Yes — in this ai weight loss app study 2026, the people who engaged early lost the most weight. Adults using an AI-enabled digital health tool as part of a dietitian-led program lost an average of 7.1% of their body weight — about 17.0 pounds — by six months, and food logging in the first 90 days was the strongest predictor of that result. Steps and physical activity, surprisingly, were not.
That finding landed at ObesityWeek 2026 in Washington, D.C., and it cuts against the way most of us think about weight-loss apps. We tend to picture the magic in the step counter, the workout streak, or the clever AI meal plan. This research points somewhere far less glamorous and far more doable: opening the app, logging what you ate, and doing it again tomorrow — especially in the first three months.
So what does that actually mean for you? If you have ever downloaded a weight-loss app, used it hard for a week, and then quietly let it slide, you are not alone. The question this study asks is different, though. It is not whether you can be perfect for six months. It is whether a strong start can carry you even when life gets in the way later. And the early answer, at least in this real-world group, looks encouraging.

What the AI Weight Loss App Study 2026 Actually Found
Let us get the basics straight first, because the details matter here. The study looked at real-world data from adults with overweight or obesity who were using Welldoc’s AI-enabled digital health tool on mobile devices. They were not using the app on its own in a vacuum. They were part of a dietitian-led weight management program at a single health system, which means a human professional was involved alongside the technology.
The results were presented on Sunday, 15 November 2026, at 2:30 p.m. ET as poster 538 at ObesityWeek 2026, the conference that ran from 14 to 17 November in Washington, D.C. The work was presented by Holly F. Lofton, MD, of NYU Grossman School of Medicine, and co-authored with Mansur Shomali, MD, Welldoc’s Chief Medical Officer. The announcement was shared publicly on 16 November 2026 via BusinessWire.
That context is important for two reasons. First, this was a poster presentation of real-world data, not a peer-reviewed journal paper — at least not yet. Second, one of the authors works for the company behind the tool. Neither fact makes the results wrong. Both facts mean we should read them as promising early evidence, not as the final word.
The researchers analyzed two groups: participants who reached three months (720 people) and participants who reached six months (355 people). That drop in numbers is worth noticing. It does not mean the other people failed. In real-world programs, people leave for all sorts of reasons — schedules change, insurance shifts, life gets busy. But it does mean the six-month results describe the people who stayed in the program that long, and we should keep that in mind when we talk about averages.
The headline numbers
At three months, average weight loss was 3.8%, or about 9.3 pounds. At six months, it was 7.1%, or about 17.0 pounds. For many people living with obesity, a loss of 5% or more is considered clinically meaningful, because it is often linked with improvements in blood pressure, blood sugar, and other health markers. A 7.1% average at six months clears that bar — and it did so in a real-world setting, not a tightly controlled trial where every meal is handed to you.
Still, averages hide a lot. Some people lost much more. Some lost less. Some may have gained. That is why the researchers dug into who lost the most, and what they did differently in those first 90 days. That is where the story gets interesting.
| Measure | At 3 months | At 6 months |
|---|---|---|
| People analyzed | 720 | 355 |
| Average weight loss | 3.8% | 7.1% |
| Average weight loss in pounds | 9.3 pounds | 17.0 pounds |
| Strongest predictor of weight loss | Early food logging in the first 90 days | |
| Not a predictor of weight loss | Step counts and physical activity | |
Read that table once more, slowly. Food logging predicted weight loss. Steps did not. If you have spent years believing that 10,000 steps a day is the main lever, that line probably stopped you in your tracks. It stopped a lot of people at the conference, too — or at least, it should have.
Early Engagement Predicted Six-Month Weight Loss
Here is the finding the researchers put front and centre: early engagement in the first 90 days significantly predicted weight loss at six months, with a p-value of less than 0.0001. In plain English, that is a very strong statistical signal. It means the link between early app use and later weight loss is very unlikely to be a coincidence in this dataset.
But what did “engagement” actually look like? The researchers compared people at the two ends of the spectrum. Participants who went on to lose more than 10% of their body weight averaged 9.0 app engagements per day during that early period. Participants who gained weight averaged 4.9 engagements per day.
Think about that gap for a second. It is not a gap between superhuman effort and laziness. Nine engagements a day sounds like a lot until you break it down: log breakfast, log lunch, log dinner, check in once or twice, maybe glance at a prompt or a message from the program. That is a handful of small taps spread across a day, not an hour glued to your phone.
