Perform Your Visit as Usual: Start a store visit or survey on the FieldKo app. You’ll see your tasks and questions (e.g. a checklist or scorecard to fill out). Some questions might be standard (like “Is the promotional display set up?” with Yes/No), and others might hint that they use AI. For instance, a question might say “Shelf compliance (auto-calculated after photo)” or an instruction like “(Photo will be analysed to generate this answer)”.

Capture Required Data (e.g. Photos): When you reach a question that relies on AI, you’ll usually be prompted to provide input – typically a photo. Using the AI Shelf Recognition workflow described earlier, you take a photo of the relevant shelf or display. In some cases, the question might not explicitly ask for a photo if the app automatically triggers the camera elsewhere, but usually you’ll know because the survey will have a photo upload field or instructions.

AI Runs in the Background: Once you’ve added the required input (like the photo), you can move on. FieldKo will send the photo to the AI service in the background. There might be a note on that question like “Analysing photo, please wait…” or you might just know from the process that the AI is working (as explained, you might even get a notification when done). The key is, you don’t have to manually fill that particular answer right now – it’s pending the AI result.

Auto-Population of Answers: After a minute or two, when the AI analysis is complete, the app will automatically populate the survey answer for you. The previously blank or “waiting” field will now have a value. Using the earlier example, the Shelf Compliance Score question might now show “92%” as the answer, which was calculated by the AI from your photo. If the question was a Yes/No type (e.g. “All products present on shelf?”), the AI might fill in “No” (with a comment like “Product X missing”) or “Yes” based on what it found. Essentially, the app behaves as if a very fast and diligent assistant went ahead and answered the question for you, using the data from the AI’s analysis.

Double-Check the AI’s Work: Go through the survey and identify which answers were filled by AI (they might be pre-filled without you touching them). Read them and make sure they make sense given what you observed in the field. For example, if the AI set “Compliance = 92%” but you know a large section of the shelf was empty, you might question that. Usually the AI is accurate, but it’s good practice to verify.

Tap to Edit if Needed: If you find an AI-generated answer that you believe is incorrect or you simply want to adjust it, tap on that answer field just as you would with any normal response. The field will let you change the value or select a different option. You can override the AI suggestion by selecting your own answer. For instance, if the AI answered “Yes” to “Is the promotional display correctly executed?” but you noticed a minor issue that the AI might not catch (maybe a small sign was missing), you might choose to change the answer to “Mostly – minor issues” (if that’s an option) or “No” depending on how the survey is structured. Likewise, if an auto-count seems off, you can type the number you counted manually.

Partial Edits: You don’t have to reject everything the AI did – you can selectively edit. Maybe it filled 5 different fields and you agree with 4 of them, but one needs tweaking. Simply change that one. The others can remain as is.

Add Comments (if the survey allows): In many cases, surveys or scorecards have a comment field for context. If an AI-filled answer was adjusted, you might leave a brief comment why, e.g. “Adjusted compliance score due to end-cap not in photo.” This isn’t strictly necessary, but it can be useful for transparency if your managers review the report later. Of course, if the AI was spot on, no comment is needed – you can just accept its answers.

Ensure all required questions have answers (either filled by you or by AI).

The AI-filled responses are treated just like any other response when you submit. There’s no special step needed – simply hit Submit or Complete Visit (whichever button finalises the form in your app).

The submitted report will include the AI-populated answers. In the data that managers or the back office see, there may be an indication which answers were AI-generated, but on your end it’s just part of the form.

Did the AI fill everything it was supposed to? (If something expected to be filled is blank, you might still have time to retry analysis. This is rare, but if the AI couldn’t determine an answer, you can capture another photo or fill it manually. Usually, though, the app wouldn’t leave it blank without telling you – it would likely show an error or request another try.)

Are you comfortable with all answers? If yes, you’re good to go.

Reduced Manual Data Entry: You no longer need to count products on a shelf with pen and paper or toggle between camera and notes to record details. The AI captures that information from the photo and fills it in. This frees you up to focus on qualitative observations and actual corrective actions rather than clerical work.

Time Savings: A task that might have taken 5-10 minutes (counting items, comparing with a planogram diagram, calculating percentages) is done in seconds by the AI. Over a full day of store visits, this can save a lot of time, helping you visit more stores or spend more time on selling and relationship-building activities.

