Private-Sector Delegates Who Speak as Stakeholders, Not Principals

Summary

• WSIS established an important role for the private sector in the technical and economic development of the Internet. It did not make a company delegate the authorized voice of all companies, operators or users.
• Representation has a traceable principal-agent structure. An employee may state an employer's approved position. A trade association may state a position adopted under its membership rules. A forum appointee serving in a personal capacity may contribute expertise and stakeholder linkages. These mandates are distinct and should not be blended.
• The “private sector” contains conflicts that matter to policy: access networks and platforms, large incumbents and small entrants, vendors and buyers, registries and registrants, advertising businesses and subscribers, employers and workers. A common label cannot establish consensus among them.
• ICC BASIS, GSMA, WITSA and ICANN constituencies show the value of organized business participation and the boundaries created by membership eligibility, sector, geography and internal procedure. A broad network can be influential without becoming a universal chamber of commerce.
• Forums should publish a representation card beside each intervention: employer or association, members covered, authorization method, issue scope, dissent, financial interests and whether the speaker acts personally, institutionally or in a formal representative role.

Presence, expertise and representation are different assets

Private-sector participation is sometimes defended with one sweeping assertion: business runs much of the Internet, therefore business must be at the table. The assertion is directionally right and institutionally incomplete.

Companies own or operate access networks, data centres, undersea systems, exchanges, cloud platforms, content services, payment systems, devices and security tools. They know implementation costs that governments and advocates may underestimate. They possess telemetry and operational experience unavailable elsewhere. Excluding them would produce poorer policy and invite rules detached from technical and commercial reality.

That establishes the case for presence and expertise. It does not establish representation. A mobile operator knows its network and market. It does not automatically know or carry the interests of fixed networks, small Internet service providers, community networks, equipment purchasers, cloud customers or people using a prepaid handset. A global platform can explain its systems. It cannot turn billions of account holders into constituents merely because they use the service.

The distinction resembles a principal-agent problem. A principal authorizes an agent to act within a defined scope. A company's board and management can authorize an employee to state corporate policy. Association members can authorize officers through bylaws, votes, committees or agreed consultation. Without such a principal, the speaker may still offer expertise, prediction or advocacy. The honest verb is “argues,” “reports,” or “recommends,” not “represents.”

Forums frequently blur the verbs. A speaker is listed under “private sector,” a moderator asks for “the business view,” and a later summary records that “business supported” an option. Each step expands the original mandate. By the end, a statement by one employee may appear to be a sectoral position.

A representation audit reverses that expansion. It asks who authorized the voice, which organizations are covered, which issue was approved and how dissent was handled. The audit does not reduce business influence. It tells readers what kind of influence they are observing.

WSIS assigned a role to a category, not authority to every delegate

The 2003 Geneva Declaration of Principles recognized important roles and responsibilities for governments, the private sector, civil society and international organizations. The formulation was a political opening. It made clear that building the information society required cooperation beyond states.

The 2005 Tunis Agenda became more specific. Paragraph 34 defined Internet governance as the development and application by governments, the private sector and civil society, in their respective roles, of principles, norms, rules, decision-making procedures and programmes shaping the Internet. Paragraph 35 said the private sector had and should continue to have an important role in the Internet's development in technical and economic fields.

The Working Group on Internet Governance had described private-sector functions including self-regulation, best practices, policy proposals, research and development, standards, legal and policy contributions, innovation, arbitration and capacity building. Its 2005 report mapped activities. It did not designate a single business electorate or say how a delegate obtained authority over the category.

The phrase “respective roles” is consequently a floor for inclusion, not a warrant for collective nouns. A software vendor making a technical standard has a role. A carrier investing in infrastructure has a role. A bank relying on secure connectivity has a role. Their roles overlap but are not interchangeable.

The WSIS language also did not erase national law. Companies remain legal persons governed by boards, owners, employees, contracts and regulators. Trade associations remain voluntary bodies with chosen eligibility and internal rules. A summit badge changes neither structure.

This historical boundary matters because later multistakeholder processes inherited the categories. The IGF could balance panels and advisory bodies among government, private sector and civil society without creating formal private-sector constituencies capable of binding every firm. The category made participation legible to institutions. It did not create a principal.

A corporate employee's clearest mandate is the employer's interest

The simplest private-sector intervention is a company speaking for itself. Its delegate can describe corporate operations, costs, experience and an approved policy position. The company's name tells readers whose interest is being advanced. That is legitimate even when the interest is commercial.