And 4.9 engagements a day is not zero, either. The people who gained weight were still opening the app about five times a day. They were trying. They just were not engaging quite as often, or quite as consistently, in those first three months — and in this study, that difference lined up with very different outcomes six months later.
Does that prove that tapping the app more caused the extra weight loss? No. This was an observational study, which means it shows an association, not causation. It is entirely possible that the people who engaged more were also more motivated, had more support at home, faced fewer barriers, or had other advantages we cannot see in the numbers. Engagement might be a marker of readiness as much as a driver of results.
Still, the pattern is hard to ignore. In behavior-change research, early habits matter. The first weeks of a new routine are when the routine either takes root or quietly falls apart. If an app can help someone get through those first 90 days with regular check-ins, it may be doing its most valuable work right at the start — even if the user barely notices it at the time.
Why Food Logging Beat Step Counts in This Study
Of all the findings in this ai weight loss app study 2026, this one may change how you use your own app tomorrow morning. Food logging was the strongest predictor of weight loss. Step counts and physical activity were not.
That does not mean exercise is useless. Far from it. Movement matters for your heart, your muscles, your mood, your sleep, and your long-term health in ways that go far beyond the number on the scale. Anyone who tells you to stop walking because a study did not link steps to weight loss is missing the point entirely.
What it does suggest is something weight-loss researchers have said for years: when it comes to the scale itself, what you eat tends to move the needle more than how many steps you take. You can walk 12,000 steps and still eat more than your body needs. You cannot easily out-walk a diet that is consistently too large, too frequent, or too heavy in calorie-dense foods.
Food logging works, when it works, for a few simple human reasons. First, it makes the invisible visible. Most of us underestimate how much we eat — not because we are dishonest, but because we forget the handful of crisps, the second coffee with sugar, the sauce on the side. Writing it down, even roughly, closes that gap between what we think we ate and what we actually ate.
Second, logging creates a pause. That ten-second pause before you eat — do I want to log this? — is often enough to ask a better question: am I actually hungry, or just bored, stressed, or passing the kitchen? You do not need willpower made of steel. You need a small moment of awareness, repeated often enough to matter.
Third, logging builds a pattern you can actually learn from. After two or three weeks of honest logs, most people can see their own habits clearly for the first time: the late-night snacking, the oversized portions on weekends, the sugary drinks that add up quietly. A dietitian can see it too, and that is exactly why this program paired the app with a dietitian. The log is not just data. It is a conversation starter.
So why did steps not predict weight loss here? One likely reason is that step counts vary for reasons that have little to do with effort — a desk job, a bad knee, a long commute, a week of rain. Another is that many people compensate without realizing it: they walk more, feel hungrier, and eat a little more in return. The study does not say steps do not matter for health. It says that in this group, in this program, steps were not the thing that separated the people who lost the most weight from the people who lost the least.
If you use a weight-loss app, that is a useful reordering of priorities. Keep walking — your heart will thank you. But if you only have five minutes a day to spend in the app, spend them logging your food honestly, not chasing a step badge.

What Happened to BMI by Six Months
Weight in pounds tells one part of the story. Body mass index, or BMI, tells another — how that weight sits relative to height, and how many people moved into a lower BMI category over time.
By six months, the share of participants with a BMI over 40 fell from 31.3% to 22.5%. That is a meaningful shift, because a BMI over 40 is the range often described as severe obesity, where health risks tend to be highest and where even modest weight loss can make a real difference to daily life — climbing stairs, sleeping better, moving with less pain.
At the other end, the share with a BMI under 27 nearly doubled, from 7.6% to 14.6%. A BMI under 27 is still above the range often labelled “healthy” (18.5 to 24.9), but moving toward it from a much higher starting point is progress worth noting. It means more people were getting closer to a weight where everyday health risks start to fall.
We should be careful with BMI, though. It is a blunt tool. It does not measure muscle, bone, fat distribution, or fitness. Two people with the same BMI can have very different health profiles. The researchers reported these shifts because they help show movement across a group, not because BMI alone defines anyone’s health. Still, as a group-level signal, the direction is clear: fewer people in the highest category, more people moving down toward a lower one.
And here is the part that often gets lost in weight-loss stories: the change happened gradually, over months, in a program that combined technology with human support. There was no dramatic crash, no extreme plan, no overnight transformation. Just steady engagement, honest logging, and a dietitian in the loop.
