Telephone Operators
43-2021.00Median wage $41,740/yr3,430 employed (US)Rank #21 of 923 scored · top 2% by substitution
Provide information by accessing alphabetical, geographical, or other directories. Assist customers with special billing requests, such as charges to a third party and credits or refunds for incorrectly dialed numbers or bad connections. May handle emergency calls and assist children or people with physical disabilities to make telephone calls.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution68
Exposure66
Augmentation38
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
14 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
57%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%66
panel mean rating 3.6/5 → substitution pressure 66/100
Technical feasibility todayw 20%65
panel mean rating 3.6/5 → substitution pressure 65/100
Cost vs. human wagew 15%77
panel mean rating 4.1/5 → substitution pressure 77/100
Adoption barriersw 20%inverted — strong barriers lower the score69
panel mean rating 2.2/5 (barrier strength) → substitution pressure 69/100
Sector adoption velocityw 10%64
panel mean rating 3.6/5 → substitution pressure 64/100
Task breakdown (14 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Observe signal lights on switchboards, and dial or press buttons to make connections.
100CI 100–100 · exposure 100 · augmentation 0 · importance 4.6/5 · click for rater detail
Observe signal lights on switchboards, and dial or press buttons to make connections.
100| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | The telephony and telecommunications sectors (highly digitized, information-intensive, and cost-driven) adopted automated switching and IVR systems decades ago. Adoption is essentially complete; the human telephone operator role has been almost entirely displaced in developed economies. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Telecommunications switching automation is essentially fully adopted industry-wide; this task has been automated at scale for a long time already. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | There is no meaningful role for human assistance or augmentation in automated call routing and switching. The task is fully solved by automation; human telephone operators are not part of the modern call-handling workflow in any significant capacity. |
| Augmentation potential | claude-sonnet-5 | 1/5 | Since the task itself has been almost entirely automated away, there is little remaining human task to augment with AI tools. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | The task of observing signal lights and making connections via button presses is entirely mechanical and rule-based. Modern IVR and call-routing systems have automated this function completely, delivering well over 50% time savings at equal or superior quality—connections are faster, more reliable, and require no human intervention. |
| Task automatability | claude-sonnet-5 | 5/5 | This is a highly routinized, physical-signal-monitoring and connection task that automated switching systems and modern telecom infrastructure have essentially eliminated the need for human labor to perform.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | There are no regulatory, licensing, or legal barriers requiring a human telephone operator to perform call routing and switching. Automation of this task has been legally and commercially standard for over 40 years, with minimal friction to further deployment. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-contact requirement protects this task; it has no legal or regulatory requirement for human performance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Modern automated switching and IVR systems cost pennies per call or connection in most enterprise and telecom environments, compared to the loaded wage of a human telephone operator (typically $30k–$50k annually). The cost ratio is orders of magnitude in AI's favor. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated electronic switching is vastly cheaper per call/connection than paying a human operator to monitor lights and manually connect calls. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Automated call routing, IVR systems, and PBX switching have been in production at scale for decades across virtually all telephony systems. These technologies reliably perform the core task (signal detection and connection-making) with demonstrable success in millions of deployments. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated switching equipment and digital exchanges have already replaced manual switchboard operation in production telecom networks for decades. |
Listen to customer requests, referring to alphabetical or geographical directories to answer questions and provide telephone information.
100CI 100–100 · exposure 100 · augmentation 25 · importance 4.2/5 · click for rater detail
Listen to customer requests, referring to alphabetical or geographical directories to answer questions and provide telephone information.
100| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Telephone operator roles have already been largely displaced by IVR and automated directory services across telecom and customer-service sectors; adoption is complete in mature markets. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Telecom and customer service sectors adopted automated directory/IVR systems rapidly and pervasively long ago, representing one of the most complete legacy automation cases. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | While AI assistants can support human operators in complex scenarios (suggesting answers, retrieving data), the task's nature—simple lookup and information provision—offers limited productivity-enhancement upside beyond full automation. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Since the task is almost fully automated already, there is little residual human-in-the-loop task left for AI to meaningfully augment. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | This task is highly automatable end-to-end: listening to requests can be handled by speech recognition, directory lookups are straightforward database queries, and response delivery via text-to-speech or chatbot is mature. IVR systems and modern voice assistants routinely perform all components with >50% time savings. |
| Task automatability | claude-sonnet-5 | 5/5 | Directory assistance is a well-structured lookup task that conversational AI/IVR systems have automated end-to-end for decades, easily exceeding the 50% time-savings threshold. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No licensing, regulatory, or legal requirement mandates a human operator for directory and informational services; customers generally accept automated responses, and there are no liability constraints on automation. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement mandates a human operator for directory information; this task has been substituted for decades with minimal friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Cloud-based speech recognition, NLP, and directory-lookup infrastructure cost pennies per query, while a human telephone operator costs $15–25+ per hour, delivering at least a 100× cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated lookup and voice response systems cost fractions of a cent per query versus a human operator's wage, an order-of-magnitude or greater cost advantage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed products like voice assistants (Siri, Alexa, Google Assistant) and IVR systems reliably answer directory and informational queries at scale in production today, with consistently low error rates for standard queries. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated directory assistance and voice-based customer information systems have been in mature commercial deployment at scale since the 1990s-2000s, now enhanced with modern NLP. |
Keep records of calls placed and received, and of related toll charges.
100CI 100–100 · exposure 100 · augmentation 25 · importance 3.8/5 · click for rater detail
Keep records of calls placed and received, and of related toll charges.
100| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Telecommunications and enterprise IT sectors have nearly universally adopted automated call logging and toll tracking; human operators for this function are virtually extinct in production telephony systems. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Telecom billing and call-detail automation is essentially universal already, representing one of the most fully and quickly digitized back-office functions in any industry. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | Once automated systems handle the core logging and billing, there is minimal meaningful assistance AI can offer a human; the task leaves no creative or judgment-heavy residuum for human-in-the-loop augmentation. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Since this task is essentially fully automated already, there is little remaining human task to augment; any residual human role is minimal oversight rather than active use of AI tools. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | Call logging, timestamping, and toll charge tracking are entirely data-entry and database-management tasks that modern systems (VoIP platforms, CRM integrations, billing software) automate completely with high accuracy and minimal human intervention, easily exceeding 50% time savings. |
| Task automatability | claude-sonnet-5 | 5/5 | Logging calls and toll charges is a structured data-capture and record-keeping task that is fully automatable via telecom switching/billing systems that log this data automatically without human intervention. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No licensing, authorization, or regulatory requirement mandates human record-keeping for call logs; carriers and businesses have no liability obstacle to full automation of this administrative data task. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-contact requirement blocks automated call logging; it is already the default industry practice with no legal requirement for human record-keeping. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Automated call logging and billing incur only minimal cloud compute and storage costs per call, orders of magnitude cheaper than paying a human operator wage to manually record calls and charges. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated CDR/billing systems cost fractions of a cent per call versus paying a human operator to manually log and tally call records, an order-of-magnitude or greater cost difference. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Deployed telecommunications and billing systems have reliably automated call record creation and toll tracking in production for decades; call detail records (CDRs) are now standard automated outputs across carriers and PBX systems. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated call detail recording (CDR) and billing systems have been standard in telecom infrastructure for decades and operate reliably at massive scale in production. |
Suggest and check alternate spellings, locations, or listing formats to customers lacking details or complete information.
89CI 79–100 · exposure 87 · augmentation 63 · importance 4.1/5 · click for rater detail
Suggest and check alternate spellings, locations, or listing formats to customers lacking details or complete information.
89| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Customer service and telecom sectors are rapidly automating suggestion and lookup tasks via chatbots, IVR systems, and CRM integrations; this represents mainstream adoption in information-heavy industries. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Telecom and directory services adopted automated speech recognition and self-service lookup systems extensively years ago, representing one of the most thoroughly automated customer service functions. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assists operators by automatically surfacing spelling and location suggestions in real time, reducing search effort and improving accuracy without removing the human from the interaction; operators can validate, filter, and confirm suggestions efficiently. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where human operators remain, AI-assisted search and autocomplete can speed up their lookup process, though the task is largely fully automated rather than augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI systems can readily suggest alternate spellings, locations, and formats using pattern matching, fuzzy matching, and knowledge bases; this requires minimal context and produces standardized outputs. The task reaches the automatability threshold if it handles ~50% of routine cases (common misspellings, standard location variants) without human judgment, which current NLP and lookup tools do reliably. |
| Task automatability | claude-sonnet-5 | 5/5 | This is a directory-assistance style lookup and fuzzy-matching task that current conversational AI and search systems handle natively, meeting or exceeding the 50% time-saving bar with off-the-shelf tools. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few legal or licensing barriers exist; the task is routine data-handling. Adoption barriers are mainly organizational (customer preference for human confirmation in some contexts) and operational (integration into existing phone systems), not regulatory or liability-driven. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or regulatory requirement mandates a human operator for spelling/listing lookups; this has already been largely automated industry-wide. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-powered spelling and location suggestion (API calls, in-house NLP models) cost pennies per query; the human wage for a telephone operator performing this task is much higher, making the cost ratio decisively favorable for automation. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated lookup/query systems cost fractions of a cent per query versus a live operator's wage, making AI dramatically cheaper for this narrow task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (spell-checkers, autocomplete, location-suggestion APIs, CRM systems) demonstrably perform this in production at scale across customer service platforms. Error rates on simple spelling or location variants are low; the main limitation is ambiguous or unusual edge cases. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated directory assistance, voice search, and chatbot-based lookup systems already perform spelling correction and fuzzy listing matches reliably in production (e.g., voice assistants, 411 automation, search autocomplete). |
Update directory information.
88CI 79–97 · exposure 87 · augmentation 50 · importance 4.1/5 · click for rater detail
Update directory information.
88| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Telecommunications and related sectors have rapidly adopted automated directory systems, CRM integrations, and data management platforms; directory updates are now predominantly automated in major carriers and service providers. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Telecom and directory services have largely digitized and automated such updates for decades, with modern AI/automation further accelerating adoption in this administrative niche. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-suggesting updates, flagging inconsistencies, and batch-validating entries, which meaningfully raises productivity for any remaining human review roles, though the core task is easily fully automated. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Where humans still perform verification or exception-handling for directory updates, AI tools can pre-populate and validate data, improving efficiency, though the task is mostly fully automatable rather than augmented. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Directory information updates are highly structured data-entry tasks that current AI systems can perform end-to-end with significant time savings. The task involves parsing updates, validating entries, and modifying databases—all well within the scope of current automation tools and APIs. |
| Task automatability | claude-sonnet-5 | 5/5 | Updating directory records is a structured data-entry task that current AI and automation systems (databases, scripts, LLM-assisted data pipelines) can perform end-to-end with substantial time savings and equal or better accuracy. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Minimal barriers exist: no licensing requirement, no legal mandate for human involvement, and privacy/compliance concerns are manageable through standard data governance. Organizational adoption friction is the main friction point. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There is no licensing, regulatory, or liability barrier to automating directory updates; it's a routine clerical function already largely computerized. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | AI-driven data entry and directory management costs are orders of magnitude cheaper than human telephone operator labor when amortized across volume, making this highly cost-favorable. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated data entry/update via scripts or AI agents costs a small fraction of a cent per record compared to a human operator's wage for the same repetitive task. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed products (RPA platforms, data management systems, and ML-driven validation engines) reliably perform directory updates in production environments today, though some human oversight and error-checking remain common practice. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Directory management systems with automated update pipelines, APIs, and CRM integrations are already deployed at scale across telecom and enterprise systems, requiring minimal human involvement. |
Perform clerical duties such as typing, proofreading, and sorting mail.
85CI 72–97 · exposure 83 · augmentation 75 · importance 4.0/5 · click for rater detail
Perform clerical duties such as typing, proofreading, and sorting mail.
85| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Information-intensive sectors (finance, telecom, healthcare administration) have already widely adopted automated typing, proofreading, and mail-sorting systems; displacement of telephone operators for clerical tasks is well underway. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Office/administrative functions have moderate AI adoption—proofreading and document tools are common, but comprehensive clerical automation including physical mail remains uneven across organizations. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools like autocomplete, grammar checkers, and mail flagging systems provide meaningful productivity assistance to human operators on these tasks, though the low complexity of the work means augmentation is incremental rather than transformative. |
| Augmentation potential | claude-sonnet-5 | 5/5 | AI tools like grammar/spell checkers, transcription software, and automated sorting/routing significantly boost clerical worker productivity while a human remains involved in review and physical tasks. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | All three subtasks—typing, proofreading, and sorting mail—are routine clerical operations that current AI systems (OCR, spell-check, document classification, RPA) can perform end-to-end with substantial time savings and equal or better quality than human operators. |
| Task automatability | claude-sonnet-5 | 4/5 | Typing, proofreading, and mail sorting are highly structured digital or semi-digital tasks that current AI (OCR, NLP proofreading tools, document classification/routing systems) can perform with substantial time savings, though physical mail handling requires some human/robotic support. |
| Adoption barriers | claude-haiku-4-5-20251001 | 1/5 | No licensing, regulatory, or legal requirement mandates human performance of clerical typing, proofreading, or mail sorting; organizational adoption is primarily driven by cost, with minimal friction beyond legacy process inertia. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, legal, or liability barriers restrict automating typing, proofreading, or mail sorting; these are routine administrative functions. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Infrastructure costs for automation (cloud OCR, RPA, NLP APIs) are orders of magnitude cheaper per task than the loaded wage of a human operator performing these repetitive clerical duties. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-based proofreading and typing tools cost a small fraction of a human clerical wage per unit of output, though physical mail sorting still requires human or mechanized handling that tempers full cost savings. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 5/5 | Mature, production-deployed systems already handle typing (speech-to-text), proofreading (grammar and spell-checking tools), and mail/document sorting (machine learning classifiers in enterprise systems) reliably at scale across many organizations. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Products like Grammarly, AI transcription tools, and automated mail-sorting systems are deployed in production, but full end-to-end clerical workflows still typically involve human oversight and mixed digital/physical tasks. |
Operate paging systems or other systems of bells or buzzers to notify recipients of incoming calls.
81CI 61–100 · exposure 72 · augmentation 13 · importance 4.5/5 · click for rater detail
Operate paging systems or other systems of bells or buzzers to notify recipients of incoming calls.
81| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Most call notification has already been automated across sectors over the past two decades; telephone operator roles have largely disappeared. The slow remaining velocity reflects only legacy systems and niche use cases still in operation, not new adoption. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | This function has already been almost fully automated across nearly all sectors; legacy manual telephone operator roles for this task have essentially disappeared. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | Call notification is a rote, already-automated task with little room for human-AI collaboration. Modern systems have no need for a human operator in the loop, making augmentation opportunities minimal. |
| Augmentation potential | claude-sonnet-5 | 2/5 | Since the task itself is nearly always fully automated already, there is little remaining human-in-the-loop process for AI to augment. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Modern phone systems and voicemail handle most incoming call notification automatically without human intervention. However, legacy paging systems and niche contexts still require manual operator input, and the task lacks the complexity to achieve 50% time savings over existing automation already deployed in most organizations. |
| Task automatability | claude-sonnet-5 | 5/5 | Operating paging or notification systems for incoming calls is a simple, rule-based routing function that automated telephony/PBX and call-routing software has handled for decades, fully meeting the time-saving threshold. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory or legal barriers prevent automation of paging and call notification; most organizations have already replaced manual operators with system automation. Customer preference for human operators is minimal for this function, and organizational inertia is the main remaining friction. |
| Adoption barriers | claude-sonnet-5 | 1/5 | There are no licensing, legal, or liability barriers preventing automated paging/notification systems from replacing manual operation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Fully automated call notification systems cost pennies per thousand calls and run continuously without labor. A human telephone operator wage is orders of magnitude higher, making AI/automation vastly cheaper for any residual notification tasks. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated PBX/paging systems cost a small fraction of a human operator's wage per call handled, running continuously with minimal marginal cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Contemporary PBX systems, VoIP platforms, and automated call routing are production-standard and reliably perform call notification at scale. AI-enhanced systems can intelligently route and notify; however, the task as stated (manually operating buzzers/bells) is largely obsolete in modern deployments, limiting feasibility of the traditional task itself. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated call routing, auto-attendants, and paging integrations are mature, widely deployed products used at scale across virtually all industries today. |
Monitor automated systems for placing collect calls and intervene for a callers needing assistance.
