Almost Timely News: Terraforming the AI Use Case Desert

How to help AI take more of your job (the parts you don't want)

Mar 29, 2026

Almost Timely News: 🗞️ Terraforming the AI Use Case Desert (2026-03-29) :: View in Browser

The Big Plug

👉 I’ve got a new course! GEO 101 for Marketers.

👉 Just updated! The Unofficial LinkedIn Algorithm Guide, March 2026, now with new information straight from LinkedIn!

Content Authenticity Statement

95% of this week’s newsletter content was originated by me, the human. You’ll see output from Openwork and Minimax M2.7 in the opening section as demo output. Learn why this kind of disclosure is a good idea and might be required for anyone doing business in any capacity with the EU in the near future.

Watch This Newsletter On YouTube 📺

Click here for the video 📺 version of this newsletter on YouTube »

Click here for an MP3 audio 🎧 only version »

What’s On My Mind: Terraforming the AI Use Case Desert

In this week’s newsletter, let’s talk about something that truly baffles me: the AI use case desert. I read with interest a survey recently that said one of the top reasons for lack of AI adoption was lack of use cases.

Boy did that make me scratch my head. I’m in the opposite camp; there are SO many use cases for generative AI that there’s no way I can tackle more than a small fraction of them. Deciding which use cases make the most sense to tackle is the hardest part of any client engagement because there’s so many. Even putting together this newsletter every week is less a question of “what should I write about?” and more “which of the dozen major developments this week is worth demonstrating in some way?”

What’s behind this chasm? A few things - lack of understanding about what the technology is capable of today, lack of understanding about where AI ends and infrastructure begins, lack of knowledge about what data you have access to, and lack of imagination.

If these sound familiar, it’s because they’re part of literally every keynote talk and workshop I’ve delivered over the past 5 years. What sets apart successful folks from less successful folks when it comes to generative AI is critical thinking, creative thinking, and contextual thinking.

So let’s dig into how to terraform the use case desert into the use case garden - using AI, of course.

Part 1: The Sexy Block

The top mental block for a lot of folks - consultants especially - is that most use cases for AI are going to be boring. Consultants - and I count myself in this group - often need to show flashy and sexy because that’s what pays the bills. A senior leader sees “the art of the possible” with a big flashy show and wants to sign right away, but very often the real value is in the boring stuff.

This is compounded by the fact that the majority of AI use cases are optimization, not innovation. I define this as doing what you’ve always done, but bigger, better, faster, and cheaper - and most folks focus on the faster and cheaper. There’s a maxim in business strategy that you can’t cut your way to growth, and that holds true even in the age of AI. Yes, absolutely reduce the amount of time you spend on less valuable tasks, but if that’s all you’re doing, you’re more or less rearranging the deck chairs on the Titanic.

Value comes from optimization to clear up space and time, then innovation - doing something new, doing something out of your normal skill set personally and organizationally, something that adds value in a new and different way.

The often silly example I use is taking something traditionally boring, like an RFP response, and turning it into a country song. I, as a non-musical person, can generate reasonably good country music from the most unusual sources. And to be clear, this takes away no paying work from human musicians, because I never even thought to do this before. I’m not taking a task from a human and giving it to a machine.

But that’s the difference between optimization and innovation - doing what you’ve always done versus doing something you’ve never done before.

Part 2: The ROI Block

The second major block, typically from more conservative or even resistant folks, is the familiar refrain asking for the ROI of something like an AI initiative. I’ve gone on long rants about this in the past on two fronts - first, people using the phrase ROI instead of results, because ROI is fundamentally a financial calculation. Second, people asking to measure ROI on something they don’t measure today, so they have no basis of comparison. Both are silly.

Here’s the reality of ROI. Because it’s a financial formula - (earned - spent) / spent - it takes some time to understand and calculate what you spent and what you earned. And in a field like generative AI, where the technologies change literally every week, by the time you complete a project, there’s a good chance that the old project methods and ways are obsolete. That in turns makes an ROI measurement less valuable, because part of the purpose of measurement is to achieve repeatable, reliable results, and you can’t do that if everything is changing.

