School’s out! And if this past year taught us anything, it’s that the final exam wasn’t about what we learned. It was about what we were willing to forget.
Every team building with AI right now carries habits that made perfect sense six months ago. Thorough prompts. Careful research before every experiment. Those instincts felt responsible. They also quietly became the ceiling.
At Slingshot, we asked our team to reflect on the AI habits they let go of this year. Not the new tricks they picked up, but the old ones they actively unlearned. The answers had a surprising throughline: almost every breakthrough came from trusting the tools more and overthinking less.
Summary
The breakthroughs came from trusting the tools more and overthinking less: shorter prompts, raw materials instead of pre-chewed context, and editing skills that matter more than prompting skills. The biggest shift was moving from one-off prompts to reusable systems and bringing AI into higher-stakes decisions, not just safe, scoped tasks. And the team that got the most from AI was the one that stopped giving the tools all the credit and started taking ownership of the work.
1. Stop Pre-Chewing the Problem
There’s an instinct to set the stage before asking AI for help: spell out all your goals, constraints, history, and edge cases. It’s the responsible thing to do, especially when the stakes are high and the tool is still relatively new.
“Early on, I’d write paragraphs of context and pre-chew the problem before I even got to the ask,” said Sarah Bhatia, Director of AI Product Innovation. “It felt thorough. It was actually just slowing me down. The setup work didn’t improve the output. It was making me a bottleneck.”
That instinct comes from treating AI as a new hire who needs the full download before doing anything useful. But the tools have changed. Hand over the raw material, state what you need, and let the conversation fill the gaps. If supplementary context is required, AI will ask. Most of the time, it doesn’t need to.
David Galownia, Slingshot’s CEO, noticed the same pattern from a different angle. He described dropping what he called “permission and politeness scaffolding,” along with meta-questions about capability. The prompts got shorter, the answers got sharper. The habit of over-preparing was protecting comfort, not quality.
2. Stop Asking If It’s Possible. Just Try.
Most professionals learn to scope before they act; understand the tool, map the constraints, confirm feasibility, then execute. In almost every other context, that discipline is a strength. With AI, it’s a speed trap. Tools change so quickly that by the time you’ve researched last month’s model, this month’s one already does it differently.
“The cost of trying is basically zero,” Sarah said. “And the answer I get from actually attempting is sharper than the answer I’d get from reading about it. If it can’t do something, it’ll say so. Turns out the hesitation was the bigger bottleneck.”
Savannah Cherry, Director of Marketing, described a similar shift: “I used to ask things like, ‘How do I upload a video and get a transcript?’ Now I don’t ask what’s possible; I just throw things at it and see what happens.”
And the behavior compounds. Once you stop gatekeeping yourself on what the tool can handle, the range of problems you’re willing to bring to it expands fast. That’s where the real acceleration starts.
3. The Real Skill Is Editing, Not Prompting
Job postings now require “prompt engineering,” and LinkedIn is flooded with prompting frameworks. Companies have invested in training to improve instructional writing. Although not entirely wrong, this focus overlooks the other half of the workflow.
“I used to think the magic was in writing the perfect prompt, like there was some secret combination of words that would produce finished copy,” said Whitney Powell, Slingshot’s Sales and Marketing Coordinator. “AI is never going to hand you the final version, and chasing that is a trap. The prompt gets you in the building, but your judgment gets you done.”
Sarah echoed this from her own workflow: “The first output is a starting block now. The real value shows up in the iteration: punchier, shorter, try it in my voice, give me three more angles. The work I’m actually proud of has been through five rounds, not one.”
The implication for business leaders is significant. When evaluating your team’s AI readiness, don’t measure how well they prompt; measure how well they edit. The people producing the best AI-assisted work aren’t crafting clever instructions. They know what good looks like and refuse to ship anything less.
4. From Crafting Prompts to Designing Systems
There’s a natural evolution once you stop obsessing over individual prompts: you start thinking about repeatable infrastructure instead.
Doug Compton, Principal AI Developer, described this shift in concrete terms. Early on, MCP servers were the default way to connect AI agents to external tools and services. They still matter for deployed agents, where security boundaries are critical. But Doug unlearned the assumption that they’re always the right approach.
“MCP servers are still relevant for deployed AI agents. They provide a nice security barrier,” Doug said. “But I use command-line tools wrapped in Skills locally while developing. It uses far fewer tokens in your context window, allowing AI to stay sharper for longer.”
The distinction is practical: Skills are modular components that provide AI with reference information without overwhelming the context window. This space allows agents to remain focused during longer sessions.
Savannah described the same evolution from the marketing side: “I used to spend time crafting the perfect single prompt. Now I spend that energy building reusable systems: projects with instructions, prompt templates, and Claude artifacts that solve a repeatable problem. The skill that matters isn’t writing one great prompt, but designing a workflow that functions without one.”
This mindset is the difference between using AI and operationalizing AI: one is a single activity, the other scales.
5. The Prompts Got Shorter. The Stakes Got Bigger.
Early AI use tends to stay in a safe lane: draft this note or summarize this document. The tasks are clean, scoped, and low risk. That’s a reasonable starting point; it’s also a plateau.
“The category of problem I’m willing to think through with Claude has expanded dramatically,” David said. “Things I used to only discuss with a business partner, an attorney, or a peer CEO are now things I work through with Claude first. Not as a replacement, but as a way to show up sharper.”
He summed it up in one line: “The prompts got shorter. The stakes got bigger.”
That progression matters because it reflects a genuine shift in trust. When leaders start using AI for strategic thinking rather than just task execution, they move faster in the rooms that count. The analysis is more layered, and the decisions are better informed. Critically, the AI didn’t replace the human conversation, but raised the floor for it.
