Adaptive and personalized scripts and no code
At 8:57am on Monday, I saw the cursor on a new rental of intake form flashing: “Title: Engineering Manager.” Today’s US goal for stand-up is the new “confidence without oversharing.” In workplace learning, real magic occurs after “submission,” so I purposely constructed its shape to shorten the role, audience, time and outcome. My job is not to let people remember the rules. It’s about making the next conversation easier than the last one. That’s especially true in the first 90 days when onboarding feels like you’re stepping into a movie halfway through.
Designing workplace learning for upskills in work flow
Day 1: 1 screen, low heart rate
I send a script rather than a sermon. For her first stand-up, our Just-in-Time Assistant (a codeless GPT helper wired into a simple scenario library) provided a one-screen plan.
Pre-brief (10S)
“Share one sentence goal, two risks, one decision that requires one decision.” Liveline
“At high levels, we are targeting Fridays for API handoffs. Two flags: rate limiting and QA availability. Post-brief (30s)
“Did we decide? Otherwise it’s clarity, detail, or time?”
She used it as she wrote it. The decision was made in six minutes. Quietly, the system recorded three small signals to track for all learners: time to decision, follow-up required, clear/reliability of self-assessment. That is the fuel for adaptive workplace learning. It’s not a personality test, it’s just something that happened in real work.
Week 2: Not a pure pattern
By the second week, her meeting was finished on time, but the follow-up was bulging. The assistant proposed a new approach: summary and assignment. “Before we wrap: the decision is A with a spike in B.
Seven days later, the follow-up was half-shrunk. If you look behind the curtain, you will see “small data.” Each script attempt will result in a small record of context, tone, and result clarity of 1-5. The assistant will compare this week to last week and suggest practicing your next microskills. This is an uplifting experience as an improvement loop, not a course catalog.
Another learner, another path
Kenji, a product lead on another floor, wrote a beautiful document but stagnated with live debate. His intake: “Stakeholder review. Looking for a healthy challenge without conflict.” The assistant routed him to a push-pull set:
invite
“I’m probably missing an angle. What would make this stronger?” narrow
“If you have to decide today, which risks are most concerned about you?”
“Commit to X now and make Y’s spikes into the timebox by Thursday.”
After two reviews, his ratio of discussion and decisions improved. The system noticed a more advanced audience on his calendar and shifted him to the opener of Executive Smarry. Same audio, better sequence. It is movement workplace learning. As the environment changes, guidance will be adapted.
What I actually built (and why it works)
People think I have a team of machine learning engineers. There was a spreadsheet, LLM, and stubborn clarity.
Scenario Library
Kickoff, Status, Stakeholder Review, Feedback, 1-1, Onboarding Synchronization Slot Based Template
A skill graph with small tone, audience seniority, and timebox variables
Framing, enquiries, summarizing, facilitating decisions, handing over a handful of signals.
Decisions are made, follow-up is required, clear/confidence, and routing rules for decision
It said, “Given this intent and these results, we recommend one script, one reflex, and a 48-hour trace experience.”
Delivery is intentionally boring, with Slack/Teams messages sent before the meeting, a concise LMS practice pack and a single mobile page where the entire script fits a single screen. Just-in-time beat just-in-case.
We will not ship without guardrails
Just before the meeting, she taps her assistant on the phone and chooses a tone that feels directly, neutral or warm, and the assistant remembers. Behind the scenes, I keep the agreement simple and ethical. There is no minimum data, clear opt-in, or secret profiling. What she sees is pure clarity, not cleverness. One screen, no scrolling, precise lines she can use now. And each script has a single red flag, one has a gentle “don’t do this” cue, so her brain maintains the light and her voice remains her.
As data got caught up and debriefs piled up, a clear pattern was formed. Confidence can now be measured in real time. The self-assessment quickly climbed in the first two weeks, then settled into a steady rhythm as the prompts disappeared from sight. That gentle plateau told me that habits had been formed. The assistant could retreat and the action remained.
The smallest trace volume has begun to deteriorate. 20 minutes from the time box. It causes objections before the alignment slips. Name the owner and checkpoint before the rooms are distributed. Each movement looked small in its own right, but together they reduced the time to decision, reduced the need for follow-up, and gave the meeting a cleaner arc. This is a quiet engine of growing learning in the workplace. It’s a stack of small levers that lift the entire team.
Clarity consistently created a sense of belonging. Learners reported feeling more understood and able to manifest themselves as themselves. The line I hear most is, “I didn’t change who I was, I changed the way I started.” That shift retains voice and identity, removing friction from the first 30 seconds of the conversation.
Three weeks later, the engineering manager sent me a note after a rigorous cross team review. “I used the “Three Options, One Proposal, An Objective” script. They decided in 15 minutes. It felt like I did. She didn’t need a new personality. She needed a reliable opener and a path to a decision that everyone could live with.
This is the heartbeat of workplace learning for onboarding and upskills for new employees. Don’t change people. Change the initial movement. One vivid start sets the tone of the room, decisions, and the habits that follow.
How to: If you’re building this for your team
Start smaller than you think.
Lists the top 10 recurring scenarios. Write one script for each (two-tone variant + one short timebox version). Capture three signals: decision (y/n), follow-up (y/n), and transparency (1–5). Route the next nudge from these signals: Deliver it in Slack/Teams.
Run it for 4 weeks in one team. It shows that fewer meandering meetings, more decisions in the room, and fewer calendar worries. It’s an upskill you can feel by Friday.
By the 60th to 90th, I usually see the assistant quietly for most learners. It’s not that they’ve “finished the course.” Pass settled into muscle memory. They stop needing a certain prompt as confidence takes over. I sometimes open the script for board deck previews, tricky feedback, or high stakes demos, but the training wheel is off.
That’s my North Star as a mentoring designer. Design the smallest nudge possible to unlock your next conversation. In workplace learning, these fine-tuning for upskills, one crisp opener, one clean decision, one confident voice at a time.
