"AI" is the most overused word in technology marketing right now. Every software company claims AI features; most of them mean rule-based automation or basic statistics. For pool service technicians and operators, cutting through that noise matters — because the real AI applications in this industry are genuinely useful, and the fake ones are a distraction.
Here's what's actually working, what's mostly marketing, and what's worth watching over the next few years.
The best route management software uses ML-based optimization to sequence stops more efficiently than humans typically do manually. This is real, useful, and the "intelligence" behind Skimmer's route sequencing. The algorithms consider stop proximity, time windows, traffic patterns, and historical service time. It's not magic — it's operations research — but it produces measurably better routes than most manual planning.
Dolphin's Nautilus line, Polaris, and Hayward's robotic cleaners use sensor arrays and coverage algorithms to navigate pool floors and walls. The newer premium models (Dolphin Premier, Polaris 9650iQ) use adaptive mapping to learn pool geometry over successive runs. This is genuine machine learning applied to a physical hardware problem, and it works — these cleaners are significantly more thorough than fixed-pattern predecessors.
Platforms that have accumulated years of chemical reading data for individual pools are beginning to offer predictive alerts — flagging pools whose chemistry pattern suggests they'll need intervention before the next scheduled visit. This is useful for commercial operators managing dozens of pools and wanting to prioritize service calls. For residential routes, it's an emerging feature that will likely become standard in route software within 2–3 years.
AI-generated service summaries, follow-up messages, and recommendation alerts are already in use in some platforms. The technology here is mature (GPT-style language models are well-suited to templated service communication) and the customer-facing quality is high. Skimmer and similar platforms are likely to integrate LLM-generated service communication within 1–2 years.
Several companies have marketed smartphone-compatible water testing devices that use computer vision to read test strip colors and provide chemistry recommendations. The concept is sound; the execution has been inconsistent. Test strip reading via phone camera is sensitive to lighting conditions, strip age, and camera quality — conditions that vary dramatically in field environments. FAS-DPD wet chemistry (Taylor K-2006 style) remains more reliable for professional use. Digital photometers (LaMotte, Hach) are accurate but use light transmission, not computer vision.
Devices like the pHin, WaterGuru SENSE, and similar connected pool monitors claim to continuously monitor chemistry and provide AI-driven recommendations. For homeowners doing their own maintenance, these are useful. For professional pool service, their accuracy at high CYA levels (common in outdoor pools) and their inability to measure CYA at all make them unreliable as primary testing tools. Use them as a supplemental alert system, not as a replacement for professional testing.
Startups are working on smartphone-based diagnostic tools that use computer vision to assess equipment condition — pump seal wear, filter condition, impeller damage — from photos taken in the field. The training data requirements are significant (you need thousands of labeled images of each failure mode), but the technology is feasible. For a tech who needs to quote a repair without being certain of the part, a camera-based diagnostic tool would be a meaningful upgrade.
IoT sensors on pool equipment (pump vibration sensors, pressure transducers, temperature monitors) combined with ML models trained on failure data represent the most compelling near-term AI application for pool service. Pentair's IntelliCenter and Hayward's OmniLogic are already collecting operational data from thousands of installations. The jump from data collection to failure prediction is a known ML problem — it's a matter of when, not if.
Integrating weather forecasts (heat waves increase chemical demand) and bather load signals (HOA pools before a holiday weekend) into automated scheduling recommendations is another near-term application. This would allow route software to proactively move a service visit forward when the algorithm predicts chemistry will need attention sooner than the standard schedule.
The most valuable AI tools for pool techs in 2026 are already available: route optimization in Skimmer and Pool Brain, and accurate chemistry calculations in tools like PoolLens. The emerging tools are worth watching, but the fundamentals matter more right now.
Pool service is a physical business. The tech who shows up, tests accurately, doses correctly, and communicates proactively will outperform an AI-assisted competitor who skips the fundamentals. The tools above are amplifiers of good practice, not replacements for it.
PoolLens: offline dosing calculators and field reference for every stop. No AI required — just accurate chemistry.
Open PoolLens Free →Yes — AI features are active in route optimization, water chemistry trend analysis, automated customer communication, and robotic pool cleaner navigation. The most visible applications are in automation hardware and route management software.
Not currently, and not any time soon. Physical pool service requires hands-on chemical testing, equipment diagnosis, and manual cleaning that cannot be fully automated. Robotic cleaners handle one task; everything else still requires a person.
Skimmer uses AI/ML for route optimization suggestions, automated scheduling adjustments, and pattern detection in chemical readings to flag pools trending out of range. These are useful but incremental improvements.
AI-assisted chemistry tools can analyze historical readings to predict when a pool will need treatment and flag unusual patterns. This is primarily useful for large commercial operations with large datasets of historical readings.
Near-term AI applications include: computer vision for water clarity assessment, AI-driven scheduling based on weather and usage patterns, natural language customer communication, and predictive equipment failure detection via sensors.