01
You describe your home
Tell RightRobot about your floors, pets, clutter, budget and what matters most. You can write naturally rather than learning a long filter system.
Methodology
How RightRobot works
RightRobot is designed to answer a simple question: which robot makes the most sense for your home? The AI helps understand your request. It does not decide which product wins.
AI interprets→Hard requirements filter→Evidence rules rank→Confidence is shown
The matching process
From a normal conversation to a defensible shortlist
02
AI interprets — it does not pick the winner
AI converts what you said into a structured household profile. The recommendation itself is then produced by RightRobot's fixed catalogue, evidence and hard-gate rules.
03
Must-haves are checked first
A hard budget, maximum robot height, minimum threshold ability, required mopping or required self-emptying can exclude a robot before any preference scoring happens.
04
The remaining robots are matched to your priorities
RightRobot weighs the areas that matter for your home — such as hard-floor cleaning, carpet, pet hair, obstacle avoidance, mopping, automation, reliability, noise and value.
05
Evidence strength stays separate from Home Fit
A robot can look like a strong fit while still having developing evidence. RightRobot shows that uncertainty instead of silently turning missing evidence into an average score.
Two numbers, two different jobs
Home Fit is not evidence confidence
Home Fit is how well a robot matches the requirements and priorities you gave us. It is household-specific — not a universal review score.
Evidence confidence describes how much trustworthy, exact-model evidence supports the assessment. It can be reduced by thin coverage, too few distinct sources or meaningful disagreement.
Home Fit86%Strong match for this home
EvidenceMediumUseful evidence, still developing
What counts as evidence
Different evidence is treated differently
Manufacturer specifications
Useful for factual capabilities such as dimensions, dock features and stated runtime. Manufacturer claims do not count as independent performance evidence.
Measured performance
A third-party source reports a measurement under a stated test method or context. Different methods are not averaged as though they were identical.
Observed performance
A behaviour is demonstrated or repeatedly observed during real use, such as cable avoidance or a dock failing to empty reliably.
Reviewer judgement
A qualitative assessment stays qualitative. RightRobot does not turn an opinion into a made-up percentage.
Long-term ownership
Months of ownership can reveal maintenance, app, dock and durability issues. Time owned alone is not treated as proof of reliability.
Owner signals
Individual owner reports can flag possible failure modes, but a single complaint or praise does not create a reliability score or independent consensus.
Source quality & conflicts
We do not make disagreement disappear
Source diversity matters
Several videos from one creator do not become several independent sources. RightRobot looks for distinct sources and exact-model relevance, not just a large link count.
Sponsored or incentivised sources are labelled
Known sponsorship, supplied-product or affiliate signals are retained as provenance and can reduce the weight placed on a source. Useful factual evidence is not automatically discarded just because a commercial relationship exists.
Negative evidence stays in
Evidence is not selected simply because it supports a recommendation. If credible sources repeatedly expose a weakness, or materially disagree, RightRobot preserves that caveat and can cap confidence.
Product imagery
Product imagery is source-tracked
RightRobot uses exact-product imagery from official brand/manufacturer sources or approved partner routes where available, and keeps the source provenance recorded internally. Approved affiliate product-feed imagery is treated as the preferred commercial source when a programme supplies it; ECOVACS UK and MOVA UK feed assets are now used on matching models. Partner-feed, press and licensed assets progressively replace interim brand-hosted imagery as permissions become available. Images from unverified third-party, retailer or marketplace sources are withheld. Image availability never changes Home Fit or recommendation ranking.
Prices & availability
A recommendation still needs to be buyable
Prices and stock can change quickly. RightRobot records when an offer was checked and is moving toward approved retailer/product feeds for live availability. Where current UK availability cannot be verified, the safer behaviour is to hold or remove the purchase recommendation rather than guess.
How RightRobot makes money
Affiliate commission does not change the ranking
Some purchase routes may be affiliate-funded. Where a retailer link is monetised, RightRobot labels it before you click; Amazon Associates routes are also identified as a paid link. If you buy through that link, RightRobot receives a commission from the retailer. That does not increase the price you pay and it does not enter RightRobot's Home Fit, evidence, confidence or ranking calculations.
A retailer paying a higher commission does not receive a better product score or recommendation position. Where more than one suitable retailer is available, price, stock and route quality can determine which purchase options are shown.
Corrections
Evidence can change
Robot software, prices and product availability change. Sources can also be corrected or retracted. RightRobot keeps provenance so that evidence can be revisited. If you spot a factual error, use the contact details on the site and include the robot model and the point you think needs checking.
Editorial policy · Affiliate disclosure · About RightRobot · Contact
RightRobot is operated by Brown Bear Productions Ltd · Company no. 11412367 · Registered in England and Wales · Registered office: Fircot, Elm Lane, Reading, England, RG6 5UQ.