Blog · Automated insulin delivery / software logistics

AID Mythbusting: What Automated Mode Does (and Does Not)

Educational · Community themes + manufacturer algorithm notes · Sep 2026

Public diabetes forums are full of automated insulin delivery (AID) frustration that sounds clinical but is often a settings map problem. Threads ask whether Automated Mode is “too safe” overnight, which knobs actually move Omnipod 5 SmartAdjust or Tandem Control-IQ behavior, why the system paused for hours on a fake low, and why highs do not get aggressive auto-corrections the way a person might dose by hand.

This post is AID orientation and mythbusting. It pulls from manufacturer algorithm FAQs and common community themes. It is not a settings recipe, not a comparison that crowns a winner, and not permission to change someone else’s targets, ratios, or basal programs. Systems, firmware, and labels change. Your clinic and your device’s current user guide own what fits you.

AID in one paragraph

Most modern AID systems pair a continuous glucose monitor (CGM) with a pump or pod. On a short cadence (often every few minutes), the algorithm reads sensor glucose and trend, predicts where glucose is headed, and then increases, decreases, or pauses automated insulin delivery toward a target or target range the system is designed around. Names differ (Automated Mode, Control-IQ, SmartAdjust, and others), but the shared idea is closed-loop adjustment on top of, not instead of, your care plan.

What that does not mean: the system replaces meal awareness, written sick-day rules, site reliability, or clinic judgment. Manufacturer materials still emphasize meal boluses, sensor accuracy limits, and clinician-guided settings.

Myth: “Just raise the basal program to fix Auto Mode”

This is one of the most common Omnipod 5 forum moves, and manufacturer language pushes back hard. Insulet’s Omnipod 5 SmartAdjust FAQ states that Target Glucose is the only adjustable setting that directly impacts automated insulin delivery in Automated Mode. Basal Programs are for Manual Mode basal delivery. Changing basal segments after system initiation, per that FAQ framing, does not steer SmartAdjust the way people hope when they are stuck in Automated Mode overnight.

Bolus calculator settings (carb ratio, correction factor, duration of insulin action, and related SmartBolus fields) still matter for suggested meal and correction boluses in Automated and Manual modes. Over time, boluses contribute to total daily insulin (TDI), and Omnipod materials describe adaptation that uses TDI across Pod changes. That is an indirect path, not “edit basal program, Auto Mode rewrites tonight.”

Safer framing: if Automated Mode feels mismatched, bring overnight CGM reports and Auto Events (or your brand’s equivalent history) to clinic. Ask which levers your exact system actually uses. Do not treat a forum basal edit as an algorithm fix.

Primary starting point: Omnipod 5 FAQs: The Omnipod 5 Algorithm / SmartAdjust. Confirm against the current user guide for your model and region.

Myth: “AID will cover meals if I forget to bolus”

Manufacturer guidance for Omnipod 5 is explicit: you still need to bolus for carbohydrate-containing meals, and you should also bolus to correct highs as advised by your clinician. Predictive micro-adjustments are not a substitute for announced meals. Other AID brands rhyme on the same idea even when the UI labels differ: carb entries and meal boluses remain user work.

Forum “it should have known I ate pizza” threads are usually expectation mismatch, not a broken Pod. Late or missed boluses, high-fat meals, and site absorption limits all sit outside what a 30- or 60-minute prediction can fully erase.

Myth: “If it suspended, I must be truly low”

AID systems often pause or reduce insulin when predicted or measured glucose looks low. That safety behavior is useful when the low is real. It is expensive when the sensor is wrong.

Compression lows (pressure on a CGM site during sleep) are a classic trap: a sharp drop on the graph, pump or pod suspends or throttles, real insulin delivery falls behind, then you wake high. Communities describe that cascade constantly. Pattern recognition and verify-before-treat logistics belong in Compression lows 101. Overnight multifactorial highs (including after suspends) belong in Overnight highs: clinic-prep guide.

This post’s point is narrower: a suspend event is a system decision based on sensor inputs, not proof that fingerstick glucose matched the cliff. Treat suspend history as clinic data, not as a moral failure of the algorithm.

Myth: “More aggressive auto-corrections should behave like my old MDI corrections”

AID algorithms are usually designed with constraints. They may increase delivery toward a target or range, deliver microboluses on a timer, or layer Sleep and Activity (Exercise) features that temporarily change aggressiveness. They are not open-ended “correct like I would with a pen” engines. Manufacturer comparisons (including Omnipod’s own algorithm FAQ) note different prediction horizons and whether a system aims at a customizable target versus a target range. Those design choices explain why some nights feel conservative even when the graph looks high.

Wanting faster corrections is a legitimate clinic conversation. Freestyle forum maxing of every aggressiveness toggle is not a substitute for that conversation, and this site will not publish numeric AID settings.

What usually does change the overnight story (logistics, not recipes)

• Sensor truth. Placement, compression, warm-up, and known interference patterns. Bad input, bad automation.
• Site and Pod or infusion reliability. Silent site failure looks like “Auto Mode will not bring me down.” Pair with Site change day survival guide and T1 go-bag: pump failure backups.
• The settings your brand actually documents as Auto levers. For Omnipod 5 Automated Mode, manufacturer materials spotlight Target Glucose as the direct automated-delivery setting, with bolus settings affecting suggested doses and, over time, TDI-based adaptation. Other brands document different primary levers. Read yours.
• Feature modes. Activity / Exercise and Sleep-style modes (names vary) temporarily change how conservative the system is. Using the wrong mode overnight is a logistics error, not proof the algorithm “hates you.”
• Meal and correction habits the algorithm can see. Consistent carb entries and clinician-guided corrections help systems that adapt from delivered insulin. Skipping boluses then blaming Auto Mode mixes cause and effect.

Clinic packet: better questions than “make it more aggressive”

Bring a short packet instead of a vibe:

• Two weeks of overnight CGM (or your download portal’s equivalent), with nights marked for compression suspicion vs true low symptoms
• AID event history: suspends, limited/manual exits, Activity or Sleep mode use, Pod or set changes
• Meal timing notes for late dinners and high-fat meals on rough nights
• Site locations and any adhesion or absorption problems (see Pump and CGM skin toolkit)
• One sentence on the goal: fewer alarms, fewer morning highs, safer exercise, or fewer decision points

Ask your clinician which settings for your system are worth reviewing, whether a feature mode change fits, and whether the graph points to sensor, site, behavior, illness, or true algorithm mismatch. That is pattern triage, not DIY dosing.

Related reading on this site

• Compression lows 101: false overnight cliffs and verify-before-treat
• Overnight highs: clinic-prep guide: dawn vs delayed meals vs suspend vs failed site
• Site change day survival guide
• Software and Hardware hubs

What this post is not

• Not dosing advice, ISF/ICR/basal recipes, or target numbers to dial in
• Not a recommendation to start, stop, or switch any AID system
• Not a claim that one brand’s algorithm is superior
• Not permission to ignore your device IFU, CGM accuracy limits, or clinic instructions
• Not a guarantee that changing any single setting will fix overnight highs