How a Solo Founder Uses Evening Planning with AI to Run Two Products Without Burning Out

The subject of this case study is a solo founder — let’s call her Mara — who has been running two B2B SaaS products simultaneously for eighteen months. One product is in growth mode with active customers and a small contractor team. The other is in early development, with Mara as the primary builder.

This is not an unusual profile among founders who have found early traction and moved to expand before revenue fully supports a team. The cognitive load is significant: customer support, product decisions, contractor coordination, sales conversations, and active development all compete for attention in the same workday.

Mara’s problem, before building the evening planning habit, was not that she worked too little. It was that she could not tell, on any given day, whether she was working on the right things. Work bled into evenings and weekends not because she was inefficient but because the day never closed — she was always mildly aware of something unfinished.

The Trigger

The decision to build an evening planning practice came after a specific incident: Mara arrived at a customer call unprepared because she had forgotten a key follow-up she had promised in the previous call. The item had lived in her head for three days, visible enough to cause anxiety but not captured anywhere it would surface at the right moment.

The fix was obvious in retrospect: capture the thing. But the deeper problem was structural. Her end-of-day had no systematic capture process. Items lived in her memory until they either surfaced organically or were lost.

She started with the Three Questions approach — what did I complete, what’s still open, what’s my first move tomorrow — because it was the smallest change she could make. For the first two weeks, it took under five minutes.

Discovering the Gap in Simple Capture

The Three Questions gave her operational closure. She stopped losing commitments. The first-move habit improved her mornings measurably — within a week, she was starting focused work within 20 minutes of sitting down instead of 45.

But she noticed a gap. The format captured what was happening but not why. She had recurring patterns that the task-level review was not surfacing: she was consistently underestimating how long product decisions took, and she was systematically overloading Mondays.

Both of these were structural problems she could have fixed if she had noticed them. She did not notice them because there was no mechanism in her nightly check-in for looking at the pattern level.

That’s what pushed her toward an AI-assisted session with a genuine Reflect phase.

The Current System

After three months of iteration, Mara’s evening planning session has a consistent shape.

Trigger: The moment she sends her last Slack message of the day or closes her final browser tab related to work — whichever comes first. She has trained herself to recognize this as the session cue rather than the end of the workday. The session ends the workday; it is not optional post-work activity.

Duration: 18-22 minutes on normal evenings. A 4-minute minimum version on constrained evenings. She has not fully skipped a workday session in six weeks at the time of this writing.

Tool: She uses Beyond Time as her primary planning interface. The key feature for her use case is the conversation history, which she reviews every Friday as part of a weekly retrospective.

The Close Phase (7 minutes)

Mara opens Beyond Time and uses a version of the brain dump prompt that she has customized for her two-product context:

“End of workday capture. I’m running two products — [Product A, growth mode] and [Product B, early dev]. Ask me one at a time: What’s unfinished in each product that needs attention? What did I promise customers, contractors, or collaborators that I haven’t done? What ideas came up today that are floating unanchored? What’s the one thing I’m most anxious about right now? Compile everything into a sorted list by product.”

The product-specific framing matters for her because without it, the two products blur together during the brain dump. Separating them in the prompt forces her to mentally context-switch through each one systematically.

The compiled list is then triaged: tomorrow, this week, defer, or drop. She estimates that 20-30% of items get dropped at triage — things she felt compelled to include during the dump but that on reflection are not actually important.

The Reflect Phase (7 minutes)

This is the phase that has had the largest impact on her strategic decision-making over time.

The prompt she uses:

“Look at what I’ve shared across this session. Three things: First, what patterns do you notice — things that came up more than once, or issues that seem related? Second, what’s the highest-leverage thing I did today across either product? Third, ask me the one question that would produce the most honest assessment of where I’m wasting effort.”

The third component — the AI generating the most relevant diagnostic question — produces different questions every session. Some have been: “You’ve mentioned the contractor coordination problem three times this week — what’s the actual constraint there?” and “The customer call prep issue came up again — at what point does this become a process problem rather than a memory problem?”