Imagine you started a project with the equivalent of a campfire and some sticks. A quarter of the way through the meal, someone invents metal and now you have a cooking pot and a fire. Halfway through the meal prep, someone invents natural gas capture and metal manufacturing, and now you have a gas cooktop and cooking pots and pans. Three quarters of the way through the project, someone invents rooftop solar and induction ranges. By the time you’re done with a standard 2-year enterprise project, literally everything is different when it comes to AI - and if you’ve been adapting, then what you started with can’t be used as a measure of where things are now.

Most of the time, when people are thinking about ROI, what they really mean is measuring change, and that’s a lot more immediate. It’s much easier to measure change. The same general formula applies - (new - old) / old - but instead of money, it can be any quantifiable unit. It can be time, leads generated, NPS numbers, anything where there’s new numbers and old numbers.

Especially for a fast-moving technology like AI, this is a better system of measurement. It’s not ROI, but it is reasonably objective and defensible - and because the cycles of feedback can be shorter than the entire lifecycle of something (which ROI requires), you can know the impact of AI faster.

Part 3: Decomposing the Job Description

A couple weeks ago, we talked about how to determine the level of exposure in your job from AI by decomposing a job description. The exact same process applies for use cases - and in fact, much of the prompt is reusable, including using the Trust Insights TRIPS Framework. But rather than judge how exposed our job is, we pivot it to focus on what the most useful use cases are for us.

The Trust Insights TRIPS Framework is:

Time: How much time does a task take? The more time a task takes, the better a candidate it could be for AI.
Repetitiveness: How repetitious is a task? The more repetitive a task is, the better it might be for AI to handle it.
Importance: How economically important is a task? The more valuable a task is economically, the more sense it makes to have AI do that task. Assuming the same quality of output but dramatically increased speed, AI can add to the bottom line faster for more economically important tasks.
Pain: How painful is a task? The more painful a task is, the better candidate it is for AI. AI adoption goes up as painful tasks go down.
Sufficient Data: How many examples do you have of successful task outcomes? The more success examples and data you can provide, the better AI will do at a task.

Here’s how. Taking your current job description or perhaps an aspirational job description as a plain text file, combine it with a great big honking prompt like this and the TRIPS Framework in the agentic AI tool of your choice (e.g. Claude Cowork, Claude Code, OpenCode, Openwork, Google Antigravity, OpenAI Codex, Qwen Code, Microsoft GitHub Co-Pilot):

You’re a career and HR consulting expert. Based on the job descriptions in {your text file name}, infer a list of the top 25-100 individual tasks a person in this job is most likely to perform, in descending order of frequency. We define a task as a singular work output. An example of a task is creating a weekly report. Part of your analysis should include 3-7 broad categories of tasks from the job descriptions; this field should be named task_category. Format your response as YAML array with the task name, task category, a concise one sentence description of the task, the percentage frequency that the task occurs weekly in that job, and five columns from the Trust Insights TRIPS framework, scoring the task on an integer score of 1-5 for each of the 5 dimensions, inferring scores based on the job description and your background knowledge. Create an averaged TRIPS overall score as well called trips_average, averaging the 5 component fields. Use your web search tools if you are not confident in your estimates. Return your result as a YAML array and write it to the output folder as trips.yaml. If there are more than 50 tasks, truncate at 50. Then using the existing YAML array, augment it with a 0-10 integer score estimate of how likely it is generative AI is capable of performing the task with an 80% success rate; this should include regular generative AI and agentic AI system (examples: Claude Code, OpenClaw, etc.; use your web search tools to obtain any additional needed background context). Name this field ai_likelihood. For context, the current top scores on generative AI evaluations as of 2026 are: METR 50%: 12 hours. METR 80%: 90 minutes. GDPval-AA (ELO-500/2000): 58%. Terminal-Bench Hard: 58%. Tau-2 Bench Telecom: 98%. AA-LCR: 74%. AA-Omniscience: 55%. Humanity’s Last Exam: 45%. GPQA Diamond: 94%. Scicode: 59%. IFBench: 83%. CritPT: 30%. MMMU Pro: 82%. Remote Labor Index: 4.17%. Generative AI includes text generation, image generation, audio and video generation, tool use, and agentic harnesses. Infer your score based on the task description and TRIPS scores, paying special attention to repetitiveness and sufficient data as significant indicators that a task can be successfully performed by AI. Use your web search tools if you are not confident in your estimates. Use a scratchpad to reason with, then conclude from your reasoning what the score would be. Amend the YAML array with your scores and update output/trips.yaml. Do not include the scratchpad in your scores. Finally, based on TRIPS average scores and success rates, create a scaled average of ai_likelihood and trips_average (trips_average will need to be rescaled 0-10 to match ai_likelihood) named ai_use_case_average, extract out the top 5 best candidate use cases for generative AI (including agentic AI) in descending order by ai_use_case_average. With these 5 candidates, create a new YAML output file called output/tasks.yaml that contains the task name, task category, task summary, percentage frequency, TRIPS scores, ai_likelihood, and a three part explanation of how to implement generative AI and AI agents for the task - strategy, tactics, and execution plan (why, what, and how). Never return tables; always return valid YAML.

