What would stop you from trying Soba (https://soba-app.com/) this month? Be specific about how you handle this today. Also say the maximum you would pay per month in USD (0 if nothing).
Page reviewed
Soba: Carb, Bread Unit and Glycemic Index Food Scanner for iPhone
https://soba-app.com/ (opens in a new tab)
10 fictional AI personas. Exploratory perspectives, not a survey of real customers. Methodology & limitations
Simulated interest
AI synthesis, not measured demand
Suggested price range
Median $5 · simulated willingness to pay
Response sentiment
What this panel suggests
The main thing that would stop people from trying Soba this month is trust: respondents do not feel comfortable relying on photo-based carb estimates when the result could affect diabetes management or insulin timing. That skepticism is sharper for mixed plates, large portions, and local dishes they believe an AI may not recognize well, especially foods like dosa or idli. Some also said they already have a working process today—manual weighing, MyFitnessPal, Apple Notes, spreadsheets, and Apple Health—so Soba has to beat an established routine rather than replace nothing. The page’s own claims about testing and median error do not fully resolve that concern for these users; they still treat the output as an estimate that needs verification. On pricing, willingness is modest: many would only try if the free option is enough, while the center of gravity is around a small monthly spend, with some willing to pay more only if accuracy and convenience clearly reduce manual work.
Pricing & positioning
Position Soba as a confidence-boosting helper for estimation, not a replacement for weighing or established carb-counting workflows. Lead with the free scan allowance and make the paid tier easy to justify through repeated use, better portion correction, and convenience. For higher-stakes users, emphasize where the app is strongest and where manual verification is still recommended.
The stated willingness to pay clusters low, with a median around $5/month. Some respondents would pay nothing unless accuracy improves enough to replace parts of their current process; a smaller segment would pay more if it consistently handled their meals well.
Recurring themes
“The biggest thing stopping me is accuracy.”Read William Towne’s response →
“The main hesitation I have about trying Soba this month is the accuracy for local Indian dishes.”Read Anjaneya Butt’s response →
“Right now, I'm already using a multi-pronged system with photos, MyFitnessPal, Apple Notes, and a detailed spreadsheet for carb counts and insulin timing.”Read Dolores Thompson’s response →
Objections to investigate
- Perceived insufficient accuracy for medical needs
- Lack of proven performance on local/specific cuisines
- Current manual system is already sufficient
- Unclear value versus free/manual alternatives
Use these findings to plan customer interviews or experiments. How to interpret AI research
What the panel said
Click a name to open that persona’s full profile.
Dolores ThompsonneutralRight now, I'm already using a multi-pronged system with photos, MyFitnessPal, Apple Notes, and a detailed spreadsheet for carb counts and insulin timing. It's very specific to my Type 1 diabetes management and insulin regimen. While Soba's photo estimation sounds interesting for quickly checking carbs, I'm hesitant to switch if it doesn't offer the granular control I need. The fact that it's free to download with in-app purchases is appealing, but I'd need to see if the 'Soba Pro' subscription offers more than my current setup. I would pay up to $5 per month for a tool that truly streamlines my data without sacrificing accuracy.
Simulated interest: 60/100
William TownenegativeThe biggest thing stopping me is accuracy. I rely on my kitchen scale and NHS carb tables, which I know are correct. This app claims to estimate from a photo, but the page itself says "The first result is an estimate" and that "the error grows with the portion." I need absolute certainty for Leo's diabetes, not a guess that could be off by 10g or more. I'd need to see proof of its accuracy with a variety of meals before I'd even consider paying. For now, I'd pay $0. I'm happy with my current system, even if it's manual.
Simulated interest: 10/100
Elisa WünscheneutralWhat would stop me from trying Soba this month? My main concern is accuracy and integration with my existing workflow. Currently, I meticulously check client meal photos against my own portion templates, which is time-consuming but ensures a certain level of precision. While Soba's ability to estimate from a photo is intriguing, I'd need to rigorously test its accuracy against my established methods. The page mentions a median error of 5.2g for meals under 15g of carbs, which is concerning for clients with very strict targets. Furthermore, it's unclear if Soba can export or sync data in a way that complements my Google Sheets and paid food-logging app, or if it would create another data silo. I would pay $10 per month for a tool that significantly reduces my manual cross-referencing while maintaining high accuracy.