These are the questions Mara acknowledges she would not have asked herself, because they require noticing a pattern across multiple days of sessions.

The Set Phase (4 minutes)

“Given the open loops and today’s reflection, what should be my single first task tomorrow? It needs to be: startable within 30 minutes of sitting down, meaningful progress possible in 90 minutes, and the highest leverage across both products right now. One option, specific output.”

After the AI’s suggestion, Mara edits it to account for any context the AI might not have (meeting schedule, energy prediction, outstanding deadlines) and writes the final version in a sticky note on her monitor.

The sticky note is non-negotiable. She tried putting the first move in her task manager and in the app’s interface, but the physical visibility of the sticky note has a different psychological effect — it is the first thing she sees when she sits down, before she has opened any applications.

What Changed After 90 Days

The changes Mara reports fall into two categories: immediate operational improvements and slower strategic shifts.

Immediate (weeks 1-4):

The reactive morning scramble disappeared. Previously, the first 20-30 minutes of most mornings involved inbox triage, Slack review, and ad-hoc priority setting. After the first move became a reliable nightly output, she starts executing within minutes of sitting down.

Commitments stopped falling through. The Close phase’s systematic capture of promises made — to customers, contractors, anyone — created a category of task that previously had no home in her system.

Sleep improved. She is cautious about attributing this entirely to the practice, but she reports falling asleep faster and with less nighttime cycling through work problems. This is consistent with the Scullin et al. findings on task list writing and sleep onset.

Slower shifts (weeks 5-12):

The Monday overloading pattern — identified through the Reflect phase after about three weeks of sessions — was corrected by a simple rule: Monday first moves are always existing-project maintenance, never new-initiative starts. The decision-making cost of starting something new competes badly with the reorientation cost of the weekend gap.

The product decision time estimation improved. Recognizing the pattern — consistently underestimating how long product decisions take — led to a rule change: product decisions get their own dedicated blocks and are never scheduled as “30 minutes between calls.”

Both of these changes were invisible to Mara until the AI’s pattern recognition in the Reflect phase surfaced them. They are the kind of structural problems that look like motivation or discipline issues from the inside, but are actually scheduling and process problems that respond to structural fixes.

What Does Not Work

Mara is direct about the failure modes she has hit.

The Friday session consistently gets abbreviated or skipped when a deliverable is due at end of week. She has partially addressed this by building a simplified Friday version that front-loads the Set phase — the weekend-gap-specific capture — before the full Close.

The Reflect phase produces lower-quality output on the days she is most stressed. When cognitive resources are depleted, the honest self-assessment becomes harder and the session drifts toward a task list with a thin veneer of reflection on top. She does not have a full solution to this; she accepts it as a known limitation and notes that even a lower-quality Reflect phase is better than none.

The dependency concern is real. On a recent trip without her usual setup, she found that unassisted reflection felt noticeably less structured. She describes this as a reasonable trade-off — “I traded some independence for a significantly better daily practice” — but acknowledges it as something to watch.

The Transferable Lessons

Not every element of Mara’s system transfers directly to a different context. But several structural choices are broadly applicable:

The product-specific (or domain-specific) framing in the Close prompt matters whenever your work spans multiple distinct contexts. Without it, the brain dump produces a flattened list that loses structural information you need for triage.

The “one question the AI generates” approach in the Reflect phase consistently produces the most useful insights. It breaks the pattern of asking the same reflective questions in the same order every session.

The sticky note for the first move is simple but consistently mentioned by people who have tried multiple approaches. Physical visibility beats digital notification.

The 4-minute minimum version — one capture, one first move — is what keeps the streak alive. The sessions done on the hardest evenings are not wasted sessions. They are what make the habit durable.

For the framework behind the Shutdown Sequence Mara uses, see The Evening Planning Framework. For a product walkthrough, see Beyond Time Evening Planning Walkthrough.

Your action: Take one structural change from Mara’s system — the domain-specific framing in the brain dump, the AI-generated question in the Reflect phase, or the physical sticky note for the first move — and apply it to your next session.