APPY result in infringement and invalidity proceedings relating to "Builder" trade marks for app-building software

The Intellectual Property Enterprise Court (IPEC) in Engineer.AI Global Ltd v Appy Pie Ltd¹found in favour of the defendants, by dismissing the infringement allegations and partially invalidating the claimant's "BUILDER" trade marks registered for app-building software. This decision also provides a useful summary of the law on the key principles of trade mark disputes and also a look at targeting, trade marks relating to AI and costs capping regime in the IPEC.

Background

The parties are “no code” application development platforms. The claimant, Engineer.ai Global Ltd, provides composable software services, where users specify elements and functionality, and Engineer.ai builds the requested app for the user. The defendants (together Appy Pie), primarily offer “no code” drag-and-drop app building software, where users can build their own app.

Engineer.ai owns several UK registered trade marks (together the Marks) for variations of "Builder", primarily in classes 9 (software for building computer applications and computer programs) and 42 (software as a service).  The Marks included three word marks – BUILDER, BUILDER.AI and BUILDER PRO STUDIO – and the following figurative marks (registered in colour and black & white formats):

Engineer.ai claimed infringement by Appy Pie of the Marks (or parts of them) in two categories:

• “Category 1" infringement: use on the Appy Pie website, to describe various of Appy Pie's products (e.g. App Builder, Chatbot Builder, Website Builder), in each case using a capital "B" for 'Builder', which Engineer.ai alleged showed that the use was not descriptive and was identical or similar to the Marks, pursuant to ss. 10(2)(b) and 10(3) of the Trade Marks Act 1994 (TMA). Engineer.ai argued that its BUILDER word mark when used with one or more descriptive words, would be perceived by the relevant public as a family of marks,² and so would associate Appy Pie's use of "Builder" as forming part of the series, and creating a likelihood of confusion; and

   

• "Category 2" infringement: in a LinkedIn post titled "7 of the best no code app builders in 2022", which Engineer.ai claimed was targeted at customers in the UK, was not permitted advertising under Regulation 4 of the Business Protection from Misleading Advertising Regulations 2008/1276 (2008 Regulations), and therefore infringed the Marks pursuant to ss. 10(1) and 10(3) TMA.

Appy Pie denied infringement, on the basis that it did not use the mark "BUILDER" solus, and only used the signs complained of in a generic or descriptive way to describe the services provided (the capitalisation of "B" in “Builder” was stylistic rather than indicative of use of a trade mark). Appy Pie also denied that the LinkedIn post targeted consumers in the UK or could be characterised as a comparative advert under the 2008 Regulations. Alternatively, Appy Pie argued that it was permissible under the 2008 Regulations, in that it was not misleading and did not denigrate Engineer.ai's products, and relied on the descriptive use defence under ss. 11(2)(b) and (c) TMA.

Appy Pie also counterclaimed for invalidity of the Marks in respect of certain goods and services in classes 9 and 42, which it submitted were purely descriptive or highly allusive of the products and services provided, pursuant to ss. 3(1)(b), 3(1)(c) and 3(1)(d) TMA.

Engineer.ai denied the counterclaim on the basis that each of the Marks were inherently distinctive and/or had acquired distinctive character and so were not liable to invalidation under s. 47(1) TMA.

Decision

In dismissing the infringement claims and allowing Appy Pie's counterclaim to partially invalidate the Marks, HHJ Melissa Clarke provided a useful recap on the law in relation to the key issues including the relevant dates for assessing, and evidential requirements for proving, inherent and acquired distinctiveness and reputation of trade marks, the principles applicable to families of marks, and targeting of websites to UK consumers. The key findings are summarised below.

Distinctiveness – inherent and acquired

The judge first considered the distinctiveness of the Marks and whether they were inherently distinctive or had acquired distinctiveness through use. For these purposes, the parties agreed that the average consumer was "anyone who wishes to create an app without coding, whether businesses, other organisations or individuals". The judge found that there were "no special features of the average consumer to weigh in this case, save that the level of attention would be higher than average given that the consumer is selecting a product that will enable them to build something of value".

In relation to inherent distinctiveness of the Marks, Engineer.ai argued that:

• The blocked in "B" in each of the stylised BUILDER Marks was a memorable and striking feature and the Marks should therefore not be equated just with the words which they contained;
  
  
• It was wrong to strip down the Marks that contain phrases and treat the Marks' constituent elements in isolation, as the composite marks were not directly descriptive of their products or services provided under the Marks; and
  
  
• Current enthusiasm for generative artificial intelligence (AI) should not skew an assessment of the BUILDER.AI word mark and the Builder.ai figurative mark at the relevant date (being the date the Marks were filed).

Appy Pie argued that the Marks were inherently non-distinctive for material parts of their specifications on the basis that:

• They designate the kind, intended purpose and broad characteristics of the goods and services, and are therefore descriptive and non-distinctive under s. 3(1)(c) TMA;

• Terms such as "Builder", "Cloud", "Pro" and "Studio" are generic, laudatory terms commonly used in the industry for these products and services, or indicate the quality of the goods under s. 3(1)(d) TMA; and

   

• The “ai” element would be disregarded by the average consumer in the same way that the eye would skip over any top-level domain associated with a sign (such as “.co.uk” or “.org”) or if it was noticed, it would be understood to be a reference to a 'builder' product that made use of AI.

HHJ Melissa Clarke conducted an analysis of each of the Marks, and found the terms “Builder” and “builder” to be "used widely and interchangeably in the software industry as a tool to enable the creation or development of software, whether used with a further descriptor or not, since at least 2013", and that the weak level of stylisation and blocked-in “B” did not save the Marks. Similar findings were also made in relation to "Cloud", "Pro" and "Studio".