Because this is a series of complex tasks, you’ll get the best results from agentic tools rather than regular chat tools. If you use regular chat tools (like ChatGPT), you’ll want to do this in individual pieces rather than all at once. If you’re using an agentic tool, be sure to put in Plan mode first.

What you should end up with is a detailed set of explanations about the conversion of those tasks into generative AI use cases. It’s balanced between necessity - the TRIPS score average - and capability, the ai_likelihood average. Once you’ve got the 5 use cases in a YAML file, you can then use your favorite AI agent to start building a more thorough implementation plan (such as with the 5P Framework by Trust Insights) to bring it to life.

Part 4: An Example Brought to Life

Let’s look at a practical example of this. Say you’re the food and beverage manager at a restaurant chain, and the restaurant company is running into tough times. Costs are up, diners are down, and economic pressures are all around. Your job is a tough one, trying to create great culinary experiences that are still profitable for the company.

The challenge you face is that there’s so much administrative work in the role that you don’t spend nearly enough time on the floor to see the customer experience up close. So how could we use this process? We start with the job description; I’ll use one I found on LinkedIn to decompose the process overall.

For today’s build, I’ll use OpenWork, an open source competitor to Claude Cowork, and Minimax M2.7 on the Plus token plan as the AI provider. You can, of course, use the system of your choice, but these two represent the best for the budget-conscious. Minimax’s plan is $200 per YEAR versus Claude’s $200 per MONTH, and OpenWork integrates seamlessly with it.

We put the giant honking prompt in to get all the research and documents. Then, as we often do with agentic systems, especially for research tasks, we walk away and let it do its thing. Once it comes back with the results, we can start figuring out implementation.

I recommend watching the video version of this newsletter to see this in action.

Here’s what Minimax M2.7 returned:

  - task_name: "Analyze revenue and cost performance metrics"
    task_category: "Financial Management"
    task_summary: "Review financial reports, identify trends, and recommend actions to improve profitability."
    weekly_frequency: 30
    trips_average: 4.0
    ai_likelihood: 8
    ai_use_case_average: 8.0
    implementation:
      strategy: |
        This task is highly repetitive (monthly/weekly cycles), data-rich (historical financial data),
        and economically critical (directly impacts profitability). The pain score is moderate because
        manual analysis is time-consuming. AI excels at pattern recognition across large datasets
        and can surface insights that humans might miss. Automating this frees the director to
        focus on strategic decisions rather than data compilation.
      tactics: |
        Deploy an agentic AI system with access to POS data, inventory systems, and financial reports.
        Configure the agent to generate weekly variance analyses with natural language explanations.
        Set up automated data pipelines from Delphi (property management), inventory systems, and
        accounting software. Use prompt engineering to ensure AI outputs are formatted for executive
        presentation. Implement human review checkpoints before distribution.
      execution: |
        1. Select a platform: Microsoft Copilot Agents, Claude Code with data tools, or a custom
           LangChain agent with hotel industry connectors.
        2. Build data connectors: API integrations with POS (Toast, Square), inventory (MarketMan),
           and accounting systems (QuickBooks, Sage).
        3. Create prompt templates: Standardized weekly/monthly P&L analysis, year-over-year comparisons,
           segment performance breakdowns (restaurants, bars, events).
        4. Establish review workflow: AI drafts reports, F&B Director reviews and approves,
           final report distributed to leadership.
        5. Train the model: Feed 12+ months of historical data for baseline patterns. Use
           few-shot learning for industry-specific terminology and brand voice.
        6. Start with advisory mode: AI suggests insights, human makes final recommendations.
           Gradually increase AI autonomy as confidence builds.