Consistency and Accuracy: The AI applies the same criteria every time. It won’t miscount because of fatigue or overlook a product tucked in a corner – as long as it’s visible in the photo, the AI will account for it. This consistency means the data collected is more uniform. For example, two different reps might have slightly different ways of judging “compliance score,” but the AI uses a standard method for all, making the results more comparable across stores and reps.

Real-Time Feedback: With automated scoring, you get immediate feedback on store conditions. If the compliance is low, you find out right there and can take action. Without AI, you might only realize later when compiling a report that a store was non-compliant. Now it’s flagged on the spot.

User Control Preserved: Importantly, you maintain control. Automation doesn’t mean you’re cut out of the loop. FieldKo’s design ensures you can always intervene. This is crucial because it recognises that AI is there to aid your expertise, not replace it. If something is off, you can fix it. The result is a partnership: the AI handles the heavy lifting of analysis, and you handle the judgment and nuanced decision-making.

Provide Good Inputs: The old saying “garbage in, garbage out” applies. The AI’s ability to fill answers correctly largely depends on the input it gets. So, when taking photos for AI analysis, make sure they’re clear and comprehensive (as discussed in the Shelf Recognition article). If an AI field is supposed to be filled from a photo, a poor photo could lead to a missing or wrong answer. Similarly, if there are other AI-driven inputs (like maybe audio or text analysis in other contexts), ensure clarity in those too.

Be Patient for a Moment: The auto-fill usually happens within a minute or two. If you notice a field is still blank but you know it should be AI-filled, give it a little time or ensure that the photo upload went through. You’ll usually get a notification or visual cue when it’s done. It’s often faster than you expect – by the time you’ve moved to the next section of the survey, it might already pop in.

Stay Aware of Edge Cases: AI is smart but can have trouble with unexpected scenarios. For example, if a shelf has an entirely new product that wasn’t in the system, the AI might not recognise it and could potentially mark a known product as “missing” when it’s actually there but just new. In such cases, use your understanding to correct the record. Also, extremely crowded or messy shelves might confuse the AI slightly – another reason your oversight matters.

Override When Necessary: Do not hesitate to override AI answers if you have reason. The company trusts your field expertise. The AI is there to assist, not to second-guess you. An overridden answer won’t get you in trouble – on the contrary, it shows you are actively ensuring accuracy.

Learn from AI’s Insights: Sometimes the AI might pick up on things you missed. For example, it might identify a missing product in the middle of a long shelf that you didn’t notice at first glance. In that sense, it can make you even more effective by highlighting details. So use the AI responses as a second set of eyes – you might discover an insight that you can act on (like “Oh, actually that variant is gone, I should reorder it!”).

You navigate to the Cereal Aisle Compliance section of the survey. It asks you to take a photo of the cereal shelf and has several questions: “Planogram Compliance %”, “Number of facings for Top 5 brands”, and “Are all expected products present?”

You tap Add Photo, snap a clear picture of the entire cereal shelf, and save it. Under the hood, FieldKo uploads the photo and begins AI analysis.

While that’s happening, you continue to the next part of your survey (maybe checking the dairy fridge). A minute later, you get a notification icon that the shelf analysis is done.

You scroll back to the cereal section of the survey. Now you see the answers filled in: Planogram Compliance % is showing 88%, the facing counts for the top 5 brands are each populated (e.g., Brand A – 8, Brand B – 6, Brand C – 0, etc.), and the expected products present? question is marked No (because the AI noted Brand C is missing).

You recall that indeed, Brand C’s space was empty – a popular item sold out. The 0 facings and the “No” make sense. However, you also notice the compliance 88% seems a bit low. You review the shelf photo and realise the AI counted that missing product as a significant compliance hit. In your judgment, the rest of the shelf was perfect, so maybe you would rate compliance a bit higher if it weren’t purely formulaic. The survey allows you to adjust the percentage, so you tap it and change it to 95% with a note “One product out of stock, but all others perfectly placed.”

You leave the facing counts as-is (no need to count again – those numbers look right to you). And the presence question remains “No,” which is correct.

You finish up any remaining questions and hit Submit on the survey.

Embrace the time savings and consistency – let the AI do the boring parts.

Always review the automated inputs – your insight is crucial.

Enjoy the faster workflow – maybe you’ll even finish your route a little early that day thanks to the boost!