77CI 74–81 · exposure 75 · augmentation 50 · importance 4.2/5 · click for rater detail
Monitor automated systems for placing collect calls and intervene for a callers needing assistance.
77| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 5/5 | Telecommunications and call centers have been rapidly and deeply automating operator functions for over two decades, with widespread production deployment of AI-driven IVR, monitoring, and routing systems across most major carriers and service providers. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Telephone operator roles are a shrinking, largely legacy occupation with automation already deeply embedded for decades, but the occupation itself is in structural decline rather than seeing fast new AI adoption. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI assists remaining human operators by filtering routine calls, flagging priority cases, and providing real-time information about system status; however, the primary use case is replacement rather than productivity enhancement of human operators still performing the task. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted triage and monitoring dashboards can help remaining human operators focus attention on calls that truly need intervention, improving efficiency for the residual human-handled cases. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | Most of the task—monitoring systems and detecting when human intervention is needed—can be automated with current AI. Automated call routing and triage systems can handle the majority of cases, though complex edge cases or difficult customer interactions may still benefit from human oversight, preventing a full 5 rating. |
| Task automatability | claude-sonnet-5 | 4/5 | Modern IVR and speech AI systems already handle most collect-call routing autonomously, with human intervention needed only for edge cases, meeting the time-saving threshold for the bulk of the task.a |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Few regulatory barriers exist to automating call monitoring and triage itself; however, customer preference for human assistance in certain situations and the need for at least some human oversight of edge cases create modest friction to full displacement. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists for this task, though some liability/accessibility regulations (e.g., ADA, emergency services) require a human fallback path for callers needing assistance. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | Modern automated monitoring and routing systems cost significantly less per call handled than employing human telephone operators; the infrastructure is commodity-grade cloud service with minimal ongoing labor, making it orders of magnitude cheaper than human wages. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated call routing systems cost fractions of a cent per call versus a staffed operator's wage, making AI drastically cheaper for the bulk of routine call handling. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed IVR (Interactive Voice Response) and call-routing systems already monitor automated call placement and escalate to humans when needed. These are production systems at scale in telecommunications, though some integration complexity remains for full end-to-end automation across all edge cases. |
| Technical feasibility today | claude-sonnet-5 | 4/5 | Automated collect-call and IVR systems are mature, widely deployed telecom products (e.g., prison/inmate calling systems) that operate at scale today with human backup only for exceptions. |
Promote company products, services, and savings plans when appropriate.
57CI 45–70 · exposure 45 · augmentation 75 · importance 3.6/5 · click for rater detail
Promote company products, services, and savings plans when appropriate.
57| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Contact centers and telecoms are early-to-mid adoption, with IVR and chatbot promotions increasingly common, but high-quality end-to-end replacement in production remains limited; pilots outnumber widespread deployments. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Telecom and customer service sectors have rapidly adopted AI-driven voice bots and chat-based upsell tools, with production deployments increasingly common. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can draft or suggest targeted promotions, highlight relevant products from inventory, and flag optimal timing based on call context, meaningfully boosting operator productivity while the human operator retains judgment on appropriateness and customer rapport. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI can supply real-time prompts, next-best-offer suggestions, and script generation to human operators, meaningfully boosting their promotional effectiveness. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Promoting products and services requires nuanced judgment about customer needs and timing, which current AI struggles with reliably. While chatbots can deliver scripted promotions, they cannot consistently discern 'when appropriate' or adapt to customer context well enough to meet the ≥50% time-saving bar. |
| Task automatability | claude-sonnet-5 | 4/5 | Scripted upselling and promotional messaging during calls can largely be generated and delivered via conversational AI/IVR systems, meeting the time-saving bar for the promotional-speech component of the task. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Some regulatory oversight (FTC rules on promotions, do-not-call lists) exists but does not legally require human sign-off; customer preference for human operators and organizational hesitation to fully automate customer-facing sales provide moderate friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for this task, but customer preference for human interaction and brand risk from mishandled sales pitches create moderate friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Inference cost for a promotional message is minimal (cents), and inference+integration overhead remains far below the loaded wage of a human operator (~$25–40/hour all-in), making AI substantially cheaper per promotion delivered. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated voice/chat promotion systems cost a fraction of a per-minute human operator wage once built, though integration and monitoring add some ongoing cost. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Deployed systems can execute templated promotional scripts and basic product information delivery, but reliably gauging appropriateness and customer sentiment remains error-prone in production; few organizations rely on AI alone for this without human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Conversational AI and voice bots are deployed in call centers for upselling and promotions, but conversion quality and natural handling of objections still lag skilled human operators, so scope is narrower than full replacement. |
Operate telephone switchboards and systems to advance and complete connections, including those for local, long distance, pay telephone, mobile, person-to-person, and emergency calls.
