This article is intended for general information only and does not constitute legal advice. The application of English law to any particular dispute will depend on its facts, evidence and contractual context. Parties should obtain advice from a solicitor qualified in England and Wales where appropriate.
Introduction
LinkedIn and similar professional networking platforms are now routinely used for commercial introductions, business development and the commissioning of professional services. Communications which would previously have been conducted by email or telephone may begin, and sometimes proceed substantially, through platform messaging. At the same time, automated messaging systems and large language models are increasingly used to generate, assist with or manage commercial communications.
This combination can give rise to an unusual evidential and contractual problem. A business or individual may receive messages which appear to have been sent by a representative of a company, or from an account associated with that company. The correspondence may contain affirmative statements, indications of interest and apparent agreement to a proposed course of work. The style may be highly consistent because the messages have been generated or heavily assisted by an automated system. The recipient may consequently spend substantial time preparing a detailed business plan, market assessment or competitive analysis.
If the work is subsequently delivered and an invoice issued, the company may then contend that no agreement was made and that the individual who sent the messages lacked authority to commission the work. The fact that automation was involved does not, by itself, determine the legal outcome. The principal questions remain whether a contract was formed, what its terms were, and whether the circumstances otherwise give rise to a claim for payment.
The position is fact-specific. The use of LinkedIn rather than email does not ordinarily make a communication legally ineffective, nor does the use of an automated tool necessarily prevent a communication from having legal consequences.
Formation of contracts under English law
English contract law does not generally require a contract for services to be contained in a formally executed document. Subject to particular statutory or contractual requirements, a binding agreement may arise through an exchange of oral or written communications.
The conventional analysis considers offer, acceptance, consideration, an intention to create legal relations and sufficient certainty of terms. These requirements are not necessarily applied as a mechanical checklist. The courts take an objective approach, considering what the parties’ communications and conduct would reasonably convey rather than relying solely on their undisclosed intentions.
In a commercial setting, there is ordinarily a strong presumption that the parties intend to create legal relations. That presumption is not conclusive, and the particular circumstances may demonstrate that discussions were preliminary or subject to a later formal agreement. Nevertheless, the absence of a signed contract is not, in itself, determinative.
Writing is likewise not generally essential. An agreement may be evidenced by a sequence of emails, text messages, meeting records or other communications. LinkedIn correspondence can, in principle, perform the same evidential function. What matters is the substance of the communications and whether they objectively establish the necessary elements of a contract.
The principal difficulty in an informal messaging dispute may therefore be less concerned with the medium than with the content. A message stating, for example, that a company wishes to commission specified work for an agreed fee may be materially different from a message merely expressing an interest in receiving a proposal. Similarly, an affirmative response to a detailed scope of work may have greater significance than a general expression of enthusiasm.
Consideration is ordinarily straightforward in a commercial services arrangement: one party promises or provides services and the other promises payment. More difficult questions may concern whether the scope of the services and the price were sufficiently certain. It is possible for a contract to arise even though every operational detail has not been specified. Equally, courts may conclude that negotiations had not progressed beyond an agreement to agree where essential matters remained unresolved.
Conduct after the communications may also be relevant. Delivery of the requested work, its receipt without objection, requests for amendments, or subsequent use of the work product may all form part of the evidential picture. None is necessarily conclusive, but each may assist in determining objectively what the parties understood their arrangement to be.
Attribution of automated or AI-generated communications
Automated systems and large language models do not have legal personality. They do not independently become contracting parties merely because they generate a message. The legal significance of an automated communication therefore ordinarily has to be considered by reference to the person or organisation that deployed, controlled or held out the system.
This does not mean that every statement produced by an automated system will bind the organisation behind it. Authority remains important. A company may have authorised an employee or system to conduct preliminary marketing without authorising it to enter contractual commitments. Conversely, the circumstances may support an argument that the company represented that the account or system had authority to communicate on its behalf.
The ordinary principles of agency provide a useful framework. Actual authority may arise from an express or implied authorisation. Apparent or ostensible authority may arise where a principal has represented, through words or conduct, that an agent has authority and a third party relies upon that representation in circumstances where the law recognises the reliance.
The precise application of those principles to automated systems remains fact-sensitive. The relevant question is not simply whether an AI system generated a particular sentence, but how the system was deployed and presented to the counterparty. If a company knowingly operates an automated account under its branding, permits it to conduct negotiations and gives no indication that its communications are non-binding, those circumstances may potentially be relevant to questions of authority and reliance.
Recent analysis of artificial intelligence and English private law has considered responsibility for outputs generated by automated systems through established doctrines, including contract and agency. There is not, however, a settled body of English case law specifically determining when a LinkedIn message generated by a large language model constitutes an authorised contractual acceptance. It would therefore be inappropriate to treat the technology itself as creating a new and definitive category of liability.
A company may also have contractual or internal arrangements governing the use of automated tools. Those arrangements may be relevant to internal responsibility but will not necessarily determine the company’s position as against an external counterparty who was unaware of them. The distinction between internal limitations on authority and the appearance of authority to a third party can therefore become important.
Alternative restitutionary claims
If a claimant cannot establish a concluded contract, there may in some circumstances be an alternative claim based on unjust enrichment, including a claim commonly described as being founded upon quantum meruit.
Broadly, an unjust enrichment claim requires an enrichment of the defendant, obtained at the claimant’s expense, in circumstances in which the enrichment is legally regarded as unjust. The precise doctrine is more nuanced than a simple principle that anyone who receives a benefit must pay for it.
