Weight loss with meal planning: my story (−12 kg)

Losing weight with meal planning is less about willpower and more about not having to make decisions when you're tired and hungry. Weight loss requires a caloric deficit — but the hard part is finding the energy to plan meals day after day. Here's how I solved it: I let an AI build the week's dinners around the food I was already buying, and went from 110 to 98 kilos in three months. This is my story, not a diet plan — but the method is copyable.

The problem: I only cook with a recipe#

I've always enjoyed cooking — but only with a recipe. If I follow a recipe it turns out well, to the letter. If I don't have a recipe, I don't cook at all. And somewhere in there the problem grew — I discovered it in late 2025: my eating habits were really bad.

I ate what I cooked for the kids — sausages, falukorv, pasta, rice. Portions too big, too often, without a plan. The gym came in waves, a little down and a little back up. I'm not a sweets or crisps person and have no craving for desserts, but I'm weak for soda. At the end of December 2025 I weighed around 110 kilos.

My wife had been nagging me for years: see a dietitian, a nutritionist, do something. I did none of it. Instead, I built.

The turning point: build instead of booking#

Short answer: In December 2025 I built my own AI nutrition coach, nutritionist and personal trainer — instead of booking an appointment. I'm technically minded and had been building with AI since June 2025, so the step was short.

I'd been experimenting with AI since June 2025, and my first project is still alive today. So when the food question came up again, it felt natural to solve it my way: build a tool that could plan the food for me, instead of going to someone who did it manually. By the end of December 2025 it was ready.

The idea was simple: a system that knew what I should eat to lose weight, built the week's dinners around that, and made a shopping list automatically. No Sunday evening with pen and paper. No logging.

The system: how the weekly planning actually worked#

Short answer: The AI built weekly menus around the stores I already shopped at, with ready-made shopping lists and dinners only. It showed me what to avoid, and I never had to log anything — I just followed the plan.

On January 1, 2026 I started. And the first thing that struck me was how little I had to do.

No logging, no Sunday planning#

The AI automated everything. No sitting down on Sunday evening to plan the week, no food log. What would I log? I was just following the plan. The weekly menus were built on the stores I already shopped at in everyday life, with ready-made shopping lists. Dinners only — the rest of the day I handled myself.

It also showed me what to avoid in order to lose weight, and some of it surprised me. Things I thought were harmless were in the way; other things I thought were "unhealthy" mattered less than I'd assumed.

Full the whole way, never starving#

On weekdays I skipped lunch. That was my own thing — and the AI knew about it and planned the rest of the day's nutrition accordingly. I was properly full at the other meals, with good protein. I never starved.

This is important to be clear about: skipping lunch is a personal detail, not a recommendation. Meal planning works just as well with three or more meals a day. What decides the outcome is the whole week, not whether you skip a meal. If you're sitting below the recommended minimum levels, you're doing neither your body nor your weight any favors.

And the craziest thing of all is the food — it's so good. I've never eaten as varied food as I have since January 1. Many people think losing weight sounds boring and picture boiled broccoli and chicken, or nothing but salad. I haven't eaten much salad during my weight loss. For me, walks went a long way; I didn't train at full capacity.

The result, in numbers#

From 110 to 98 kilos between January 1 and March 31 — 12 kilos in three months, roughly 0.9 kg a week. Since then I've stayed steady around 97, with the same weekly planning and recipes I actually look forward to. No yo-yo, no starving, nothing that felt like a project with an end date.

A ready-made diet program vs your own AI meal planning#

Short answer: A ready-made, expensive diet program delivers a fixed menu you have to adapt to. Your own AI meal planning flips it: the menu is built around your everyday life, your store and your taste — usually cheaper and easier to keep up long term.

Those expensive, ready-made diet programs? For me they would have been wasted money. Here's how I see the difference:

The point isn't that all programs are bad. The point is that for most people, a plan based on the food you already eat is enough — and it doesn't cost a thousand kronor a month.

How it became an app: from my kitchen to others'#

After a while my siblings and my mom started asking if I could make the same kind of menus for them, based on what they shopped. I did, at first by hand. Then more people asked if they could get access too. That's when the idea for the app was born — what is today Smaklig.

The goal is simple: it should be affordable and accessible to everyone. We're building a community, and we're doing this together. You genuinely save money by planning — not just through the campaigns, but by avoiding the impulse buys and the food waste that otherwise creeps in. Swedish households throw away about 16 kg of edible food per person per year (Naturvårdsverket) at a cost of around 1,330 SEK per person (Swedish Food Agency). Planning is one of the most effective ways to cut it.

Common mistakes when you want to lose weight#

Short answer: The most common traps are starving yourself, thinking salad alone is enough, and forgetting protein. All three make weight loss harder than it needs to be.

• Starving yourself. Big deficits feel efficient but leave you hungry, tired and worse at sticking to the plan. A moderate, planned menu that fills you up beats a starvation menu every week.
• Thinking it has to be boring. Boiled chicken and broccoli forever is a recipe for giving up. Varied, tasty food isn't in the way of weight loss — it's what makes it possible to sustain.
• Forgetting protein. Protein fills you up and preserves muscle in a deficit. Without enough of it you get hungrier and lose more muscle than you want.
• Chasing only campaigns. Campaigns are good, but the big saving is in the planning itself — knowing what you'll eat means you buy the right things and throw away less.

Getting started: a simple checklist#

Short answer: Start with the store you already shop at, plan just the dinners, build the menu around the week's campaigns, and let the structure handle the calorie balance. Adjust after two weeks.

1. Pick your store. Start from the one you already shop at — that makes the shopping list real, not theoretical.
2. Plan dinners first. Dinner is the meal that most often goes off the rails. Get it in place and the rest falls easier.
3. Build around the campaigns. Let the week's deals drive the main ingredients. That cuts the cost without costing nutrition.
4. Prioritize protein and fullness. A menu that fills you up is a menu you'll keep. Put protein in every dinner.
5. Track the trend, not the daily weight. Weigh yourself a few mornings a week and look at the average over time. Adjust after two weeks — not after one bad day.

That's how simple it was for me. No magic solution, no expensive product — just a plan built around my everyday life, and the patience to let it do the work.

Related reading#

• Weekly menu for weight loss — how to build the menu around a caloric deficit
• Calorie deficit: how big does it need to be? — the numbers behind the weight loss
• Stabilisation after weight loss — how to keep the weight off without continued dieting
• Meal box vs your own planning — what's actually cheapest
• Glossary: meal planning and nutrition — definitions of TDEE, macros and GI

About the author#

Alexander Eriksson went from 110 to 98 kg between January and March 2026 with AI-planned weekly menus and has held steady around 97 kg since. He is the founder of Smaklig. He built the app after seeing how Swedish households throw away about 16 kg of food per person per year and spend thousands of kronor on unplanned grocery shopping. Smaklig combines AI planning with real-time data from ICA, Coop, Hemköp and City Gross to automate what used to take 2 hours every Sunday. Alexander has worked with data-driven optimization since 1998 and started Smaklig in 2026 to make smart meal planning accessible to every Swedish household.

My journey is one example — not medical advice.

Get started free on Smaklig →