50CI 0–100 · exposure 50 · augmentation 0 · importance 4.5/5 · click for rater detail
Operate telephone switchboards and systems to advance and complete connections, including those for local, long distance, pay telephone, mobile, person-to-person, and emergency calls.
50| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Adoption velocity is near-zero because the displacement happened in the pre-AI era through direct automation of switchboard technology. Modern sectors do not hire telephone operators to manually route calls. |
| Sector adoption velocity | claude-sonnet-5 | 5/5 | Telecommunications is a deeply digitized sector where automated switching has already achieved near-total penetration and displacement of manual operators. |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI cannot meaningfully augment a task that is not performed by humans. The underlying function (call routing) is handled entirely by telecom infrastructure with no human-in-the-loop role to enhance. |
| Augmentation potential | claude-sonnet-5 | 1/5 | Since the task itself has been almost entirely automated away, there is little remaining human task to augment; few human operators still perform manual switching. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Modern telephone systems are fully automated; the task of manually operating switchboards and advancing connections is largely obsolete. Current AI has no meaningful role in a function that technology has already eliminated in most jurisdictions. |
| Task automatability | claude-sonnet-5 | 5/5 | Modern telecom infrastructure already fully automates call routing and connection via digital switching systems, VoIP, and automated exchanges, exceeding the 50% time-saving threshold trivially. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Historically, the barriers were very high—switchboard operation required licensing and certification. Although the occupation has declined due to technical displacement, regulatory licensing once protected the role from competition. |
| Adoption barriers | claude-sonnet-5 | 1/5 | No licensing, liability, or human-contact requirements protect this task; it has already been automated across virtually all telecom systems. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | This comparison is moot; automated switching systems have a marginal cost per call far below any human wage, but the task itself is not performed by humans or AI in modern practice. |
| Cost vs. human wage | claude-sonnet-5 | 5/5 | Automated digital switching costs fractions of a cent per call versus a human operator's wage per connection, an order-of-magnitude or greater cost advantage. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product performs this task because the occupation itself has been automated away by underlying telecommunications infrastructure over decades. The task no longer exists as a human occupation in production systems. |
| Technical feasibility today | claude-sonnet-5 | 5/5 | Automated switching systems have been in mature, reliable, large-scale production for decades, essentially eliminating the need for manual switchboard operation in normal circumstances. |
Provide assistance for customers with special billing requests.
47CI 32–61 · exposure 38 · augmentation 63 · importance 3.9/5 · click for rater detail
Provide assistance for customers with special billing requests.
47| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 3/5 | Telecommunications and utilities have deployed AI-based billing support systems widely, but adoption is mixed: basic automation is common in information-heavy sectors, yet special billing requests remain frequently handled by humans due to complexity and risk. Pilots are common but full replacement is rare. |
| Sector adoption velocity | claude-sonnet-5 | 4/5 | Telecom, utilities, and financial services sectors have rapidly adopted AI chatbots and virtual agents for billing support, though operator-specific roles are shrinking rather than being fully replaced overnight. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist operators by retrieving account history, suggesting relevant policies, and drafting responses, meaningfully speeding up the handling of special requests while keeping the human in the loop for judgment and final decision-making on exceptions. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI tools can pull account history, suggest resolutions, and draft responses, significantly speeding up how a human operator resolves special billing cases even when full automation isn't appropriate. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can handle routine billing inquiries and provide standard policy information, special requests typically require understanding context, negotiation, and exceptions that are difficult to automate end-to-end. Current AI systems struggle with non-standard cases that demand judgment and flexibility, limiting meaningful automation to well-defined sub-steps rather than the full task. |
| Task automatability | claude-sonnet-5 | 3/5 | Routine billing inquiries and standard adjustments can be handled by AI-driven IVR/chat systems, but 'special' requests often involve exceptions, judgment, and negotiation that current systems handle poorly without human escalation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Regulatory requirements around billing practices, consumer protection, and financial accuracy create moderate friction. Additionally, customer expectations often favor human contact for special requests, and organizations may face liability if AI errors affect customer accounts without proper human sign-off. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists for this task, though some billing disputes may involve regulatory consumer protection rules requiring documented human review or appeal processes. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI infrastructure for billing assistance is relatively inexpensive to run, but the need for human review, escalation, and override of AI decisions for special requests keeps total cost per resolved case near or above a loaded operator wage when integration and oversight are factored in. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | AI-driven support systems are far cheaper per interaction than staffed telephone operators for the routine portion of billing requests, though special cases still require costlier human review. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Chatbots and IVR systems can handle basic billing questions in production, but special requests often require human escalation, judgment calls, and account-specific decision-making that current AI cannot reliably handle without significant human oversight. Deployed systems typically fail on edge cases and complex negotiation. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Deployed conversational AI and billing chatbots exist widely at telecom and utility companies, but they typically handle straightforward cases and escalate special or complex billing disputes to humans. |
Provide relay service for users who are deaf or hard of hearing.
26CI 4–49 · exposure 25 · augmentation 50 · importance 3.8/5 · click for rater detail
Provide relay service for users who are deaf or hard of hearing.
26| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Relay service remains a tightly regulated, non-profit or government-contracted sector with minimal pressure and low investment incentive to automate. Adoption of AI-assisted tools is nascent, and meaningful replacement is not occurring at scale. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Automated captioning technology has been adopted in telecom and consumer accessibility products, but adoption within formally regulated relay services has been slower due to compliance and reliability requirements. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could potentially assist human relay operators by auto-suggesting clarifications or providing real-time transcription aids, but the core task's requirement for human judgment, emotional intelligence, and legal accountability means augmentation benefits remain limited. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-powered live captioning and voice recognition significantly boost operator efficiency and accuracy, allowing human operators to manage more calls or focus on quality assurance rather than manual transcription. |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | Relay service requires real-time bidirectional communication mediation between deaf/hard-of-hearing users and hearing parties, with precise message fidelity, emotional tone capture, and immediate context-switching. Current AI cannot reliably perform this conversational intermediation at the required quality and speed without human judgment. |
| Task automatability | claude-sonnet-5 | 3/5 | Speech-to-text and text-to-speech AI can handle much of the transcription and voicing, but real-time relay requires handling accents, interruptions, and emotional nuance that still benefit from human judgment, so full automation with equal quality isn't fully proven. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Federal regulations (ADA, FCC rules on relay services) require that relay services meet specific accessibility and accuracy standards, and many jurisdictions have explicit human-operator requirements or mandates for certified relay service providers, creating hard legal barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 4/5 | Relay services are federally mandated (FCC in the US) with strict accuracy and confidentiality requirements, often requiring certified operators for certain call types, creating substantial regulatory barriers to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI infrastructure (speech recognition, text-to-speech, voice synthesis) has become cheaper, but integration, quality assurance, and real-time mediation oversight still require significant human involvement, keeping total delivered-service cost comparable to or slightly below traditional relay operators. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated speech recognition and synthesis are far cheaper per call than employing human relay operators, though some human oversight or fallback is still needed to ensure accuracy and compliance. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed product reliably performs full relay service substitution for deaf/hard-of-hearing users. While speech-to-text and text-to-speech exist separately, integrating them into a live relay service that maintains message integrity, handles interruptions, and preserves nuance remains research-stage only. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | AI-based captioning and voice synthesis products (e.g., automated captioned telephone services) exist and are deployed, but many relay services still rely on human operators for accuracy and regulatory compliance, especially for complex or emergency calls. |
Interrupt busy lines if an emergency warrants.
