's a security incident waiting to hit your inbox.

so how do you scale agentic ai without it turning into a compliance nightmare. after doing many of these, the answer is super straightforward: security by design.
"if you start on a real platform, the GUARDRAILS come baked in. you don't bolt them on later." you wire governance, security, and visibility straight into the core platform.

✅ every single model interaction gets scanned for prompt injection tricks = both what goes out AND what comes back in.

✅ agents only see what the human behind them is allowed to see = delegated access locks them to the right data and nothing else.

✅ weird behavior gets flagged automatically = anomaly detection catches compromised or drifting agents across hundreds of connected apps.

✅ every reasoning hop, every action, every call is logged so you can replay exactly what happened and why.

that is how teams graduate from toy ai demos to AI YOU CAN ACTUALLY TRUST.

because if governance isn't built in from day one, your smartest agent can quietly turn into the weakest link in your whole stack.

it is possible to ship ai fast AND stay in control at the same time.

now i'm curious:
are today's enterprises really designing ai to be safe from the start or are most still waiting to learn after something breaks in production?

you know what? using ai at work is way simpler now than pretending you're 'not technical'.

instead of saying 'i need to learn ai someday', start saying 'i use ai to handle the boring 80% so i can focus on the important 20%'.

here's what that looks like in real life:

- drop a messy email thread into your ai tool and ask: "summarize the key decisions and next steps in 5 bullets."

- paste your rough notes and say: "turn this into a clear, no-fluff update for my manager. keep it under 150 words."

- give it your meeting agenda and ask: "suggest smarter questions i should ask in this meeting based on this context."

seriously, you'll look back and wonder why you were manually grinding through all of this.

these days, the people who quietly use ai for drafts, summaries, ideas and structure are the ones who look insanely organized and fast.

most of them are not 'technical' at all.

they just stopped treating ai like a sci-fi project and started treating it like a slightly over-caffeinated intern.

sometimes, a tiny workflow tweak like this is all it takes to flip your whole workday.

trust me, you'll be glad you start letting ai take the first pass on your everyday tasks.

taplio

The Battle

Dale Thomas

David Arnoux

Followers

Posts

Engagements

Likes

Performance ScoreDavid Arnoux's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Dale Thomas

David Arnoux

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);}David Arnoux's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreDavid Arnoux's LinkedIn Performance