The Surprise: Falling Engagement Did Not Erase the Gains
Now for the finding that might matter most to real life — because real life is messy, and motivation fades.
Some participants saw their engagement fall by more than half after the first 90 days. If you have ever started strong in January and drifted by March, you know this pattern intimately. The app stops being novel. Work gets busy. The novelty wears off, the notifications get swiped away, and the daily taps quietly dry up.
Here is what happened to those people anyway: they still achieved a median six-month weight loss similar to sustained users — 6.9% versus 6.8%. Read that again. The people whose app use dropped by more than half after three months lost about the same, on the median, as the people who kept going.
That is a hopeful result, and it fits with what behavior-change research has long suggested: habits formed early may support maintenance even when the original prompt fades. If the first 90 days taught you roughly how much a portion looks like, which snacks quietly add up, and how to build a plate that keeps you full, you carry some of that learning with you even after you stop logging every bite.
It is not a licence to quit after three months and expect the weight to keep falling on its own. The study does not say that. It says something subtler and more useful: a strong early stretch can build skills and routines that keep working for you, at least for a while, even when your app use drops off.
Think of it like learning to ride a bicycle. The training wheels matter most at the start. Once you know how to balance, you do not need them every minute — though it still helps to check in now and then, especially when the road gets bumpy.
For anyone who has ever felt guilty about “failing” an app because they stopped logging daily, this finding is worth sitting with. Consistency early on may matter more than perfection forever. The goal is not to be a perfect logger for life. The goal is to learn enough, early enough, that healthier choices start to feel normal — even on the days you forget to open the app.
AI Weight Loss App Study 2026: What Omada’s Research Adds
Welldoc was not the only company sharing new weight-loss app research at ObesityWeek 2026. Omada Health presented related work on 14 November 2026, and it tells a strikingly similar story — from a different tool, with a much larger group, and with a focus on how AI feedback shapes the simple act of tracking a meal.
Omada’s AI-powered Meal Map tool gives members feedback on their meals. In a randomized study of 13,852 members, Meal Map members were more likely to track at least three meals in their first month than members without the tool — 41.1% versus 36.0%. That may not sound dramatic, but across thousands of people, a five-point gap in early tracking is a real shift in behavior.
The gap did not fade after month one, either. Meal Map members had 34% greater odds of tracking at least three meals every month through 16 weeks. In other words, the tool did not just create a first-week burst of enthusiasm. It helped people keep tracking, month after month, through the period when most app habits quietly die.
And tracking, once again, lined up with weight loss. Members who tracked at least three meals in month one lost more weight by week 16 than those who did not — 3.47% versus 3.02%. That is a modest difference at 16 weeks, and we should not overstate it. But the direction matches the Welldoc findings exactly: early meal tracking is linked with better results a few months down the line.
Two different companies, two different tools, two different study designs — and the same core message. The technology matters less than the habit it supports. An AI meal map, a food log, a dietitian check-in: all of them are scaffolding for the same simple behavior, which is paying attention to what you eat, early and often enough for it to become routine.
Omada’s GLP-1 results: a different scale of weight loss
Omada also shared 12-month outcomes from its GLP-1 Care Track, and the numbers there are on a different scale entirely. Among 969 adults without diabetes, 71% stayed on their medication for a full year, and those who stayed on it lost 18.4% of their body weight. The full cohort — including people who stopped earlier — lost 16.3% at 12 months.
Those figures reflect the power of GLP-1 medicines, which have become one of the biggest health stories of 2026. Weight-loss apps and GLP-1 medicines are a huge topic this year, and they are increasingly used together: the medicine reduces appetite and food noise for many people, while the app and coaching help build the eating habits that support long-term maintenance.
It is important not to compare the two studies head-to-head as if they were a contest. The Welldoc program and the Omada GLP-1 track involved different people, different support, different medicines, and different timeframes. What they share is a lesson about staying power: whether the tool is an app alone or an app plus medication, the people who keep engaging — especially early — tend to do better.
That staying-on-medication figure, 71% for a full year, is also worth pausing on. GLP-1 medicines only work while you take them, and many people stop because of cost, side effects, supply problems, or simply because they feel they have reached their goal. Support that helps people stay on treatment, understand side effects, and build habits alongside the medicine may be just as valuable as the prescription itself.
For readers who want the fuller picture of how obesity care is changing, it helps to place these app studies alongside the wider shift in how doctors now think about obesity itself. You can read more in our earlier look at the new clinical definition of obesity and what it changes for diagnosis, which explains why weight alone is no longer the whole story.