Simulated interest: 60/100
Anjaneya ButtneutralThe main hesitation I have about trying Soba this month is the accuracy for local Indian dishes. My current method involves meticulously documenting examples in Google Sheets, cross-referencing with photos shared via WhatsApp, and consulting with the clinic's dietitian. While I appreciate the photo-based estimation Soba offers, I worry it might not capture the nuances of dishes like dosa or idli prepared in various ways. I'd need to see how well it performs with our specific cuisine before relying on it. If it proved reliable, I might consider paying up to $5 per month for the subscription.
Simulated interest: 60/100
Kirk WisozknegativeThis Soba app sounds interesting, estimating carbs from a photo. But honestly, I've got my system down. I jot down what I eat at restaurants in my notebook, and for packaged stuff, I check the label or use the free barcode scanner I already have. It works for me and doesn't cost a dime. I'd need to see this app do something significantly better than what I'm doing now to even consider paying. I wouldn't pay anything for it.
Simulated interest: 0/100
Aline GuillaumeneutralMy current process is quite manual; I use a notebook for initial ideas, a spreadsheet for detailed tracking, and I cross-reference with a paid app to ensure accuracy in carb counts and portion sizes. While Soba's ability to estimate from a photo sounds intriguing, especially for complex dishes, I'm concerned about its accuracy, particularly with the portion adjustments. The page mentions "free download with in-app purchases" and "Soba Pro provides additional photo and text scans," but the exact pricing structure and what constitutes a "scan" isn't fully clear to me. I would be willing to pay a maximum of $5 per month for a service like this, provided it significantly streamlines my workflow and proves highly accurate, reducing the need for my current multi-app approach.
Simulated interest: 60/100
Kelley BergstromneutralAs a registered dietitian, accuracy in carb counting is non-negotiable, especially for my prediabetic and diabetic clients. While Soba's claim of estimating carbs from meal photos is intriguing, I’d need to see how it handles differentiating between whole-grain and refined carbs—something my current platform struggles with. Currently, I spend 3 hours weekly correcting inaccuracies, so any tool must demonstrate comparable or better precision. For example, the image shows an error of 5.2g carbs from a single photo, but that may vary with complex meals like chicken tikka. My practice’s £3,200 annual expenditure on the current platform suggests a monthly budget of up to $100/month for a more reliable tool, but only if the error rate is under 5% and it integrates seamlessly with my workflow. Without LiDAR compatibility in my older iPhones, I’m also uncertain about its effectiveness across devices. I’d need to trial it to confirm it meets my clinical standards.
Simulated interest: 60/100
Adam van der MeernegativeThe primary concern preventing me from trying Soba this month is the accuracy of its photo-based estimation. Today, I meticulously log my meals by taking photos, then estimating the bread units and carbs myself, often cross-referencing with online databases before manually entering the totals into Apple Notes and my Excel spreadsheet. The accuracy figures provided – "68.5% of meals within 10 g of the actual carbs, from one photo" – seem insufficient for my precise needs, especially given the "median error from one photo" of 5.2g. I need a higher degree of confidence before I can rely on an app to replace my current, albeit time-consuming, method. I would be willing to pay up to $5 per month, provided the accuracy is significantly improved and demonstrable.
Simulated interest: 30/100
Girija NairnegativeBas! I would try the free version to see how accurate it is. Right now, every evening I write down exactly what my daughter ate, weigh portions with my manual scale, and count carbs by hand. If the Soba app can recognize Indian food accurately by photo or description, it would save me a lot of time. But I'm doing this for my child's health, so I will not pay $5.99/month as mentioned on the faq page for Soba Pro. My husband’s salary and my teaching pay don’t cover such luxuries. I am very strict with my budget. He should try another way, maybe offer it for real free. My time is valuable, but my savings are even more irreplaceable.
Simulated interest: 10/100
Sonia Madrigal UríasneutralThe biggest obstacle to trying Soba this month is the accuracy for local dishes. Today, I manually convert patient meal photos into carb exchanges for our shared clinic spreadsheet. The product mentions testing on research datasets, but I need to see how it handles 'escudella i carn d'olla' or 'pa amb tomàquet' with our specific ingredients and portion sizes. If it can accurately estimate these, I might consider paying up to $5 per month for the Pro version, depending on the daily scan limits.
Simulated interest: 40/100
Keep exploring.
Find your next question or see which topics divide AI panels.
What would you ask next?
Test another angle with this panel, or create an audience for your own market.