The judge also considered that in the context of the software industry at the time of the registration of each of the Marks, the average consumer would to be aware "first, that 'ai' was short for artificial intelligence, and second, that it was used as an adjective for anything autonomous". Therefore, it was held that Engineer.ai's use of ".ai" would be understood as being descriptive of a builder tool which utilises artificial intelligence to introduce an autonomous element to the build. This was supported by documents from Engineer.ai's website, which referred to: "Builder.ai is a Human-Assisted AI platform for building tailor-made software" and "AI has become a wider used adjective for anything autonomous".

Turning to acquired distinctiveness, the judge looked at the Engineer.ai's evidence of use and whether the Marks were being presented to the market as a linked “family” of marks. On the basis that the core element of the Marks was "Builder", which HHJ Melissa Clarke had found to be descriptive and non-distinctive, the family of marks argument failed.

In relation to Engineer.ai's evidence of use, the judge found that while Engineer.ai's witness was "entirely credible", much of the evidence adduced was "problematic", as it related to use after the relevant date for acquired distinctiveness, to use outside the UK, did not apply to some of the Marks relied on or was not supported by documentary evidence.

In the circumstances, the judge found that none of the Marks had acquired distinctiveness and were therefore held partially invalid for various goods and services including: software for building computer applications, downloadable computer software, software as a service and software development tools, amongst others.

Given the judge's findings that the Marks were devoid of inherent or acquired distinctive character, and that a finding of reputation requires a certain degree of knowledge of the earlier mark³, she was not satisfied on the evidence that any of the Marks had a reputation in the UK, or an enhanced character capable of impacting the assessment of infringement.

Category 1 Infringement

Given HHJ Melissa Clarke's finding that the Builder Word Mark and Builder Home Mark lacked inherent or acquired distinctiveness and were partially invalid, the Category 1 alleged infringement claim failed and was dismissed.

Category 2 Infringement

Similarly, based on the judge's finding that the Builder.ai Figurative Mark lacked inherent or acquired distinctiveness and was partially invalid, the Category 2 alleged infringement claim also failed and was dismissed.

HHJ Melissa Clarke briefly considered the principles of targeting, with reference to the recent Supreme Court decision in Lifestyle Equities v Amazon⁴. On the facts, the LinkedIn post was not considered to target UK consumers or be a clear expression of an intention to solicit custom in the UK⁵. Persuasive factors in reaching this finding included: six of the seven prices featured being in US dollars (with just one in GBP), and geographic statistics adduced by Appy Pie, showed that the viewers of the LinkedIn post were based largely outside of the UK (and primarily in India). The fact the post was in English was not considered enough to displace these other factors, especially given that English is a main language of business in India (where the second defendant is based) and is also widely spoken around the world.

Given this finding on targeting, it was not necessary to consider whether the LinkedIn post amounted to a comparative advert pursuant to the 2008 Regulations.

Costs

While not recorded in the judgment, at the Form of Order hearing, HHJ Melissa Clarke provided useful guidance on the recovery of costs in the IPEC.

The judge confirmed that the IPEC costs capping regime will only be lifted in "truly exceptional circumstances"⁶, involving cases of fraud, falsification of evidence or an abuse of the court's process. The test for displacing the costs cap is higher than the test for assessing indemnity costs. The judge considered the costs capping regime to be an important part of the IPEC that needed to be protected, as it provides cost certainty to all parties involved in IPEC claims, even if it may result in significant irrecoverable costs for the winning party.

The judge also held that actual incurred costs exceeding the relevant scale costs for each stage of the proceedings did not automatically entitle the recovering party to the maximum stage cap – a summary assessment will be conducted for each stage of the proceedings and the reasonableness and proportionality of the costs claimed will be considered.

Comment

While this case does not create new law, it provides some interesting practical points for practitioners and brands to bear in mind:

• Evidence is key: as has been made clear by various previous decisions, evidence is key in demonstrating acquired distinctiveness. In this case, HHJ Melissa Clarke found that the evidence adduced by Engineer.ai in support of its position on inherent and acquired distinctiveness was lacking in various respects. Further, its own website and witness undermined its position that the words contained within the Marks were distinctive and non-descriptive of its goods and services. Conversely, Appy Pie adduced evidence of widespread use of terms including "Builder", "Cloud", "Pro" and "Studio" in the industry at the relevant dates. Therefore, practitioners should not underestimate the value of compelling supportive and adverse evidence in trade mark cases – and should ensure that any evidence adduced covers the relevant period, the territory in question and the marks relied on, and is supported by documentary evidence.

   

• "AI" is descriptive: this decision provides interesting – albeit not surprising – confirmation that "AI" in registered trade marks itself will not be considered distinctive or save otherwise descriptive marks. This was particularly the case here, as it was considered that the ".ai" suffix would indicate that the products and services in question utilised AI (supported by Engineer.ai's website copy) or would otherwise be considered to be a top-level domain (in this case, for Anguilla). Given the continued interest in and adoption of AI, trade mark owners should be mindful that references to "AI" in registered trade marks is likely to be considered banal and non-distinctive (similar to how software is regarded), and should instead focus on other, distinctive elements of brands.

   

• Costs recovery in IPEC: the judge confirmed that the IPEC costs cap will only be displaced in totality in exceptional circumstances, and not where parties advance weak claims or where there are deficiencies in the claim advanced (here, in the case of the evidence adduced). However, parties should bear in mind that it may be possible to recover the costs of interim applications in addition to the totality of the overall cap where a party has behaved unreasonably in respect of that application. Defendants who are confident in their position should also consider whether it is appropriate to apply to transfer the claim to the High Court or whether preliminary relief (such as summary judgment or strike out) is appropriate, to seek to reduce the level of irrecoverable costs if successful in defending a claim.

This article was written for The Reporter

RPC acted for the Defendants in this case.