21CI 0–43 · exposure 20 · augmentation 13 · importance 3.9/5 · click for rater detail
Interrupt busy lines if an emergency warrants.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Telephone operator roles are declining sharply; modern telecommunications rely on automated switching, emergency services direct access, and legal frameworks that do not contemplate AI decision-making for call interruption. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | placeholder |
| Augmentation potential | claude-haiku-4-5-20251001 | 1/5 | AI offers no meaningful assistance here because the task fundamentally requires human authority, judgment, and legal accountability for an emergency determination that affects third parties. |
| Augmentation potential | claude-sonnet-5 | 2/5 | placeholder |
| Task automatability | claude-haiku-4-5-20251001 | 1/5 | This task requires real-time judgment about what constitutes an 'emergency warrant' sufficient to interrupt a call, which demands contextual reasoning, legal knowledge, and ethical accountability that current AI systems cannot reliably execute end-to-end. Additionally, the technical infrastructure and authentication required to interrupt calls has been largely obsoleted by modern telephony. |
| Task automatability | claude-sonnet-5 | 3/5 | Detecting emergency signals and interrupting a line is a rule-based process that automated telecom systems can execute, but it requires reliable emergency-classification and integration with switching infrastructure that isn't a simple drop-in replacement for human judgment in ambiguous cases.telephone.rrationale. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Hard legal and regulatory barriers exist: only authorized human operators or emergency services personnel can legally interrupt calls, and liability for wrongful interruption would create severe error-cost asymmetry favoring human accountability and authorization. |
| Adoption barriers | claude-sonnet-5 | 3/5 | placeholder |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | This task is virtually obsolete in modern telecommunications due to automatic call routing and emergency services infrastructure; the comparison is moot, but any AI system capable of safely performing this would require extensive oversight and integration costs exceeding the minimal human cost of the rare instances when it occurs. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | placeholder |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs emergency call interruption with the necessary legal authority, liability protection, and contextual judgment. This task involves regulatory authority and human accountability that cannot be delegated to current AI systems. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | placeholder |
Offer special assistance to persons such as those who are unable to dial or who are in emergency situations.
18CI 18–18 · exposure 25 · augmentation 50 · importance 4.0/5 · click for rater detail
Offer special assistance to persons such as those who are unable to dial or who are in emergency situations.
18| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | Emergency services are among the most conservative sectors; adoption of AI in life-safety contexts remains minimal due to regulatory capture, union protection, and the catastrophic cost of failure, limiting real-world displacement. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Telephone operator roles are a shrinking, low-digitization occupation with minimal AI agent deployment for emergency-specific assistance. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI tools can assist operators by suggesting rapid response protocols, auto-detecting keywords for urgency levels, and providing information lookup—materially useful but human judgment and compassionate response remain essential. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI can help triage, transcribe, or provide language translation support during calls, but human operators remain essential for judgment and escalation in emergencies. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can recognize certain emergency keywords and route calls, the task requires genuine human empathy, real-time judgment of vulnerability, and ability to provide reassurance to distressed callers—capabilities current AI systems lack reliably. Only narrow slices (keyword detection, call routing) can be automated without quality loss. |
| Task automatability | claude-sonnet-5 | 2/5 | This requires real-time emergency judgment, physical dialing assistance, and empathetic human interaction that current AI cannot reliably substitute for end-to-end without significant risk. |
| Adoption barriers | claude-haiku-4-5-20251001 | 5/5 | Strong regulatory barriers exist: emergency services are heavily licensed, human operators are legally mandated in many jurisdictions, and liability for failures (missed emergencies, misdirected critical calls) creates hard constraints against full automation. |
| Adoption barriers | claude-sonnet-5 | 5/5 | Emergency assistance carries high liability, regulatory oversight (e.g., 911/emergency service standards), and requires human judgment and accountability, making substitution legally and ethically constrained. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | The infrastructure to deploy AI-driven emergency assistance (24/7 monitoring, failover to humans, liability coverage, training) remains more expensive than existing human operator models; error costs in emergency contexts are asymmetrically high. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | While AI voice systems are cheap per interaction, the liability and oversight costs for emergency handling substantially offset savings compared to a human operator. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably handles this end-to-end; emergency dispatch systems exist but still require human operators to assess situations and provide appropriate support. Early-stage chatbots exist but fail on edge cases and emotional nuance that define this task. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Voice assistants can route calls or provide basic conversational support, but no deployed product reliably handles emergency assistance calls for vulnerable users at scale. |
Related occupations — Office & Administrative Support
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.