I had it synthesize all the data for a Korean fried chicken restaurant and follow this process, generating a report for a stakeholder, and it came up with a really solid output - a financial report with forecast.

You could take the tasks YAML file and have your AI agent convert it into a lovely deliverable, like a pitch deck to your boss to explain why they should invest in AI. You could also feed it directly to your agentic system of choice and instruct it to start coming up with implementation scripts. Really, the sky’s the limit there.

Part 5: Wrapping Up

This is how we terraform the use case desert. We don’t wait for someone to hand us use cases on a platter, because by the time someone does that, it’s probably too late. The purpose of this exercise is to get ahead of the machines and the managers who don’t understand what can and can’t be done by AI. We want to be the ones running the machines, instead of someone making shortsighted decisions about what they think AI should be able to do based on business magazines and LinkedIn hustle bro scam artists. Yes, those folks who take terrible advice will in all likelihood crater their companies, but we’re more interested in protecting and advancing our careers.

By using today’s most powerful agentic tools - and in the video version, you’ll see how easy Openwork is (it’s very similar to Claude Cowork at 1/12th the price for the AI part with Minimax), you’ll start to create use cases for AI that really matter, that are connected to outcomes you get paid for.

Shameless plug: want help terraforming your company’s specific AI use cases in more detail than just this newsletter? My company, Trust Insights, does exactly that. Click here to learn how.

How Was This Issue?

Rate this week’s newsletter issue with a single click/tap. Your feedback over time helps me figure out what content to create for you.

Here’s The Unsubscribe

It took me a while to find a convenient way to link it up, but here’s how to get to the unsubscribe.

[](https://almosttimely.substack.com/action/disable_email?utm_source=almost-timely-newsletter&utm_medium=email&utm_campaign=almost-timely-2024- 12-28)

If you don’t see anything, here’s the text link to copy and paste:

https://almosttimely.substack.com/action/disable_email

Share With a Friend or Colleague

Please share this newsletter with two other people.

Send this URL to your friends/colleagues:

https://www.christopherspenn.com/newsletter

For enrolled subscribers on Substack, there are referral rewards if you refer 100, 200, or 300 other readers. Visit the Leaderboard here.

ICYMI: In Case You Missed It

Here’s content from the last week in case things fell through the cracks:

On The Tubes

Here’s what debuted on my YouTube channel this week:

Skill Up With Classes

These are just a few of the classes I have available over at the Trust Insights website that you can take.

Premium

Free

Advertisement: New GEO 101 Course

When I talk to folks like you, being recommended by AI is one of your top marketing concerns in 2026.

We’ve taken everything we’ve learned from OpenAI’s documentation, Google’s technical papers, patents, sample code, plus our years of experience in generative AI to assemble a high-impact 90-minute course on GEO 101 for Marketers.

In this course, you’ll learn:

The three distinct phases of GEO and how they work
How to optimize for each phase (they’re different!)
How to measure your GEO efforts in a meaningful and valid way

This course is meant to be used. In addition to the course itself, you’ll also receive:

Your 90 day GEO action plan
How to set up Google Analytics for measuring GEO traffic
How to join Google Search Console data with GEO intent data
How to use our free AIView tool to improve your content and site for one of the three phases of GEO
A certificate of completion from TrustInsights.ai

And best of all, this is our most affordable course yet. GEO 101 for Marketers is USD 99 and is available today.

