đ This is a free edition of the MPL Build Newsletterâresources for running your people team like a product team. Become a paid member to get access to exclusive guides, templates, and case studies.
The instinct, when youâve bought a shiny new platform, is to open the platform. Go ahead, have a play, learn the features, enjoy yourself. Play! Work is funner that way, and youâll get to know what youâre working with.
But then, resist it.
The fastest route to a good Sana course is to not touch Sana for the first fortnight. I know how that sounds coming from someone whose entire last post was about how building has never been easier. But âthe build is easyâ and âstart by buildingâ are not the same claim, and conflating them is how you end up with forty beautifully formatted modules that teach the wrong thing very efficiently. The best AI prompting, as validated by Anthropicâs AI Fluency Index, is not the first thing that comes out of a system, it requires understanding and human-in-the-loop interrogation, which requires preparation.
For the first few weeks, the unit of work youâre aiming towards isnât the course. Itâs the brief. Get the brief right and the course nearly builds itself. Get the brief wrong and no amount of slick authoring will save you.
Buy the thinking before you build the thing
The single most important decision I made on this project was spending real money on the design, not the production. I brought in Luca Oake via my good friends at Union HR, who used to build onboarding programmes at Uber, to partner with me on the architecture. Her job wasnât to make pretty slides. It was to go and find out what we were actually trying to do.
So she interviewed every managerial stakeholder and our CEO. Not with the usual âwhat does a new joiner need to knowâ survey that gets you a list of policies and a vague gesture at âculture.â She was digging for two different things at once: what does a recruiter genuinely need in their first month to be good at this job for our clients, and what would the aspirational version look likeâŠ. the onboarding that makes a new joiner think, three days in, âoh, these people mean business.â
That second question matters because onboarding is a brand moment. Itâs the most concentrated, highest-attention window you will ever have with an employee, and most onboarding squanders it on a parade of compliance modules and a Slack channel nobody posts in after week one. If you treat it as a box to tick, or a moment to glue someone to unskippable compliance videos, people clock that immediately, and youâve burned your best chance to set a standard and build a learning culture. Lucaâs interviews were how we made sure we were building toward the aspirational thing and not just the adequate one.
The shape she came back with
What Luca built was a four-week programme with a deliberate centre of gravity for everything that came after it. Iâll share a quick overview of how it worksâŠ
Week one is an intensive online âremote classroomâ, the high-density week. Every section is pinned to an explicit learning objective, so nothing is in the room just because itâs traditionally in the room. If a section couldnât name what a recruiter should be able to do by the end of it, it didnât make the cut, from candidate care to AI-fluency. We ran the gambit and totally redesigned what it took to be a modern recruiter.
Weeks two, three and four we drop the intensity and pick up the âon the jobâ application. Each week has two sessions: a kick-off deep-dive at the start, and a Friday workshop with a piece of submitted work at the end. So the rhythm is learn the thing, go and do the thing, come back and show your thing. Itâs the difference between a programme you sit through and a programme youâre actually metabolising.
This architecture is a fluid blend between on the job tasks, human interaction, and self-directed learning. Once you have a clear map of sessions, each with a learning objective and a job to do, the rest of the work stops being âdesign a training programmeâ (a task that makes most People People break out in a cold sweat) and becomes âfill in some well-defined boxes.â Which is a task you can hand to other people.
The reaction model (or: how four hours of manager time was enough)
Here is the move that made the whole thing really challenging at Talentful. Iâve already mentioned Talentful is a business that cares deeply about their clients, and want them team to be focussed on them as much as possible, which means I had a constant job of discovering clever ways to limit the lift on them, while maximising our content in People & Ops. So, we did not ask managers to build anything.
If you ask a busy manager who owns the content to âput together an onboarding session on sourcing,â one of two things happens. Either you wait six weeks and get nothing, or you get a Google Doc with more work than you know what to do with, prompting weeks of questions, back and forth follow ups, and a final product clearly resentful of its own existence. Asking people to create from a blank page is the most expensive request you can make of them, and they know it, which is why they avoid it.
So we flipped it. We turned each of Lucaâs sessions into a brief: a structured template with the learning objective already written at the top, and clearly marked space for the things only the content owner could provide: old resources theyâd used before, useful links, half-formed ideas for a lesson plan, the war stories and worked examples that actually make a session land.
Reaction is cheap. Creation is expensive. Give someone a well-formed thing to push against and theyâll give you twenty sharp minutes; give them a blank page and theyâll give you an out-of-office. By front-loading all the structural thinking into the brief, we reduced the managersâ job to the one part of it that genuinely required them, and felt the minimum we could really push for without delivering something poor quality⊠and that is how the entire programme cost roughly four hours of manager time across the whole project, rather than four hours per session per person.
Now you open Sana (or a tool like it)
With the briefs filled in, the build is the quick part, and something I could do relatively independently, with Luca as my co-reviewer and content support system. Did this do what we were asking it to? Where could we refine an activity, or suggest a deeper reading list?
At the end, after 2 weeks of work moving each brief into Sana, reviewing the content, we delivered a package where each session brief became a Sana Live Course. The platform is built for this; youâre pouring a defined structure into a container thatâs designed to hold it, rather than wrestling a blank LMS into submission. The submitted work from those Friday workshops? That runs through Sanaâs submission feature, so the âgo and do the thingâ half of the rhythm is captured in the same place as the learning, not floating around in a separate inbox youâll lose by week three.
The mindset that kept us fast here was first draft live, then iterate. We did not try to ship a perfect, gold-plated version of each course before anyone saw it. We got a real, rough version standing up in the tool, then let the content owners look at the actual thing (in situ, as a course, not as a description of a course) and react again, give us feedback, amend things. A draft you can point at and argue with beats a blank page youâre being precious about, every single time. The second round of manager input was sharper than the first precisely because they were reacting to something concrete.
Thatâs the whole sequence: interview for the real need, design the architecture, brief the sessions, let managers react, pour the briefs into a tool like Sana, ship a rough draft, iterate live. Itâs a tool that is precisely where great thinking finally comes alive.
In part three, the bit I find genuinely satisfying as an Ops person: how all of these individual courses get rolled into one enrollable programme with a pass/fail rubric, and how the whole four-week machine ends up running every month with a single facilitator and a duplicate button.