Future of Work & Skills Observatory · Sales Development
How Will AI Change Sales Representatives by 2030?
Which sales tasks can AI assist, which relationship, negotiation and judgment skills become more valuable, and how can a sales team prepare? With a 90-day plan.
hilighter Editorial Team · Published by SkillsBoard SAS, the company behind hilighter
Published · Reviewed
How will AI change sales work in your organization?
Start from your company: the free overview ranks your departments by AI exposure, then shows which tasks of your sales roles may change and which skills to build.
Free and indicative. No lead form is required to read this report.
Sales representatives are not facing one single “AI replacement” event. They are facing a redesign of the sales cycle. AI systems can research prospects, prepare account briefs, suggest outreach, organize CRM data and summarize calls. Those capabilities may compress the time spent on preparation. They do not automatically create buyer trust, understand political dynamics inside an account or make a defensible commercial commitment.
The central question is not “Will AI replace salespeople?” but “Which parts of the selling process can be delegated safely, and which new capabilities make human sellers more valuable?” This guide answers at the task level and offers skills evidence and a 90-day transition plan for B2B and SME sales teams.
Prospecting law, the quality of contact data and buyer expectations materially change the answer. A tool that can send thousands of personalized emails is not necessarily a workflow that a European company may lawfully deploy. Productivity without trust or compliance is not progress.
Apply these findings to your own organization: Explore AI Workforce Readiness with hilighter. This resource is freely readable, with no gated report or lead capture, and the assessment is currently in English. The AI readiness journey is a planning aid, not an individual employee rating or regulatory certification.
Quick answer: what changes, what matters, and what to do
| Work dimension | Likely direction | Practical response |
|---|---|---|
| Prospect research and CRM hygiene | Faster briefs, enrichment suggestions, reminders and draft updates | Verify contacts, source provenance and CRM write permissions |
| Outreach and follow-up | AI-assisted drafts and segmentation; possible agent orchestration | Keep meaningful personalization, truthful claims and lawful channel selection |
| Discovery and qualification | Notes, question suggestions and next-step capture | Listen, validate needs, avoid invented commitments |
| Proposals and forecasting | First drafts and structured analyses from approved sources | Review pricing, margins, assumptions and factual accuracy |
| Negotiation and relationship building | More evidence and coaching support, not automatic trust | Develop stakeholder insight, negotiation judgment and accountability |
Important: the classifications in this article are original qualitative editorial judgments. They are not measured replacement probabilities, numerical risk scores or predictions of headcount reductions.
1. What the evidence actually supports
Salesforce’s 2026 State of Sales survey, involving more than 4,000 sales professionals, reports widespread AI use and high expectations for sales agents. The headline finding that 87% of surveyed sales organizations use some AI is a vendor-sponsored survey result, not an audited measure of worldwide adoption. Respondents expected agents to reduce prospect research and email preparation time, but expectations are not realized time savings. [1]
LinkedIn’s 2025 B2B sales research reports that 56% of professionals in its study used AI daily and that some sellers associated AI with better lead research or response rates. These are self-reported and observational measures; stronger performance among AI users could also reflect team maturity, territory quality or existing sales practices. [2]
Gartner’s 2025 prediction that 75% of B2B buyers would prefer experiences prioritizing human interaction by 2030 is a forecast, not a measured 2030 outcome. It is a useful warning against treating the entire purchasing conversation as an automation target. [3]
The ILO’s 2025 occupation framework evaluates exposure at the task level and emphasizes transformation rather than mechanically equating exposure with job loss. The WEF’s estimate of 39% skill change by 2030 is an economy-wide employer expectation, not a sales-representative-specific rate. [4][5] Microsoft’s 2026 Work Trend Index similarly emphasizes that organizational systems and manager support influence outcomes, beyond individual access to tools. [9]
Sales-specific evidence is still dominated by surveys, demonstrations and case studies from vendors. Teams should pilot on their own customer segments and measure qualified opportunity quality, opt-outs, buyer satisfaction and margin, not just message volume or hours saved.
How to read the evidence
Large industry surveys describe respondents’ adoption, expectations and perceptions. Controlled experiments offer stronger causal evidence but often test one tool in one workflow. Task-exposure studies identify possible technical overlap, not adoption, safety, business value or actual job displacement. hilighter does not convert any of these into a single percentage of a profession “at risk.”