What This Means If You Are Using a Weight-Loss App Right Now
So let us bring this back to your phone, your kitchen, and your next meal. You do not need a new app. You do not need a perfect streak. You need a strong, honest start — and a plan that survives the weeks when your motivation dips.
Here is what the research, taken together, actually supports:
- Front-load your effort. The first 90 days mattered most in the Welldoc data. If you are starting a program, treat the first three months as the investment period. Log most days. Check in most days. Ask your dietitian or coach questions while the habit is still forming.
- Log food before you chase steps. Walking is wonderful for your health, and you should keep doing it. But if your main goal right now is weight loss, your food log deserves the first five minutes of your attention each day, not the last.
- Aim for honest, not perfect. A rough log you actually keep is worth far more than a perfect log you abandon after a week. Log the takeaway. Log the second helping. Log the drink. The data only helps if it reflects real life.
- Expect your app use to fall — and plan for it. The people whose engagement halved after 90 days still did well on the median. That does not mean you should quit logging on day 91. It means a dip is normal, not a failure, and the skills you built early can keep helping you.
- Use the human support if you have it. This was a dietitian-led program, not an app alone in the dark. If your program offers a coach, a dietitian, a group, or even a check-in message, use it. Technology plus a person beat technology alone in this study’s design, and that pairing is probably part of the result.
One more thing worth saying plainly: weight loss is not a test of character. If you log diligently for a month and the scale barely moves, that does not mean you failed. Bodies respond differently. Medicines, sleep, stress, hormones, and health conditions all shape the outcome. The app can support the process, but it cannot override your biology — and anyone who promises otherwise is selling something.
That is also why the step-count finding deserves a careful reading. If you love walking, keep walking. If you track your steps for heart health, mood, or mobility, those are excellent reasons on their own. The study simply found that steps were not the predictor of weight loss in this group. Health is bigger than one number, and movement remains one of the best things you can do for your body, scale or no scale.
AI Weight Loss App Study 2026: How to Log Food Without Burning Out
Food logging has a reputation problem. For many people, it sounds tedious, obsessive, or vaguely joyless — a chore that turns every meal into maths. If that is how it feels to you, you are not wrong to hesitate. The trick is to make logging light enough that you will still be doing it in week six, not just in week one.
Start with the simplest version that still counts. Log the meal, not every ingredient. “Chicken sandwich, crisps, water” is enough to spot a pattern. You can add portions later if you want more detail, but you do not need a food scale and a spreadsheet to benefit from the habit.
Log in the moment, or as close to it as you can. The longer you wait, the more you forget — especially the small things. A quick photo of your plate, if your app supports it, can be a low-effort way to capture a meal in two seconds and fill in the details later.
Build logging into a moment you already have. After you sit down to eat, before you pick up your fork, open the app. Or log while the kettle boils. Or set one gentle reminder at your usual dinner time. The goal is to attach the new habit to an old one, so you do not have to remember it from scratch every day.
Review your week, briefly, once a week. Not to judge yourself — to learn. Where did the extra snacks creep in? Which days felt easiest? What meals kept you full until the next one? Five minutes of review on a Sunday can teach you more than seven days of perfect logging without reflection.
And if you miss a day, log the next meal anyway. One missed log does not erase a week of good data, and one missed day does not make you a person who “cannot stick to anything.” The people in this study were not perfect. They were consistent enough, early enough, for the habit to take hold. That is a much kinder standard — and a much more achievable one.
It is also worth remembering that food logging is not for everyone, all the time. If tracking every bite triggers anxiety, guilt, or disordered eating thoughts, stop and talk to a clinician or dietitian. There are other ways to build healthier eating patterns — regular meal times, balanced plates, mindful eating, cooking more at home — that do not require a log at all. The best tool is the one that helps you without harming you.

How This Fits the Bigger Picture on Walking, Steps, and Health
It is tempting to read this study as “steps do not matter,” but that would be a misreading — and an unfair one to the researchers. Step counts have been linked in other research with lower risks of dementia, stroke, and early death, including in work we have covered before on how low daily step counts relate to dementia and stroke risk. Walking matters. It just may matter more for your brain and heart than for the number on your bathroom scale.
That distinction is actually freeing. It means you can stop expecting one habit to do everything. Walk for your heart, your mood, and your mobility. Log your food for weight management. Sleep well for appetite control. Lift something heavy now and then for muscle and bone. Each habit has its own job, and none of them has to carry the whole program alone.
