Our Approach to AI Platform Comparison
SignalLedger's comparison work is editorially independent. No platform has paid for a position in our reviews, and no platform has provided financial incentives in exchange for inclusion, favourable treatment, or preferential ranking. Our assessments are produced on the basis of our own research and evaluation using a consistent set of criteria applied to every platform we cover. Commercial relationships are disclosed in full wherever they exist — including any affiliate arrangements — and these never influence the content of our assessments.
The goal of our comparison work is not to declare a winner or produce a definitive ranking table. Platform selection is genuinely individual: the right platform for a day trader who focuses on forex is likely to be a poor choice for someone who trades UK equities over longer horizons, and neither of those is necessarily appropriate for someone who wants to explore automated strategies. A comparison that produces a single "best" result necessarily loses the nuance that makes the comparison genuinely useful. Our aim is to give readers the information they need to make that judgement themselves, informed by a thorough and honest assessment of each platform's actual strengths and limitations.
This matters particularly in the AI platform space, where marketing claims are often difficult to evaluate without specific technical knowledge. Terms like "AI-powered," "machine learning-driven," and "intelligent signals" are applied to an extremely wide range of features — from sophisticated statistical models trained on large datasets to relatively simple rule-based alert systems that have been rebranded using current technology vocabulary. Our assessment framework is designed to cut through this ambiguity by asking specific, verifiable questions about each feature: what it does, how it works, what data it uses, and what its limitations are. A platform that markets AI heavily but cannot provide clear answers to these questions will receive an assessment that reflects that opacity, regardless of how compelling its marketing is.
We also recognise that this space is evolving. AI capabilities in trading platforms are genuinely developing, and an assessment written today may not fully reflect a platform's state in six months. Our reviews include a "last assessed" date, and we update assessments when we become aware of material changes to a platform's features, fees, or regulatory status. Our editorial methodology page explains the full process in detail.
What We Actually Compare
Our assessments cover ten consistent criteria. Each is addressed in its own subsection within every review, so that readers can navigate directly to the factors most relevant to their own evaluation.
(a) Regulatory Status and Client Protections
We verify FCA authorisation directly via the FCA register for UK-facing platforms, confirm the firm's specific regulatory permissions cover the products offered, and check the client protection framework that applies — including whether negative balance protection applies to retail accounts and whether the platform is covered by the FSCS. We note any discrepancy between a platform's stated regulatory standing and what can be independently verified.
(b) Fee Transparency and Total Cost of Trading
We evaluate whether all fee types are clearly disclosed on the platform's public website, including spreads, commissions, overnight financing methodology, currency conversion charges, withdrawal fees, and inactivity fees. We assess whether fees are accessible before account opening or only disclosed in lengthy terms documents. Where possible, we calculate the approximate total cost of a representative trade to illustrate the practical impact of the fee structure.
(c) AI Feature Specificity
For each platform that makes AI claims, we assess the specificity and verifiability of those claims. We ask what the AI feature actually does (technically, not descriptively), what data it operates on, whether its outputs are clearly labelled as analytical tools rather than trading recommendations, and whether its limitations are documented. Vague or unverifiable AI claims are noted explicitly in our assessments.
(d) Backtesting Claims and Methodology
Where platforms offer backtesting tools or cite historical performance of AI features, we assess the quality of the methodology disclosed. Key questions include: does the backtest account for realistic transaction costs? Is the testing period disclosed? Are walk-forward or out-of-sample results provided, or only in-sample results? Has any performance data been independently audited? Backtesting results without these disclosures carry very limited evidential weight.
(e) Risk Disclosure Quality
UK regulations require platforms to display the percentage of retail investor accounts that lose money when trading CFDs. We assess whether this disclosure is prominently placed (not buried), whether the platform's risk warnings are specific and meaningful, and whether the overall presentation of the product balances opportunity with an honest account of the risks involved.
(f) Educational Resources Depth
We assess the scope, quality, and independence of a platform's educational content. Specifically: does the education cover risk management and loss scenarios, or only strategy and opportunity? Is it structured as a genuine learning progression, or is it primarily a feature tour designed to encourage trading activity? Is it accessible to users at different knowledge levels?
(g) Onboarding Clarity and Demo Account Availability
We assess whether minimum deposit requirements are clearly stated before account opening, whether the account verification process is explained upfront, and whether a demo account is available — and if so, whether it provides a realistic representation of the live platform's pricing and functionality. Demo accounts with artificial pricing or limited functionality are noted.