2. Define the profession before assessing it
A job title is an imperfect proxy for what employees do. The relative importance of each task changes with company size, business model, geography, customer segment, technology stack and seniority. Analyze your own tasks before choosing training or software.
- Sales development representative (SDR/BDR): Focuses on target accounts, initial research and meetings; lead sourcing and first-draft messaging may change quickly, but consent and brand reputation are central.
- Account executive (AE): Runs discovery, complex demos, proposals and multi-stakeholder opportunities. AI can assist preparation but must not invent commercial facts or substitute for commitments.
- Account manager / renewals specialist: Maintains long-term relationships, service continuity and commercial negotiations; trust, account context and escalation quality are central.
- Inside sales / transactional sales: May handle many standardized inquiries and product configurations; AI can help with routing and proposal drafts when catalog and stock data are governed.
- Enterprise / strategic sales: Often handles high uncertainty, procurement and legal complexity; stakeholder navigation and negotiation are hard to standardize.
- Field sales / channel partner manager: Combines visits, local market understanding and partner enablement; AI assistance may improve administration but not physical interactions and local context.
3. Task-by-task AI impact map
How to read the map: each task has two separate ratings. Change potential: High workflow change means relatively repeatable activities that may be assisted or partially automated with good data and review; Medium / conditional means useful AI help depends strongly on context and systems; Human-led / consequential means tools may prepare inputs, but accountable human control remains central. Risk and trust is a separate dimension (Low, Medium, High, Critical): it addresses privacy, accuracy, rights, fairness and the cost of an error. Neither rating is a probability.
| Sales Representative task | Change potential | Risk and trust | What AI can contribute | What people must check |
|---|---|---|---|---|
| Research prospects and target accounts | High workflow change | Medium | Summarize verified company websites, annual reports and CRM history into briefings. | Verify dates and sources; do not infer private buying intent or invent organization details. |
| Update CRM records and activity notes | High workflow change | Medium | Draft call notes, proposed fields and structured follow-up tasks. | Confirm identity matches, permissions, CRM writes and sensitive information. |
| Prepare meeting agendas and discovery questions | High workflow change | Low | Suggest agendas, competitor questions and specific open-ended prompts. | Validate against opportunity stage and actual customer needs. |
| Draft personalized prospect outreach | High workflow change | High | Propose relevant subject lines and first drafts from authorized facts. | Check legal basis, ePrivacy, accuracy, tone and opt-out requirements; block mass-spam workflows. |
| Prepare meeting follow-ups and reminders | High workflow change | Medium | Summarize customer-agreed next steps and suggest follow-up wording. | Verify commitments and dates; do not claim the customer promised something they did not. |
| Prioritize leads and estimate fit | Medium / conditional | High | Identify patterns and build reviewable prioritization suggestions. | Audit for bias, stale data, proxy variables and self-reinforcing feedback loops. |
| Document discovery and objections | Medium / conditional | High | Structure recorded or written notes and surface missing questions. | Use permitted recordings; confirm quotes, context and recording notices. |
| Build proposals and product comparisons | Medium / conditional | High | Assemble first drafts from approved catalog, price and technical sources. | Confirm pricing, availability, product suitability, terms and sign-off. |
| Forecast pipeline and deal probabilities | Medium / conditional | High | Run scenarios and highlight unusual CRM patterns. | Challenge small samples, stage inflation and misleading certainty. |
| Practice objection handling and sales role-plays | Medium / conditional | Medium | Provide simulation, coaching prompts and feedback ideas. | Ensure feedback is developmental, contextual and not a secret performance score. |
| Negotiate multi-party commercial terms | Human-led / consequential | Critical | Prepare concessions, risk scenarios and alternative terms. | Human seller owns margins, commitments, ethics and stakeholder dynamics. |
| Build trust and resolve complex account issues | Human-led / consequential | Critical | Prepare customer histories and issue summaries. | Human empathy, reputation, conflict handling and accountability remain decisive. |
Five workflows worth examining first
Prospecting is moving from search to verification. Research assistants may surface company facts in minutes, but source freshness matters. A sales representative should learn to reject attractive but unsupported signals such as invented expansions, nonexistent executives or scraped personal data. The skill shift is from manually assembling lists to validating fit, timing and permitted contact methods.
Outreach volume is a dangerous optimization target. A prompt can produce hundreds of plausible emails. That does not mean the company has permission to send them, or that recipients will value them. The better workflow tests relevance, sender reputation, unsubscribe handling, reply quality and account progression.