6. Own the Work
As AI becomes part of the daily workflow, a strange new reflex has emerged. Someone compliments a presentation, a proposal, or a piece of writing, and the person behind it immediately points to the tool. ‘Oh, Claude did that.’ Or ‘It was all AI.’ The deflection is almost automatic now, and most people don’t realize what it costs them.
“These tools are part of how we work now, and the output still requires human judgment, input, and craft,” Sarah said. “We’re guiding it, shaping it, deciding what’s worth shipping. Giving the robots all the credit was a kind of self-erasure that sold our actual work short.”
Whitney framed it differently: “What’s actually useful is using AI as a thinking partner before I write a single line of copy: pressure-testing an angle, stress-testing a subject line, or finding the three reasons a campaign might not land. The output gives me clarity, and the result is still mine.”
Ownership matters because it shapes how teams value their own contribution. If everyone treats AI output as ‘the AI’s work,’ they stop investing in the editorial judgment and creative instincts that make the output worth anything. The tool is powerful, but the person pointing it in the right direction is the reason it produces something good.
7. Your Work AI and Your Personal AI Are the Same
There’s still a mental wall between ‘work AI’ and ‘everything else.’ Professional tools stay on the professional side. Personal logistics stay personal. What feels organized is actually limiting.
“I used to think my work tools were for work and I’d figure out the rest of my life some other way,” Sarah said. “Now the same workflows that help me prep for a client meeting also help me plan a trip, draft an email to my kid’s teacher, or build a checklist for a bridal shower. The pattern is the same, the trust is the same, the tools are the same. Splitting them off from each other was holding me back from getting good at any of it.”
This shift isn’t about work-life balance or using company tools for personal tasks. It’s about muscle memory. Every personal use of AI builds fluency that carries over directly to professional work. The person who uses AI to meal prep on Sunday shows up sharper for a strategy session on Monday because the reps compound across domains. Keeping the two separate means learning slower in both.
The Real Lesson Underneath All of These
Every unlearning on this list comes back to the same tension: the instincts that felt careful were actually just slow. Over-preparing, over-researching, over-prompting, under-trusting, and under-owning. The team didn’t get better at AI by adding new skills. They got better by removing the habits that were quietly keeping them from going faster, thinking bigger, and trusting themselves more.
The tools will keep changing, and the models will keep improving. But the competitive advantage won’t lie in who adopts the newest feature first. It’ll live in those who shed their old habits fastest.
Before school’s back in session, the real test is whether you’re willing to forget what you learned last semester.
How Learning Turned Into Leading
Written by: Savannah Cherry
Savannah is our one-woman marketing department. She posts, writes, and creates all things Slingshot. While she may not be making software for you, she does have a minor in Computer Information Systems. We’d call her the opposite of a procrastinator: she can’t rest until all her work is done. She loves playing her switch and meal-prepping.
Expert: David Galownia
David excels at propelling Slingshot towards their goals and oversees the strategic direction of the company. He’s been described as ‘intense, driven, caring, and passionate’ both at work and play. At work, he enjoys watching his team explore, imagine, and reinvent to do the best by their clients. At play, he drives Karts at insanely high speeds and scares his wife half to death. It’s all or nothing. Which means he gives it all.
Expert: Sarah Bhatia
Sarah Bhatia brings people together. In her decade plus of product and product-adjacent experience, her focus has been on cross-functional collaboration, asking lots of questions, and getting big results. She excels at strategy development, and getting the right brains in the room to solve big problems. Sarah would describe herself as a daredevil, because she’s not afraid to ask “dumb“ questions, get smart answers, and take (calculated) risks.
Expert: Doug Compton
Born and raised in Louisville, Doug’s interest in technology started at 11 when he began writing computer games. What began as a hobby turned into his career. With broad interests that range anywhere from snorkeling, science, WWII history and real estate, Doug uses his “down time“ to create new technologies for mobile and web applications.
Expert: Whitney Powell
Whitney earned her degree in Marketing and Management from the University of Kentucky and discovered her passion for marketing and events. Her go-getter attitude, willingness to learn, and problem-solving abilities elevate the Slingshot team. Known as a daredevil, Whitney loves trying new things and embracing challenges, whether traveling to new places or taking on new projects at work.
Frequently Asked Questions
Over-preparing a prompt with excessive context, constraints, and background can actually slow you down without improving the output. Today's AI models are designed to ask clarifying questions when they need more information. Giving the tool raw materials and a clear ask, then letting the conversation fill in gaps, tends to produce faster and sharper results than front-loading every detail.
Prompting matters, but editing matters more. The first output from AI is rarely the final version, and chasing the "perfect prompt" can become a trap. The people producing the best AI-assisted work are strong editors who know what good looks like, iterate quickly, and refuse to ship anything that doesn't meet their standard. When evaluating AI readiness on your team, editorial judgment is a better signal than prompting technique.
Using AI means writing one-off prompts to complete individual tasks. Operationalizing AI means building reusable systems: prompt templates, saved project instructions, and workflows that solve repeatable problems without starting from scratch every time. The shift from single prompts to scalable infrastructure is where teams start seeing compounding returns from AI adoption.
AI is a tool, not a co-author. The human behind the work is still guiding it, shaping it, and deciding what's worth shipping. Deflecting credit to the tool sells short the judgment and craft that made the output valuable in the first place. Owning the work keeps teams invested in the editorial instincts and creative thinking that AI depends on to produce anything good.
Yes. Every time you use AI outside of work, whether it's meal planning, trip logistics, or drafting a personal email, you're building fluency that transfers directly to professional use. The patterns are the same, and the reps compound across domains. Keeping "work AI" and "personal AI" separate means learning slower in both contexts.