(h) Support Quality and Response Times
We test customer support through the advertised channels, measuring response time and the quality of responses to technical and operational questions. We also assess whether support is genuinely technical and helpful, or whether interactions primarily focus on account management and deposit activity in a sales-oriented manner.
(i) Mobile Platform Quality
We assess the mobile application on the basis of full order management capability (not just monitoring), performance and reliability, and parity with the desktop platform's core features. A mobile app that does not allow full position management is a meaningful limitation for traders who may need to act on positions when away from a desktop.
(j) Account Requirement Transparency
We assess whether minimum deposit thresholds, account tier differences, leverage limits, and margin requirements are communicated clearly before a prospective customer begins the application process. Platforms that obscure these details make it difficult for users to make informed decisions before committing time to onboarding.
Categories of AI Trading Platforms
The term "AI trading platform" is applied to a wide range of product types. Understanding which category a platform belongs to matters substantially, because the appropriate evaluation criteria — and the relevant risks — differ between them.
Note: The category of platform matters significantly — the appropriate evaluation criteria differ between a charting tool and a fully automated execution platform. A feature that is a minor convenience in one context may be a primary risk factor in another.
(a) AI-Assisted Charting and Analysis Tools
These platforms use AI or machine learning to identify patterns in price data, generate technical analysis summaries, or highlight potential chart formations. They do not execute trades autonomously; the trader makes all final decisions. The key questions for this category are: how are the signals generated, are they labelled appropriately as analytical tools rather than recommendations, and what is the underlying model's documented accuracy? The primary risk is over-reliance on pattern recognition outputs without independent verification. Typical users include active traders looking to augment their own technical analysis rather than replace it.
(b) AI Signal and Alert Services
These services generate specific trade alerts — buy or sell signals on particular instruments at particular prices — using algorithmic or AI-driven models. They differ from charting tools in that their outputs are more explicitly directional. The evaluation questions here are sharper: on what basis are the signals generated, how is performance tracked, and how are losing signals handled? Signal services that prominently advertise accuracy rates without disclosing the full track record, including drawdown periods, should be treated with considerable scepticism. The risk is that users treat signals as reliable without understanding the model's failure modes.
(c) Copy Trading Platforms
Copy trading platforms allow users to automatically mirror the trades of selected "strategy providers" or "signal leaders." Some platforms use AI to match users to providers based on stated risk preferences. The critical evaluations here concern: the track record of strategy providers (how long, how independently audited?), the fee structure for copying (some platforms charge a performance fee on top of the underlying spread), the ease of exiting a copying relationship, and how clearly the risks of following any individual strategy are disclosed. Copy trading does not transfer responsibility for outcomes: the copier carries the financial risk of the strategy they choose to follow.
(d) Fully Automated Algorithmic Trading Platforms
These platforms execute trades autonomously based on pre-defined rules, AI-driven models, or user-configured strategies. They range from platforms that allow users to build strategies through a visual interface to those that run proprietary AI execution algorithms on the user's behalf. The evaluation considerations are the most demanding in this category: how are strategies constructed and validated, what controls exist to limit drawdown, how is the algorithm's behaviour documented, and what happens in unusual market conditions? Fully automated execution introduces the risk of rapid loss accumulation if a strategy misbehaves in unexpected market conditions, without the manual intervention that would stop it in a partly automated context.
(e) Hybrid Platforms
Many platforms offer both manual trading capabilities and varying degrees of automation or AI-assisted tools. This is increasingly the standard model. The evaluation challenge is that each feature must be assessed on its own merits — a strong manual trading environment does not validate the quality of an AI signal feature that sits alongside it, and vice versa. Hybrid platforms also introduce complexity in understanding which trades are being executed manually and which are being executed algorithmically, which is important for understanding your own exposure at any given time.
How We Assess and Present Findings
We do not assign a single score out of ten to trading platforms as a primary comparison tool. A single composite score necessarily obscures the trade-offs that a prospective user actually needs to understand. A platform that scores 8 out of 10 might be extraordinary in three criteria and weak in two others — but which three and which two matters enormously depending on your own trading context. A single number prevents you from knowing.
Instead, our reviews provide a structured per-criterion assessment. For each of the ten criteria described above, we identify what we found, note any notable strengths, flag any concerns or limitations, and describe what type of user those findings are most relevant for. We also provide an explicit summary of which types of trader the platform may be most suitable for, based on the overall assessment — and which types of trader should exercise particular caution.