Discovery becomes more analytical, not less human. Conversation tools can summarize calls and propose follow-up questions, but they can miss irony, contradictions and political context. The seller must still test assumptions, ask uncomfortable questions and obtain consent where recordings are used.
Proposals become an evidence-governance problem. Generative drafts reduce boilerplate work only if price lists, approved claims and product constraints are trustworthy. Build source-of-truth links and human approval, especially when proposals create legal obligations.
Negotiation gains information support. AI may model concession scenarios or rehearse objections, but it should not commit the company to unapproved discounts or misrepresent the customer’s negotiating position. Train sellers to see AI suggestions as hypotheses.
4. Skills and work habits that may lose relative value
Important distinction: a declining need for a routine activity does not make an employee or the underlying profession obsolete. Many foundational skills remain necessary to catch AI errors. These are activities whose standalone market value could decrease, not skills people should stop learning entirely.
- Manual company-fact compilation: Copying public snippets into briefing decks is increasingly automatable, although verification remains valuable.
- Unstructured note retyping: Routine recaps and CRM formatting may be assisted; accuracy review and permission management still matter.
- Generic email template tweaking: Basic variation is easy to generate; targeting, distinctive insight and legally compliant outreach carry the value.
- Activity-volume reporting: Counting calls and emails without assessing buyer outcomes becomes less useful when AI can amplify activity.
- Static objection scripts: Rigid scripts are weaker than adaptive questioning and contextual problem-solving; product truth must remain verified.
- Manual stage-by-stage follow-up scheduling: Reminders can be automated, while the seller retains responsibility for promises and relationship timing.
5. Eight priority skills to develop
The useful standard is not whether someone has attended an AI webinar, but whether they can demonstrate a safe, repeatable outcome on relevant work. The suggested evidence below is illustrative and should be adapted by employees and managers.
1. Evidence-based account intelligence
Verify the source, freshness and confidence of every commercially relevant fact rather than repeating fluent AI summaries. Evidence of competence: A cited account brief with corrected false positives and decision-relevant insights. First practice: Compare one AI-produced account brief with three original sources.
2. Human–AI sales workflow orchestration
Delegate narrowly scoped research, notes and drafting tasks while maintaining approval points and an audit trail. Evidence of competence: A mapped workflow showing permissions, triggers, review and rollback. First practice: Design one post-meeting follow-up pipeline with a mandatory human approval.
3. Privacy-aware and compliant prospecting
Understand the difference between public company information and personal contact data, and between technical ability to contact and legal permission. Evidence of competence: A documented channel and contact-data review with suppression and escalation rules. First practice: Review ten sample leads for provenance and channel permission.
4. Consultative discovery and business diagnosis
Ask questions that uncover real constraints, conflicting stakeholders and the cost of inaction. Evidence of competence: A recorded or simulated discovery conversation with independently scored questions and listening quality. First practice: Role-play a discovery call without reading generated questions word-for-word.
5. Product, pricing and claim validation
Detect invented numbers, product capabilities and price commitments before they reach a buyer. Evidence of competence: A signed-off proposal checklist with source references and margin validation. First practice: Audit a synthetic AI proposal seeded with five incorrect claims.
6. Commercial reasoning and forecast literacy
Challenge pipeline probabilities and data quality rather than accepting a polished predictive number. Evidence of competence: A forecast review showing assumptions, alternative scenarios and calibration notes. First practice: Compare forecast changes against actual close outcomes for one cohort.
7. Negotiation and stakeholder mapping
Build options and concessions around verified buyer needs and relationship dynamics. Evidence of competence: A stakeholder map and negotiation plan reviewed against agreed commercial boundaries. First practice: Simulate procurement objections and prepare two ethical trade-offs.
8. AI-assisted role practice and feedback
Use AI simulations to practice but judge progress through observable customer-facing behaviors. Evidence of competence: A pre/post role-play rubric with examples of stronger listening and objections. First practice: Have a manager assess two simulations independently from AI feedback.
6. Emerging responsibilities already becoming plausible
These are work responsibilities, not necessarily universally established job titles. In a small business, one employee may combine several of them.
- Sales AI-agent workflow steward. Defines approved automations, monitors failures, manages handoffs and prevents unintended CRM or customer contact actions. Maturity: emerging responsibility, not a standardized occupation.