👉 Enroll here in GEO 101 for Marketers!

Get Back To Work!

Folks who post jobs in the free Analytics for Marketers Slack community may have those jobs shared here, too. If you’re looking for work, check out these recent open positions, and check out the Slack group for the comprehensive list.

Advertisement: My AI Book!

In Almost Timeless, generative AI expert Christopher Penn provides the definitive playbook. Drawing on 18 months of in-the-trenches work and insights from thousands of real-world questions, Penn distills the noise into 48 foundational principles-durable mental models that give you a more permanent, strategic understanding of this transformative technology.

In this book, you will learn to:

Master the Machine: Finally understand why AI acts like a “brilliant but forgetful intern” and turn its quirks into your greatest strength.
Deploy the Playbook: Move from theory to practice with frameworks for driving real, measurable business value with AI.
Secure Your Human Advantage: Discover why your creativity, judgment, and ethics are more valuable than ever-and how to leverage them to win.

Stop feeling overwhelmed. Start leading with confidence. By the time you finish Almost Timeless, you won’t just know what to do; you will understand why you are doing it. And in an age of constant change, that understanding is the only real competitive advantage.

👉 Order your copy of Almost Timeless: 48 Foundation Principles of Generative AI today!

How to Stay in Touch

Let’s make sure we’re connected in the places it suits you best. Here’s where you can find different content:

My blog - daily videos, blog posts, and podcast episodes
My YouTube channel - daily videos, conference talks, and all things video
My company, Trust Insights - marketing analytics help
My podcast, Marketing over Coffee - weekly episodes of what’s worth noting in marketing
My second podcast, In-Ear Insights - the Trust Insights weekly podcast focused on data and analytics
On Bluesky - random personal stuff and chaos
On LinkedIn - daily videos and news
On Instagram - personal photos and travels
My free Slack discussion forum, Analytics for Marketers - open conversations about marketing and analytics

Listen to my theme song as a new single:

Advertisement: Ukraine 🇺🇦 Humanitarian Fund

The war to free Ukraine continues. If you’d like to support humanitarian efforts in Ukraine, the Ukrainian government has set up a special portal, United24, to help make contributing easy. The effort to free Ukraine from Russia’s illegal invasion needs your ongoing support.

👉 Donate today to the Ukraine Humanitarian Relief Fund »

Events I’ll Be At

Here are the public events where I’m speaking and attending. Say hi if you’re at an event also:

SSI, Charlotte, April 2026
The Trust Insights Generative AI Workshop, sometime this spring!
SMPS AI Conference, Austin, November 2026
MarketingProfs B2B Forum, Boston, November 2026

There are also private events that aren’t open to the public.

If you’re an event organizer, let me help your event shine. Visit my speaking page for more details.

Can’t be at an event? Stop by my private Slack group instead, Analytics for Marketers.

Required Disclosures

Events with links have purchased sponsorships in this newsletter and as a result, I receive direct financial compensation for promoting them.

Advertisements in this newsletter have paid to be promoted, and as a result, I receive direct financial compensation for promoting them.

My company, Trust Insights, maintains business partnerships with companies including, but not limited to, IBM, Cisco Systems, Amazon, Talkwalker, MarketingProfs, MarketMuse, Agorapulse, Hubspot, Informa, Demandbase, The Marketing AI Institute, and others. While links shared from partners are not explicit endorsements, nor do they directly financially benefit Trust Insights, a commercial relationship exists for which Trust Insights may receive indirect financial benefit, and thus I may receive indirect financial benefit from them as well.

Thank You

Thanks for subscribing and reading this far. I appreciate it. As always, thank you for your support, your attention, and your kindness.

Please share this newsletter with two other people.

See you next week,

Christopher S. Penn

Discussion about this post

Ready for more?