The same logic applies to the new generation of weight-loss tools. An AI app can remind you, nudge you, and show you patterns you would miss on your own. A dietitian can interpret those patterns and help you adjust without extremes. A GLP-1 medicine, for people who are prescribed one, can quiet the constant hunger that makes every other habit harder. None of these is a magic answer by itself. Together, for the right person, they can add up to something meaningful.
That is the quieter story of ObesityWeek 2026: not a single breakthrough, but a growing sense that obesity care works best as a combination — technology, human support, and, where appropriate, medicine — tailored to the person in front of you, not to an average on a chart.
Limitations: What This Study Does Not Prove
Honest reporting means saying what we do not know yet, so here it is, plainly. This was an observational study of real-world data, which means it shows an association between early engagement and later weight loss, not proof that engagement caused the loss. People who engaged more may have differed in motivation, support, health, or circumstances in ways the analysis could not fully capture.
The program took place at a single health system, so the results may not apply in the same way to people in different settings, with different support, or using the tool without a dietitian-led program. The six-month group was smaller than the three-month group (355 people versus 720), and the people who stayed for six months may not represent everyone who started.
The authors include a senior leader from the company behind the tool, which is a relevant conflict of interest to keep in mind — company-linked research is not automatically wrong, but independent replication matters. Finally, these findings were presented as a conference poster, not yet as a peer-reviewed journal paper, so they have not yet been through the full independent review that journal publication requires. Treat them as promising early evidence, and watch for the full published paper and for independent studies that test the same question.
For the full announcement, including the study details as shared by the company, see the BusinessWire announcement of the ObesityWeek 2026 poster presentation. For Omada’s related research, see Omada Health’s ObesityWeek 2026 research on Meal Map and GLP-1 care outcomes.
Frequently Asked Questions
How much weight did people lose in this weight-loss app study?
On average, participants lost 3.8% of their body weight (about 9.3 pounds) at three months and 7.1% (about 17.0 pounds) at six months. People who lost more than 10% of their body weight had averaged 9.0 app engagements per day in the first 90 days, compared with 4.9 per day among those who gained weight.
Is food logging really more important than step counts for weight loss?
In this study, yes — food logging was the strongest predictor of weight loss, while step counts and physical activity were not. That does not mean exercise is unimportant for health. Walking and movement still support heart health, mood, sleep, and mobility. It means that, for weight loss specifically in this group, what people ate mattered more than how many steps they took.
Do I have to use a weight-loss app every day forever to keep the weight off?
Not according to this research. Participants whose engagement fell by more than half after the first 90 days still achieved a median six-month loss of 6.9%, similar to the 6.8% median among sustained users. Early habits may help support maintenance even when app use drops. That said, some form of ongoing check-in — weighing yourself, reviewing your eating patterns, or touching base with a coach — still helps most people stay on track over the long term.
What is the difference between the Welldoc study and Omada’s Meal Map research?
They are separate studies from separate companies, presented at the same conference. Welldoc’s study looked at early engagement with an AI-enabled tool in a dietitian-led program and found food logging predicted six-month weight loss. Omada’s randomized study of 13,852 members found its AI-powered Meal Map tool helped more people track at least three meals in month one (41.1% versus 36.0%) and keep tracking through 16 weeks, and that early trackers lost slightly more by week 16 (3.47% versus 3.02%). Omada also reported 12-month GLP-1 results, with people who stayed on medication for a full year losing 18.4% of their body weight.
The Bottom Line
If you remember one thing from this research, make it this: the first 90 days are your window. Not because you need to be perfect, but because the habits you build early — especially the simple habit of logging what you eat — can keep working for you long after the initial excitement fades.
Open the app. Log the meal. Do it again tomorrow. Ask for help when you need it. Keep walking for your heart, even if the scale moves mostly because of what is on your plate. And if you miss a day, do not turn it into a reason to quit. Just log the next meal.
That is not a dramatic prescription. It is not a new diet, a new gadget, or a new promise. It is a small, repeatable habit — and according to the research presented at ObesityWeek 2026, it may be the most powerful one in the app.
This article is for general information only and is not medical advice. If you are considering a weight-loss program, an app, or a GLP-1 medicine, talk to your own clinician or a registered dietitian, especially if you have a health condition, take medication, or have a history of disordered eating.