- Buyer signal verification specialist. Audits intent signals and enrichment claims before they influence outreach or deal allocation. Maturity: emerging responsibility, not a standardized occupation.
- Commercial knowledge and claims curator. Maintains approved product truths, policy updates, prices and source attribution for sales assistants. Maturity: emerging responsibility, not a standardized occupation.
- Human–AI sales enablement coach. Builds responsible simulations and checks that tool feedback actually improves discovery and qualification. Maturity: emerging responsibility, not a standardized occupation.
- AI-assisted outreach trust lead. Tracks unsolicited-contact risk, customer confusion and disclosure/consent patterns across channels. Maturity: emerging responsibility, not a standardized occupation.
7. Future skills that may appear by 2028–2030 (hypotheses, not facts)
The following are scenario hypotheses. Their labels, timing and qualitative confidence are hilighter editorial assessments. They have not been independently observed at scale and should not be used to promise a career outcome.
- Multi-agent deal-workflow governance. Medium confidence (conditional, 2028–2030). Coordinate specialized research, proposal and follow-up agents while preserving customer-facing responsibility. What would validate it: Real deployments with documented accountability, safeguards and measurable buyer benefit.
- Buyer–seller AI negotiation protocol literacy. Low–medium confidence (conditional, 2028–2030). Understand negotiation where procurement and vendor agents exchange structured requirements. What would validate it: Repeatable buyer-side deployments and transparent contractual boundaries.
- Provenance-preserving personalization. Medium confidence (conditional, 2028–2030). Attach source and lawful-use evidence to every personalized commercial assertion. What would validate it: Vendor tooling and compliance reviews increasingly require verifiable provenance.
- Synthetic deal-simulation assurance. Low–medium confidence (conditional, 2028–2030). Evaluate whether simulations of buyer objections represent reality rather than a biased model narrative. What would validate it: Independent benchmark studies correlate simulations with real outcomes.
- Commercial AI decision traceability. Medium confidence (conditional, 2028–2030). Explain why a recommendation, forecast or agent action was accepted or rejected. What would validate it: Customers and enterprise procurement demand reproducible decision logs.
8. A real workflow: before and after thoughtful AI adoption
This example is a process-design illustration, not a live hilighter customer result or a promise of cost savings. Teams should compare it with their baseline and protect quality, not simply reduce handling time.
From prospect discovery to a qualified first meeting
| Stage | Conventional approach | AI-assisted approach | Control that must remain |
|---|---|---|---|
| 1. Identify target | Seller manually scans broad lists. | AI drafts a shortlist from approved ICP and permissible sources. | Seller confirms business fit and data provenance. |
| 2. Research account | Manual tab hopping and copying. | Assistant drafts cited factual account brief. | Verify facts and do not infer private intentions. |
| 3. Choose channel | Default to generic outreach. | System suggests channels by account context. | Check jurisdiction, lawful basis and opt-out status. |
| 4. Draft invitation | Edit a standard template. | AI proposes a precise, short first draft. | Human checks relevance, claims and tone. |
| 5. Prepare meeting | Collect notes and questions. | AI suggests agenda and unanswered questions. | Seller prioritizes actual buyer problems. |
| 6. Document outcome | Type recap after meeting. | AI organizes notes and proposes CRM updates. | Confirm statements, consent and CRM writes. |
| 7. Advance opportunity | Follow up from memory. | AI drafts agreed next actions. | Human owns commitments and timing. |
The redesigned process succeeds only when relevant meetings and trustworthy opportunities improve. Faster spam, fabricated personalization or unauthorized data collection are failures, even if messages per hour increase.
9. How the profession could evolve in 2026, 2028 and 2030
| Time horizon | Reasonable scenario | What remains uncertain |
|---|---|---|
| 2026: assisted selling | Research, transcription, first drafts and CRM suggestions expand in some teams. | Accuracy, buyer tolerance, regional compliance and integration costs vary. |
| 2028: governed orchestration | Some teams connect tightly scoped agents to account workflows with human approval. | Agent reliability, governance costs and procurement expectations may slow adoption. |
| 2030: relationship-centered differentiation | More standardized preparation may be automated; discovery, trust and negotiation could differentiate sellers. | Buyer preference and market power may evolve differently by segment. |
A 2030 scenario is not a forecast. Some sectors will adopt slowly because of legacy systems, customer expectations, regulation, data access or economics; others will redesign workflows quickly. A cautious employer should plan for multiple paths rather than a single predicted future.