The scope still requires precision. A policy officer may be authorized to advocate a public position but not disclose confidential operational facts. An engineer may explain network practice without authority to commit the company to regulation. A regional subsidiary may face conditions unlike the parent group. A consultant may speak from professional experience while lacking authority to state client policy.

Clear introductions solve much of the problem: “I speak for Company X's public-policy team on this proposal,” “I am offering technical experience and not an approved corporate position,” or “I have been authorized by these named companies to present the association's adopted view.” Such statements are more useful than a stakeholder label.

Corporate law and employment hierarchy supply some internal authorization, but they do not prove that the position reflects employees, customers or suppliers. Management may lawfully choose a lobbying position that workers oppose. A platform may argue for users' interests while its revenue model creates a different incentive. A network operator may invoke underserved communities while prioritizing spectrum or licensing terms valuable to shareholders.

These possibilities do not make the advocacy illegitimate. They define the principal. The company is entitled to argue that a rule will raise costs or reduce investment. It should support the causal claim with evidence and acknowledge distributional effects. It is not entitled to convert customers into represented citizens without a separate process.

The company voice is often strongest when it is narrow. Operational data from a named network, a disclosed methodology and a bounded conclusion can materially improve debate. A broad claim that “industry cannot implement” is weaker if competitors, smaller firms or adjacent sectors were never consulted.

A trade association has a mandate only as broad as its governed membership

Trade associations exist to aggregate business interests. They reduce the cost of participation, develop expertise, reconcile competing member positions and create continuity across long negotiations. Their collective voice can be more informative than a parade of identical corporate interventions.

Their authority depends on institutional facts. Who is eligible to join? Which members actually joined? How are dues weighted? Who sets policy? Does silence count as consent? Can a minority publish dissent? Does the board approve every submission, or can staff infer policy from broad principles? Are local affiliates bound by a global secretariat?

These questions determine whether a statement reflects a deliberated membership position, a dominant-member preference or professional staff judgment. All three can contribute value. They should not be described as the same thing.

Membership breadth is often cited as proof of representativeness. It is only one element. An association with hundreds of members may have a policy committee attended by ten. Large dues payers may possess formal or informal influence. Small businesses may join for events or technical services and never authorize advocacy. National associations may transmit their own internally aggregated positions, making the chain several layers deep.

A sound mandate record should therefore publish membership eligibility and a current member list; the governance body responsible for policy; the threshold used to approve the position; the number or proportion participating; material exclusions; and any recorded dissent. Confidential ballots can be protected while aggregate participation remains visible.

Issue scope matters as much as membership. A mobile association may credibly aggregate mobile-operator views on spectrum and roaming. Its authority is less obvious on labour rights, online speech or the welfare of all Internet users. A general chamber of commerce may have broad economic standing but limited technical depth. The fact that an issue affects business does not place every dimension of that issue inside a business association's representative competence.

ICC BASIS is a valuable channel with a bounded principal

The International Chamber of Commerce played a prominent role in organizing business engagement around WSIS and created Business Action to Support the Information Society after the summit. ICC says BASIS brings global business expertise into WSIS, the IGF and related processes. In its own historical account, BASIS membership comprises companies and associations that participate in Internet-governance work.

This is a meaningful mandate. ICC has national committees, chambers, companies and associations across many countries and sectors. It can convene experienced businesses, circulate drafts and sustain engagement that individual small firms could not afford. Institutions gain a stable counterpart rather than searching for business entities afresh each year.

The mandate remains bounded. ICC and BASIS can speak for positions developed through their membership and governance. They cannot empirically establish that every private company agrees, that non-member operators consent, or that users authorize the position. Phrases such as “the voice of global business” are institutional descriptions of ICC's ambition and network, not proof of a universal ballot.

The boundary becomes visible when ICC publications make specific recommendations. Its 2025 IGF business guide presents business messages on connectivity, cybersecurity, data governance and artificial intelligence. Readers can treat these as the organized positions of ICC BASIS. To assess their breadth, they need the participating membership, drafting process, approval method and dissent, not an assumption that “business” is singular.

This does not diminish ICC's institutional importance. It improves it. A trade association's comparative advantage is an accountable aggregation mechanism. Showing the mechanism protects its position from both exaggeration by allies and dismissal by critics.

The same rule should apply to every stakeholder coalition. A civil-society network does not represent humanity; a technical body does not represent every engineer; a government does not necessarily reflect every citizen. Precise authority is a condition of serious multistakeholderism, not a burden reserved for companies.

Sector associations illustrate why the membership perimeter matters

The GSMA describes itself as representing mobile operators and organizations across the mobile ecosystem. Its membership directory distinguishes operator, industry, rapporteur and community memberships. Eligibility reaches licensed mobile operators, vendors, software companies, some virtual and wholesale operators, and adjacent sectors under different categories.

This structure gives GSMA substantial standing on mobile connectivity. It also shows why “operators” is not a complete description of the principal. Different membership classes occupy different commercial positions. A licensed network operator, equipment vendor, mobile virtual network operator and financial-services company may agree on some standards while diverging on wholesale access, device rules, competition or data use.

WITSA offers another model. Its membership criteria make national or regional technology-industry associations the members of a global alliance. That creates an indirect chain: companies authorize or join national associations; those associations join WITSA; WITSA represents agreed positions internationally. The model can expand geographic reach, but each link has its own participation and authorization rules.

ICANN's Generic Names Supporting Organization makes sector boundaries more explicit. Its stakeholder groups and constituencies divide commercial interests among business users, intellectual-property interests, Internet service and connectivity providers, registries and registrars. The Business Constituency's charter defines eligible commercial users and excludes or separates some contracted parties to manage conflicts. The architecture recognizes that “commercial” is too broad for one undifferentiated principal.

These examples are not interchangeable. GSMA is organized around a mobile ecosystem, WITSA around associations, ICC around broad commerce, and ICANN constituencies around a particular policy body's functions. A delegate may belong to several at once and hold different mandates in each.

A forum should ask which hat is active. The answer determines the represented membership, subject competence and conflict. Listing every affiliation without identifying the speaking mandate can create an illusion of breadth rather than clarity.

Operator interests diverge across technology, market power and geography

The phrase “network operators” sounds coherent until a policy imposes costs. Large mobile groups can spread compliance across markets; a rural wireless provider may have two specialists. A submarine-cable consortium values stable landing rights; a retail access provider values affordable capacity. An incumbent may favour obligations that raise barriers to entry; a new entrant may prefer mandatory access. A transit provider, exchange point and content-delivery network see routing incentives from different positions.

Geography compounds the divergence. Power reliability, foreign-exchange constraints, equipment import rules, sanctions, spectrum prices and backbone competition vary sharply. A policy feasible for a European multinational may be impossible for a small operator in an island market. Conversely, a company may invoke developing-market constraints while speaking from headquarters with little consultation of local subsidiaries.

Scale also affects forum presence. Large firms employ dedicated policy teams and can attend annual meetings, intersessional calls and regional consultations. Small operators are more likely to rely on an association or not appear. The visible private-sector entities are therefore not a random sample of businesses.

This selection effect does not disqualify large-company delegates. It limits generalization. When a speaker says “operators need,” the evidentiary burden should include which operators, in what markets, under which architecture and after what consultation. A trade association should show whether small and regionally diverse members participated in the position.

Forums can reduce imbalance through travel support for small businesses, remote participation, regional evidence calls and publication of minority views. They should also avoid using one private-sector seat per panel as proof of sector inclusion. A representative panel may need conflicting business models, not simply multiple corporate logos.

The policy value of disagreement is high. Divergent company evidence can reveal where a proposed rule creates rents, shifts liability or assumes infrastructure only incumbents possess. Treating the private sector as a unified bloc removes that diagnostic information.

A company cannot derive a user mandate from a customer relationship

Corporate delegates frequently frame positions in terms of users: users need seamless services, users value privacy, users prefer personalization, users will bear compliance costs. Companies should study and advocate for customer welfare. They often hold valuable evidence about behaviour and service quality. But a commercial relationship is not political authorization.

Users choose among products under constraints. They may accept terms because alternatives are costly, because an employer selected the service or because network effects make exit unrealistic. Clicking agreement does not authorize a company to represent the user in public policy. Customer surveys measure responses to questions; they do not create a standing mandate on questions never asked.

Interests can align. A network and its subscribers may both oppose service fragmentation. A security provider and customers may both favour stronger incident response. Alignment should be demonstrated issue by issue, not presumed from the relationship.

Interests can also diverge. Advertising revenue may reward data collection that users dislike. A platform may favour rules that entrench its compliance advantage. An operator may pass regulatory fees to subscribers while protecting margins. A vendor may define safety in ways that require purchasing its product. None of these observations proves bad faith. They show why “we represent our users” requires more than market reach.

If a company wants to report user preferences, it should publish method, sample, wording, geography and limitations. If it created a user council with genuine authority, it should state selection and powers. If it merely predicts customer effects, it should label the prediction.

Users also hold interests beyond consumption. They are citizens, workers, creators, children, targets of abuse and members of communities. A corporate customer lens captures only part of that identity. Multistakeholder design exists partly because no service provider can absorb the public into its balance sheet.

Sponsorship, speaking access and representation must not be confused

Companies support forums with cash, venues, technology, staff and travel. Sponsorship can make participation possible and improve the quality of an event. It can also produce visibility: prominent branding, executive speaking slots, invitations and access to organizers.

Financial support does not itself create a representative mandate. A sponsor may speak for its company; it does not acquire additional authority over the sector because the event depends on its contribution. Conversely, sponsorship does not prove that the forum adopted the sponsor's position. The relevant facts are the benefits attached to support, the selection of speakers and any influence over agenda or outputs.

Role labels should distinguish sponsor, organizer, speaker, advisory member, association representative and entity. Each carries a different relationship to the event. A company can occupy several roles, but one should not silently amplify another.

The IGF is particularly important because it is a discussion forum rather than a negotiating chamber. Its institutional description emphasizes exchange among stakeholder groups. A corporate executive on a main stage contributes visibility and experience, but the appearance does not ratify the intervention as a private-sector consensus. Session reports should attribute positions to named speakers or organizations unless a documented group process supports broader language.

Sponsorship disclosure should include amount or range, in-kind support, benefits, agenda rights and speaking guarantees. Speaker selection should state whether the person was invited by organizers, nominated by an association, selected through an open call or attached to sponsorship. This information lets readers assess access without assuming purchased conclusions.

The rule should be symmetric. Governments host events, foundations finance civil society, technical organizations provide infrastructure and companies sponsor meetings. Every material support channel can shape visibility. Transparent role separation protects the forum from claims that a financial contribution became a proxy mandate.

MAG membership demonstrates the ambiguity of personal capacity

The IGF Multistakeholder Advisory Group prepares the annual programme. Its terms of reference say members serve in their personal capacity but are expected to have established linkages with their stakeholder groups. The MAG includes members from government, private sector, media, civil society and the technical community.

This design seeks two goods at once. Personal capacity can free members to deliberate rather than defend rigid instructions. Stakeholder linkage brings knowledge, networks and some accountability to communities outside the room. The combination is productive but ambiguous.

A private-sector MAG member may be employed by a company, nominated through a business network and appointed by the UN Secretary-General, yet formally serve personally. Who is the principal? The employer authorizes time and may hold policy preferences. The nominating network expects linkage. The appointing institution defines duties. None necessarily gives binding instructions.

The correct description is not that the member represents every business. The member contributes private-sector experience, consults relevant networks and owes duties under the MAG terms. Public biographies should identify employment, nomination, association roles and consultation practices. Meeting interventions should disclose a direct corporate interest where relevant.

Personal capacity should not become a shield against accountability. Members can report outreach, inputs received, conflicts and reasons for decisions without pretending they were delegates bound by a vote. Stakeholder groups can evaluate whether linkages were genuine and decide whether to nominate the person again.

Nor should employer affiliation be treated as conclusive control. A member may disagree with the company, and the MAG's rough-consensus practice may require compromise. Evidence of actual conduct matters more than inference from a business card.

The MAG example offers a general lesson: multistakeholder institutions often appoint knowledgeable individuals where no formal electorate exists. Legitimacy then depends on transparent selection, balanced composition, consultation, conflict rules and limited authority. It should not be manufactured by calling appointees representatives of a vast sector.

The NETmundial principle points toward issue-specific authority

The 2014 NETmundial Multistakeholder Statement said stakeholder roles and responsibilities should be interpreted flexibly with reference to the issue under discussion. That formulation is a useful correction to permanent stakeholder silos.

Issue-specific authority recognizes that the relevant business community changes with the question. Submarine-cable resilience calls for operators, landing-station owners, equipment vendors and coastal authorities. Domain-name policy calls for registries, registrars, registrants and business users. Platform-content rules call for services, advertisers, creators, moderators and affected communities. No company or association has equal standing across all of them.

Flexibility, however, must not mean informality without accountability. If roles change issue by issue, the forum needs a stronger record of why a entity is relevant and whom the person can authorize. Otherwise organizers can select a convenient company and describe the panel as multistakeholder.

An issue map should identify affected business models, implementation responsibilities, costs, dependencies and conflicts. Organizers can then invite several relevant perspectives and mark gaps. Associations can state which members fall within the issue and how the position was approved. Companies can distinguish operational evidence from broader policy advocacy.

Issue-specific authority also prevents category capture. A handful of globally visible technology firms should not become the default “private sector” on every digital question. Telecommunications incumbents should not define software markets. Vendors should not speak as purchasers. Companies headquartered in wealthy markets should not substitute for businesses operating under different infrastructure and legal conditions.

The flexible model is demanding because it requires institutional work. Its reward is better evidence and less theatre. The private-sector seat becomes a place for relevant interests, not a ceremonial token offered to whichever executive is available.

A representation scope table can make claims auditable

Every private-sector intervention can be classified through a simple scope table.

Speaking basis	Verifiable principal	Proper claim	Evidence required	Claim that exceeds mandate
Company delegate	Named employer	The company's approved position or experience	Role, authorization and relevant operating evidence	“All businesses agree”
Association officer	Current governed members	Position adopted within association scope	Member list, bylaws, approval method and dissent	“The entire sector supports”
Coalition spokesperson	Named endorsing organizations	The signatories' agreed text	Signatory list, drafting and endorsement record	“Global business demands”
MAG or forum appointee	Appointing terms; no sector-wide principal	Personal judgment informed by stakeholder linkages	Appointment, affiliations, outreach and conflicts	“I am mandated by the private sector”
Technical expert	Employer or self; expertise rather than constituency	Evidence about systems and implementation	Method, data boundary and conflicts	“Users authorize this policy”
Sponsor representative	Named sponsor	Sponsor's view, if separately invited to speak	Sponsorship benefits and speaker-selection basis	“Funding gives a sector mandate”

The table does not rank one basis above another. A technical expert may offer the most important evidence in a debate while holding no representative mandate. An association may possess a strong mandate but present weak evidence. Authority and accuracy remain separate.

The scope should be attached to the intervention, not inferred from an organization's reputation. A delegate may speak for an employer in one session, an association in another and personally on the MAG. The active role must be explicit each time.

Forums can implement the table without bureaucratic delay. A speaker form can require structured answers and publish them beside the programme. Session reporters can attribute claims using the selected basis. Corrections can update the record when a mandate was misstated.

The most important column is the last. Overclaiming usually occurs through one linguistic jump: company to industry, member association to all firms, customer base to users, appointment to representation, or sponsorship to endorsement. Naming the prohibited jump before the meeting makes later summaries more accurate.

Business consensus should be demonstrated, not staged

There are issues on which broad business agreement is real. Companies may share an interest in interoperable networks, predictable law, secure infrastructure and cross-border commerce. Associations can discover and articulate such common ground. A shared position becomes more credible when competing business models endorse it through a documented process.

Consensus should mean more than absence of objection on a mailing list. The record should show who received the proposal, how long they had to respond, which members participated, the threshold used, changes made after challenge and any reservations. For a diverse association, weighted and unweighted results may both be relevant.

Silence deserves careful treatment. Members may be indifferent, lack staff, disagree privately or assume a statement will proceed. Counting every non-response as consent inflates mandate. Associations can use formulations such as “approved by the board after member consultation; three members recorded reservations” rather than “industry unanimously agrees.”

Minority reports are not institutional failure. They reveal where regulation creates different costs. A dissent appendix can improve the majority's argument by showing that alternatives were considered. It also prevents members from leaving merely to preserve their ability to speak.

Consensus claims should expire. Technology, membership and markets change. A position approved in 2016 cannot automatically authorize a 2026 intervention merely because the language resembles it. Associations need review dates and triggers for reconsideration.

Finally, organized business should disclose who did not participate. Non-members, small firms, informal providers and businesses from unrepresented regions remain outside the mandate. A candid perimeter statement is more credible than a universal slogan.

The objective is not to make business speak timidly. It is to make collective speech valuable. A position whose principal, process and perimeter are visible can be weighed against competing interests and implemented with a clearer understanding of support.

Four authorization failures recur in multistakeholder debate

The first failure is vertical inflation. An employee's authorized company position becomes an association view, then a sector view, then “the private sector's” view as it passes through panels and summaries. No one necessarily makes a false statement at the start. The error accumulates because each retelling drops the principal. The remedy is persistent attribution from transcript to final report.

The second is horizontal substitution. A visible company is asked to stand in for adjacent business models. A platform is treated as the technology sector; a mobile group as all connectivity providers; a registry as the domain-name market; a multinational as small enterprise. Operational expertise in one layer does not authorize policy claims for another. Organizers should map the affected market before selecting the speaker and mark missing segments.

The third is downstream appropriation. A company invokes customers, creators, workers or local businesses as beneficiaries and gradually describes itself as their representative. Evidence that a policy affects those groups is legitimate. Representation requires a separate mechanism through which those people selected, instructed or could remove the speaker. Advisory councils and surveys may inform a company without making it an agent of respondents.

The fourth is procedural laundering. A forum's invitation, an advisory appointment or a place in a stakeholder-balanced programme is treated as validation of the speaker's mandate. Selection proves only that organizers chose the person under their criteria. It may confer duties to the forum and access to deliberation. It cannot supply authorization from absent companies.

These failures have different consequences. Vertical inflation corrupts the historical record. Horizontal substitution hides distributional conflicts. Downstream appropriation turns market relationships into political consent. Procedural laundering lets institutions certify their own representativeness by pointing to the people they selected.

They also require different evidence. Attribution records address inflation. Membership and market maps address substitution. Independent constituency procedures address appropriation. Appointment terms and outreach reports address procedural claims. A generic declaration that a meeting was multistakeholder addresses none of them.

The failures are easiest to detect when the claimed principal is named. If no legal person, membership body, signatory group or appointing instrument can be identified, the intervention is probably stakeholder speech rather than representative action. That is not a reason to exclude it. It is a reason to stop enlarging its mandate through shorthand.

The forum's record should follow the mandate through to the final summary

Representation can be accurate at the microphone and distorted in the report. A speaker may say “our company supports,” while a rapporteur writes “the private sector supported.” A panel with three aligned companies may be summarized as cross-industry consensus even though all share a business model. A moderator may invite “the user perspective” from a consumer platform and thereby confer a mandate the company did not claim.

Attribution rules should survive every stage. Transcripts and reports should name the speaker and principal. Broader labels should be used only when a recognized association or documented consultation supports them. Automated summaries and media releases need the same discipline.

The forum should preserve counterevidence. If another operator disputes feasibility, the disagreement belongs in the report. If small firms were absent, the report should mark the gap. If a sponsor held no agenda rights, that limitation should accompany sponsorship disclosure rather than leaving readers to infer control.

Public records can also connect interventions over time. Repeated changes in a company's position may reflect new evidence, commercial strategy or membership decisions. An association can publish revisions and member votes. Institutional memory prevents a single prominent intervention from becoming an assumed sector norm.

This record is especially important at a non-binding forum. The IGF's influence travels through reports, relationships and later policy. A misattributed collective noun may be cited by regulators long after listeners have forgotten who spoke. Accurate provenance is therefore not clerical detail; it is part of the forum's substantive integrity.

The strongest defence of broad business voice is still bounded

Supporters of expansive business representation make a serious argument. Firms employ millions of people, connect billions of users and invest enormous capital. Trade networks aggregate practical knowledge across economies. Requiring a formal vote for every intervention could make participation slow and legalistic while governments and activists speak more freely.

The answer is proportionality. A company need not poll every employee to describe its own investment plan. An association need not publish confidential deliberations to state an established policy. A MAG member need not become a delegate bound by instructions. The depth of proof should rise with the breadth of the claim and the authority of the decision.

“Our company estimates this rule will add these costs” needs method and corporate attribution. “Our association's members support this option” needs membership and approval evidence. “Business worldwide rejects the proposal” needs a far broader and more inclusive process. “Users want this outcome” needs independent user evidence and still should not be called a corporate mandate.

The same proportionality addresses speed. A standing association can pre-authorize policy staff within published principles and require later ratification for new positions. Emergency interventions can state that member review is pending. Informal coalitions can list endorsers without claiming all-sector consensus.

Precision does not burden only business. Civil-society and technical delegates should meet equivalent standards. Governments should distinguish official national positions from an agency official's view. The discipline improves all categories.

Broad business participation remains essential. What must end is the conversion of economic scale into representative title. Building infrastructure creates responsibility and expertise. Serving customers creates duties. Paying for a forum creates support. None of those relationships makes the company the principal of people and organizations that never authorized it.

A private-sector representation card has eight fields

A concise record beside a speaker biography or submission can prevent most ambiguity.

1. Active role. Company employee, association officer, coalition spokesperson, appointed adviser, consultant, technical expert or sponsor representative.
2. Principal. The named employer, association membership, signatories or appointing terms authorizing the intervention.
3. Coverage. Member classes, sectors, regions and business models included and materially excluded.
4. Authorization. Board approval, member vote, committee decision, standing policy, executive instruction or personal-capacity statement.
5. Issue scope. The questions on which the mandate applies and the date on which it expires or is reviewed.
6. Participation and dissent. The proportion involved, objections recorded and whether silence was counted as consent.
7. Material interests. Commercial exposure, sponsorship, contracts or regulatory interests relevant to the recommendation.
8. Evidence boundary. Whether claims derive from company data, member evidence, commissioned research, customer surveys or personal expertise.

The card should use plain language. “Speaking for Company X under its published policy” is enough where the claim is narrow. An association claiming a global sector position should provide deeper links. A personal-capacity appointee should say explicitly that no employer or sector-wide mandate attaches.

False declarations should be correctable. A member association may challenge an officer's claim; a company may clarify that an engineer spoke personally; a forum may amend a report that expanded attribution. Corrections should remain visible.

The card also gives journalists and regulators a better citation. Instead of “the private sector warned,” they can write that a named company, a coalition of listed firms or an association representing specified members made the claim. That accuracy preserves disagreement and prevents one intervention from becoming a manufactured consensus.

From 2003 to the present, the category has outgrown its shorthand

WSIS needed stakeholder categories to break an intergovernmental monopoly over Internet policy. The private-sector category recognized a basic truth: companies build and operate systems that public policy seeks to shape. The IGF then created a durable place where business could exchange evidence with governments, civil society and technical communities.

Two decades of institutional use have exposed the shorthand's limit. The private sector is not an organization. It has no universal membership roll, election, treasury or instruction mechanism. A delegate cannot be the agent of an abstract category. The principal must be a company, a governed association, a named coalition or an appointing institution whose terms expressly limit the role.

This conclusion does not demand a global business parliament. It demands honest attribution. Companies may advocate their interests openly. Trade groups may aggregate members efficiently. Appointed experts may deliberate personally. Forums may balance stakeholder experience. Each arrangement becomes stronger when it stops borrowing authority from actors outside its mandate.

The practical rule is straightforward: a company speaks for itself; an association speaks for the members and issues its rules cover; a coalition speaks for its signatories; an appointee speaks under the appointing terms; an expert speaks from evidence. Users, non-members, absent regions and competing operators remain outside unless they separately authorized the claim.

Once this rule is applied, disagreement becomes information rather than embarrassment. Regulators can see which business models bear cost. Smaller firms can contest incumbent claims. Users can challenge corporate assertions made in their name. Associations can demonstrate real consensus instead of relying on brand breadth.

The private sector deserves a seat because it is affected, knowledgeable and responsible for implementation. It does not become the principal of the room by taking that seat.

What to watch

The first signal is whether IGF programmes and reports replace “business said” with named and scoped attribution. The second is whether ICC BASIS and other associations publish the member process behind issue positions, including participation and dissent. The third is whether sector organizations distinguish membership classes when claiming operator consensus. The fourth is whether MAG biographies clarify personal capacity, nomination, employment, outreach and conflicts. The fifth is whether corporate claims about users disclose independent research rather than treating customer scale as authorization.

The sixth signal is representation diversity inside the business category: small and large firms, access and application layers, buyers and sellers, incumbent and entrant, different regions and market conditions. More company speakers are not necessarily more representation if all share the same commercial exposure.

The final test is linguistic. Can a reader identify the principal behind every collective noun? If the answer is no, the statement may still contain valuable evidence, but its representative authority remains unproven. Internet governance can accommodate that distinction. It should welcome stakeholders, test expertise and record interests without inventing principals who never gave a mandate.