How our comparisons work

One category, one rubric

Every category has 20 specific checks with weights totalling 100. The checks cover core workflows, workflow depth and AI. AI contributes 15 points in the standard model. Categories use different checks, so an accounting score and a CRM score are not interchangeable.

Workflow depth includes practical details such as permissions, approvals, exceptions, integration behaviour and human oversight. These are assessed as written, rather than awarded because a supplier advertises a broad feature family.

What the four evidence statuses mean

• Supported: the recorded source establishes the particular behaviour. It earns all the check's points.
• Limited: evidence establishes a real functional limitation relative to the check. It earns half the points.
• Not supported: evidence explicitly establishes the absence of the capability. It earns zero.
• Not verified: the evidence does not settle the check. It earns no documented points, but its weight remains in the possible upper score.

A feature available on a paid plan or add-on is not automatically marked “Limited”. Plan eligibility and extra cost are separate buying considerations. The headline starting price is not a promise that every scored capability is included.

Why you see a score range

A result of 62–78/100 means 62 points are supported by the evidence and 16 points remain unverified. The upper number is a possibility, not an awarded score. Rankings sort by the lower, documented number, with product names breaking ties.

Evidence coverage is the percentage of weighted checks settled as supported, limited or not supported. It measures completeness of the research, not product quality. Less public documentation may reduce the documented score without proving a weaker product.

Change the priorities

In a comparison, you can exclude a group or double its importance. We multiply the group's check weights and normalise the result to 100. Excluding all groups produces no score. The “With AI excluded” result shows the score using your other priorities with AI set to zero.

For example, 60 earned non-AI points and 9 earned AI points produce 69/100 at standard weights. Excluding AI produces 60 ÷ 85 × 100, or 71/100 after rounding. The increase reflects a changed question, not an improvement to the product.

Sources and supplier confirmations

Research uses official supplier pages, documentation, pricing and UK public-sector guidance. Sources and checked dates are attached to product records. A source is not necessarily a test result. We have not conducted hands-on performance, security or reliability testing of every product.

Where a supplier has directly confirmed a workflow, the evidence note identifies that confirmation. Pulseem's supplied confirmations cover dynamic segmentation, branching journeys, reporting/API access and supported-language AI, including English. They do not establish UK pricing, number provisioning or data residency.

Product availability, price, plans and documentation can change. Validate important workflows in a live demonstration and obtain contractual commitments where needed.

Scope and corrections

This is a researched selection, not every product on the UK market. The feature score does not include customer stars, market share, a security certification or an overall recommendation for every business. No supplier-specific score boost is used in the calculation.

Send a correction with a specific workflow, the relevant plan and a source or demonstration. We review the affected check and recalculate any changed score.