10. Governance, trust and employee safeguards
- Consent and lawful contact: EU marketing rules involve GDPR, ePrivacy and national implementation; a public email address or scraped LinkedIn profile is not universal permission for cold outreach. [7][8]
- Account and contact-data provenance: log origins and usage rights; avoid enrichment sources with unverifiable permission or unnecessary personal details.
- Agent autonomy: start with drafts and recommendations; prohibit automated sending, discounting, contract promises and CRM deletion without explicit scoped authority.
- Truthful claims: approve product features, customer evidence, pricing and proof points; flag when the model cannot verify an assertion.
- Call recording and customer privacy: assess notices, local rules, access, transcription retention and vendor data use before enabling meeting AI.
- Fair allocation: audit how AI prioritizes accounts or leads so historical coverage patterns do not silently determine future opportunity.
- Buyer transparency: evaluate whether an automated agent is directly interacting with prospects and applicable AI Act disclosure duties. [6]
- Human and business outcomes: monitor complaints, opt-outs, conversion quality, trust and lost deals, not just activity velocity.
Legal note (EU context): AI Act Article 50 includes transparency rules for certain direct interactions with AI and specified synthetic-content scenarios, with exceptions and context-specific obligations [6]. The GDPR and national rules continue to apply to personal-data processing; in sales and marketing, ePrivacy and local direct-marketing laws may independently restrict outreach [7]. A website article is not legal advice. Verify the rules relevant to your specific workflow and jurisdiction.
11. A practical 90-day action plan for teams
| Phase | Team actions | Individual learning activity | Evidence / go–no-go test |
|---|---|---|---|
| Days 1–30: map and baseline | Choose a single sales segment; map 12 tasks; assess approved sources, permissions, customer communications and current KPIs. | Practice source-checking, data minimization and a supervised research brief. | Approved scope, contact-law review, baselines for qualified meetings and complaints. |
| Days 31–60: run controlled pilots | Pilot account briefs and post-meeting drafts for a small group; run offline quality reviews before any external AI action. | Use a role-play rubric; audit an AI-written proposal and outreach sample. | No unsupported claims; documented reviewer sign-offs; buyer quality no worse than baseline. |
| Days 61–90: evaluate and scale selectively | Compare qualified opportunities, conversion, time saved, trust indicators and seller feedback; redesign failed workflows. | Share validated prompts, failure patterns and better negotiation practice. | Proceed only if business outcomes and privacy checks are acceptable; retain rollback. |
Simple organizational scorecard
| Metric | Why it matters | Common measurement error |
|---|---|---|
| Qualified opportunities per 100 lawful contacts | Measures buyer relevance instead of message volume. | Calling all replies or meetings qualified. |
| Discovery-to-proposal conversion | Signals whether preparation improved understanding. | Ignoring lead mix or territory differences. |
| Proposal error and correction rate | Protects commercial trust and margin. | Counting drafts saved without checking facts. |
| Opt-out / complaint rate | Detects buyer backlash or inappropriate channels. | Hiding complaints because emails were not opened. |
| CRM accuracy audit | Tests whether automation improves data quality. | Assuming AI-written fields are verified. |
| Seller time in live customer conversations | Tests reallocation to high-value work. | Confusing more meetings with better customer outcomes. |
Review results across case types, languages, customer groups and seniority levels rather than treating a single average as proof of improvement. Include a baseline, independent spot checks, escalation quality and an explicit rollback plan.
12. Frequently asked questions
Will AI replace sales representatives by 2030?
There is no defensible universal replacement percentage. Repeatable research and drafting tasks are more exposed than trust-intensive negotiation and complex account management. Company-specific workflow design matters more than a job title. [4]
Which sales tasks are easiest to automate?
Account summaries, routine CRM entry suggestions, meeting recaps, standardized product Q&A and draft emails are practical starting points when data and approvals are reliable.
Will cold outreach agents make human sellers unnecessary?
No such conclusion follows from current research. High message volume can harm deliverability, reputation and lawful contact requirements. Relevance and customer trust remain constraints.
What is the most important new skill for an SDR?
Verified account intelligence combined with compliant, relevant outreach. Fluency with a generation tool is useful but insufficient without source and contact-method judgment.
Can AI negotiate prices with prospects?