One potentially relevant situation is the free acceptance of services. If a person knowingly accepts substantial services in circumstances in which it would be inequitable to retain the benefit without payment, restitutionary relief may potentially arise. The circumstances in which services were provided, the defendant’s knowledge, and the parties’ expectations are consequently significant.
A claimant cannot necessarily avoid the consequences of an unsuccessful contractual negotiation by describing the same facts as unjust enrichment. Courts have recognised limits on restitutionary recovery where a claimant voluntarily assumed the commercial risk of proceeding without a concluded agreement. This may be particularly relevant where the claimant knew that negotiations remained incomplete, understood that formal approval was required, or deliberately began substantial work in anticipation of securing a contract later.
The extent to which a defendant has actually benefited from the work may also matter. A completed analysis which is received, retained and used for commercial purposes presents a different factual picture from preliminary material which is promptly rejected and never used. Even in the former situation, however, the availability and valuation of restitutionary relief cannot be assumed.
Practical considerations and evidential issues
Disputes of this kind are likely to turn heavily on the contemporaneous evidence. The precise wording of the LinkedIn messages should therefore be preserved, preferably in a form which records the dates, times, identities of the accounts and the sequence of communications.
The distinction between preliminary negotiation and contractual commitment may be particularly important. Messages containing clear references to scope, deliverables, price, deadlines and payment arrangements are potentially more significant than general statements such as an intention to proceed or an indication that a proposal looks promising.
The degree of certainty concerning price is also relevant. An agreed fixed fee is comparatively straightforward. A statement that the work will be paid at an unspecified rate, or that commercial terms will be discussed later, may leave greater uncertainty. The same applies to scope. A detailed specification incorporated into the correspondence may provide considerable clarity; a broad request to “put together a strategy” may not.
Evidence of performance and delivery should also be retained. This may include drafts, source material, correspondence transmitting the finished work, invoices, records of revisions and evidence of subsequent use. If the recipient requested amendments or relied upon conclusions in the work, those facts may potentially bear upon the parties’ understanding.
The authenticity and completeness of the message history may become contested. Screenshots can be useful but may not capture the full context. Where possible, the underlying platform records and complete conversation history should be retained. Metadata, account information and records identifying the person responsible for the account may also become relevant.
In a genuine business-to-business dispute, the Late Payment of Commercial Debts (Interest) Act 1998 may be relevant to interest and certain statutory compensation or recovery rights where its conditions are satisfied. Its application should not be assumed merely because an invoice has been issued; the underlying entitlement to payment and the statutory requirements must first be considered.
For lower-value disputes, the County Court’s small claims procedure may provide a proportionate route where the relevant financial and procedural requirements are met. The availability of that procedure does not, however, remove the need to establish the underlying legal claim. Parties should also consider the costs, time and evidential risks of litigation before commencing proceedings.
Risk management observations
For businesses commissioning external work, the simplest risk-management measure is to confirm the commercial terms before substantial work begins. A short written confirmation identifying the scope, price, payment terms, principal deliverables and any requirement for formal approval can substantially reduce uncertainty. If messages are intended to be exploratory rather than contractual, that should be made clear.
The same principle applies to consultants, contractors and other potential suppliers. Where a prospective client communicates informally through LinkedIn, it may be prudent to seek confirmation of the person authorised to commission the work and of the applicable commercial terms before undertaking significant expenditure of time or resources.
Organisations deploying automated messaging tools face a separate risk. Automated systems may produce confident or affirmative language which does not accurately reflect the organisation’s internal approval procedures. Poorly supervised tools may therefore create ambiguity about whether an apparently authorised representative has made a commitment. Appropriate controls can include limits on the matters which automated systems may discuss, human review before material commitments are made, and clear indications where communications are exploratory.
Complete records should be preserved by both sides. Deleting messages, relying solely upon screenshots, or allowing an automated platform to overwrite conversation histories may make a subsequent dispute more difficult to resolve. Organisations using automated systems should also retain sufficient records to establish how a communication was generated, what instructions governed the system and whether a human reviewed the relevant exchange.
These measures are particularly important because the use of automation does not necessarily reduce the ordinary legal significance of commercial communications. It may instead increase the need for clarity about who is communicating, with what authority, and on what terms.
Informal commercial arrangements formed through LinkedIn messaging are capable, in appropriate circumstances, of giving rise to binding obligations under English law. The absence of a signed document is not necessarily decisive, and electronic communications can evidence offer, acceptance and the other requirements of contractual formation.
The involvement of an automated system or large language model does not, by itself, determine whether a contract exists. Such systems have no independent legal personality, and the relevant questions are ordinarily concerned with the organisation or individual behind the system, the authority given to it, the representations made to the counterparty and the objective meaning of the communications.
Where a contract cannot be established, restitutionary claims may sometimes provide an alternative route, although those claims have significant limits. In all cases, the precise wording of the communications, the certainty of the agreed terms, the parties’ conduct and the evidence of delivery and use are likely to be important.
The position remains fact-specific. Businesses and individuals should, wherever practicable, confirm scope, price and payment arrangements before substantial work begins and should preserve complete records of commercial communications. Parties facing an actual dispute should obtain advice from a solicitor qualified in England and Wales rather than relying solely on general principles or the apparent informality of the communications.
This article is provided for general information and educational purposes only. It is not legal advice and should not be relied upon as a substitute for advice on the facts of a particular matter. English law may produce different outcomes depending on the precise contractual, evidential and factual circumstances.
Article provided by Levy & Co Solicitors.