Where the evidence is insufficient to reach a confident assessment on a particular criterion — for example, because a platform does not disclose enough information about its AI methodology to evaluate the claim — we say so explicitly, rather than assigning a middle-of-the-road score that implies a level of knowledge we do not have. Our editorial methodology page describes our full research and assessment process, including how we handle commercial relationships and how we update reviews when platforms change.
Review Status and Transparency
Our review programme is actively ongoing. We are building structured assessments for platforms that are relevant to UK retail traders, prioritising those with significant user bases and those where the marketing claims are most difficult for a non-specialist to evaluate independently. Reviews take time to complete properly: we do not publish assessments until we are confident they are accurate, complete, and appropriately balanced.
Platform reviews are actively being added to SignalLedger. See our Reviews section to view available assessments and understand what is in progress. We will not publish a review until we are confident in its completeness and accuracy.
The framework described in this page is the same one applied to every assessment we publish. If you are evaluating a platform that is not yet covered in our reviews, you can apply this framework directly using the step-by-step guide in our How to Compare Trading Platforms guide. That guide translates each criterion into specific questions you can answer yourself through direct research on any platform's website, terms, and FCA register entry.
We also believe in transparency about the limitations of any review programme. Our assessments represent a snapshot at a point in time. Platforms update their features, change their fee schedules, and change their regulatory status. We include the date of each assessment and flag known changes as they come to our attention. No assessment should be treated as a guarantee about a platform's current state; always verify current fees and regulatory status directly with the platform and the FCA register before making any financial commitment.
Our Comparison Framework at a Glance
| Criterion | What We Assess | Why It Matters |
|---|---|---|
| FCA/Regulatory Status | Verify authorisation via FCA register; confirm permissions match products offered | Determines FSCS, FOS access, and mandatory conduct protections |
| Fee Transparency | All cost types disclosed pre-sign-up; clarity of overnight financing methodology | Hidden or complex fees erode returns and prevent accurate cost comparison |
| AI Claim Specificity | What the feature does technically; what data it uses; documented limitations | Vague AI claims cannot be evaluated; specificity is a proxy for transparency |
| Backtesting Disclosure | Testing period, cost assumptions, out-of-sample results, independent audit status | In-sample backtests without cost adjustment overstate apparent performance |
| Risk Disclosure Quality | CFD loss rate prominently displayed; warnings specific and proportionate | Required by FCA regulation; indicates whether a platform treats risk seriously |
| Educational Resources | Scope, depth, risk coverage, learning structure, accessibility | Good education reduces harm; poor education may amplify it |
| Demo Account | Available before account opening; realistic pricing; full platform functionality | Essential for learning without financial risk; its absence is a meaningful signal |
| Support Quality | Response time, answer quality, technical competence, commercial orientation | Support is most critical at moments of difficulty; testing it in advance is valuable |
| Mobile Quality | Full order management, performance, parity with desktop features | Traders who manage positions on mobile need full functionality, not a view-only app |
| Account Transparency | Minimum deposits, tier differences, leverage limits, margin requirements — pre sign-up | Undisclosed requirements prevent informed decision-making before commitment |
What to Expect From Our Reviews
Each completed SignalLedger review of an AI trading platform includes: a structured per-criterion assessment covering all ten criteria described above; a clear identification of the platform's notable strengths and documented limitations; a "who this platform may be relevant for" section that explicitly addresses trader type, experience level, and specific use cases; a prominently placed risk disclaimer; the date the assessment was completed; and a transparency note describing any commercial relationships between SignalLedger and the platform being reviewed.
We do not use language like "our top pick" or "the best platform for X" as primary framing, because this implies a definitive judgement that is not appropriate given how individually specific platform selection ought to be. We do identify cases where a platform performs exceptionally well or poorly on specific criteria, and we note when a particular combination of strengths makes a platform well-suited to a particular type of trader — but this is always specific and qualified, not a general endorsement.
We are committed to updating reviews when material changes occur. If you notice that a platform's fees, features, regulatory status, or ownership have changed since our assessment, we welcome that information through our contact page. Our goal is accuracy and utility for our readers, and that requires an ongoing commitment to keeping assessments current. This approach is consistent with our Editorial Methodology, which we encourage all readers to review.
Educational content only. This guide is provided for informational and educational purposes and does not constitute financial advice, investment advice, or a recommendation to use any financial product. Trading and investing involve significant risk of loss. Read our full Risk Disclaimer.