Systems can simulate options or calculate permitted discounts; external commitments need explicit company authority, current pricing rules and accountable review.
Can we automatically score every lead?
Technically possible in some CRM products, but scores require quality checks, bias review, current data and careful analysis of whether they improve outcomes.
How should SME sales teams start?
Pick one narrow task, ideally account-brief preparation or post-meeting notes. Baseline time and quality, train a few users and expand only after checking errors and customer results.
How can hilighter help?
Use the AI Workforce Readiness journey to examine your real sales roles and skills gaps, then validate findings with managers and employees. The guide does not claim automatic hiring evaluation or regulated AI certification.
13. What organizations should do next
For sales teams, the opportunity is to reduce low-value preparation without degrading the customer relationship. The wrong ambition is to produce maximum automated activity. The right ambition is to create better-informed conversations, more credible proposals, stronger seller capability and measurable customer value.
An effective transition starts with task mapping, verified skills and shared controls. It should be reviewed with actual sellers, sales operations, legal/privacy specialists and customer-facing leaders. Invest in consultative discovery, negotiation and commercial reasoning alongside AI tool use.
Ready to examine your actual jobs rather than a generic profile? Explore your organization’s AI Workforce Readiness with hilighter. Identify tasks, document current skills, validate capability gaps with managers and employees, and turn the results into a practical development plan. The assessment is currently in English. It does not automatically preselect a role and does not provide an official legal-risk rating.
Methodology, limitations and editorial disclosure
This is a research-informed, scenario-based editorial resource, created for the hilighter Future of Work & Skills Observatory (SkillsBoard SAS). Task categories and future skill confidence labels are editorial classifications. The report is not a representative employer survey, an assessment of any particular employee, a labor-market forecast, a technical benchmark of AI products, legal advice or a guarantee that a job will or will not disappear. The sources include independent research, vendor-sponsored surveys and regulatory guidance. Those evidence types are not interchangeable. Any vendor-reported percentage is attributed to the respective survey, with its denominator where available. Publication date: 5 November 2026; verify material new developments before future updates.
Editorial maintenance: refresh source links, legal interpretations and examples at least annually, and whenever a material regulatory or technological change occurs. Identify substantive edits and their dates rather than silently changing the reported evidence.
Sources and further reading
- Salesforce, State of Sales Report 2026 (3 February 2026), survey of more than 4,000 sales professionals. https://www.salesforce.com/news/stories/state-of-sales-report-announcement-2026/
- LinkedIn, The ROI of AI: research on how AI is transforming B2B sales (11 March 2025). https://news.linkedin.com/2025/the-roi-of-ai--new-research-on-how-ai-is-transforming-b2b-sales
- Gartner, Prediction on preference for human interaction in B2B sales by 2030 (25 August 2025). https://www.gartner.com/en/newsroom/press-releases/2025-08-25-gartner-says-by-2030-that-75-percent-of-b2b-buyers-will-prefer-sales-experiences-that-prioritize-human-interaction-over-ai
- International Labour Organization, Generative AI and Jobs: A Refined Global Index of Occupational Exposure (20 May 2025). https://www.ilo.org/publications/generative-ai-and-jobs-refined-global-index-occupational-exposure
- World Economic Forum, The Future of Jobs Report 2025 (2025). https://www.weforum.org/publications/the-future-of-jobs-report-2025
- European Commission, Guidelines on transparency obligations for providers and deployers of AI systems (Article 50) (2026). https://digital-strategy.ec.europa.eu/en/policies/guidelines-ai-transparency-obligations
- European Commission, Can data received from a third party be used for marketing? (accessed November 2026). https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/legal-grounds-processing-data/can-data-received-third-party-be-used-marketing_en
- European Data Protection Board, Guidelines 1/2024 on processing of personal data based on legitimate interest (2024). https://www.edpb.europa.eu/system/files/2024-10/edpb_guidelines_202401_legitimateinterest_en.pdf
- Microsoft, 2026 Work Trend Index: Agents, human agency, and opportunity (5 May 2026). https://www.microsoft.com/en-us/worklab/work-trend-index/agents-human-agency-and-the-opportunity-for-every-organization
- NBER, Shifting Work Patterns with Generative AI (revised November 2025). https://www.nber.org/papers/w33795
- NIST, AI Risk Management Framework: Generative AI Profile (26 July 2024). https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence