Evidence Register

The public booking flow captures a service category, a free-text description of the task, the service address and a preferred date and time before any provider is presented to the customer.

Function supported:

Customer Task Capture

Technologies:

Task Interpretation, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Primary company source. The capture sequence is directly observable without an account and is reproducible across categories.

Provider onboarding describes selection of service categories, identity verification and background screening, selection of the work area, and configuration of the hours in which the provider will accept work.

Function supported:

Provider Skills & Onboarding

Technologies:

Skills, Availability, Geography

Published:

Continuously updated

Accessed:

18 August 2026

Establishes that capability, geography and availability are all captured as structured provider record attributes at registration.

Help documentation describes the categories a Tasker may register for, the vetting applied before activation, and the category-level task history surfaced on a provider profile.

Function supported:

Provider Capability & Verification

Technologies:

Skills, Performance

Published:

Continuously updated

Accessed:

18 August 2026

Supports capability data being held at category level. Does not establish sub-skill, credential-class or tooling-level capability data.

Public location directory sets out the metropolitan areas in which the platform operates and within which providers are presented to customers.

Function supported:

Geographic Service Area

Technologies:

Geography

Published:

Continuously updated

Accessed:

18 August 2026

Establishes metro-level operation. Sub-metro travel radius behaviour is inferred from provider documentation rather than directly disclosed.

Providers set weekly working hours and may update availability in the provider application; customer-facing time options reflect the hours a provider has published.

Function supported:

Provider Availability

Technologies:

Availability, Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Availability operates as a live constraint on the options presented, not as a display-only profile attribute.

Engagements are hourly-based, with an estimated number of hours presented at booking and invoicing reconciled against actual time worked.

Function supported:

Task Duration

Technologies:

Duration, Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

An estimated-duration value demonstrably exists and drives the booking. The derivation method — category default, provider input or model output — is not disclosed.

Following task, location and timing capture, the customer is presented with an ordered shortlist of providers showing category, hourly rate, rating, review volume and completed-task counts.

Function supported:

Provider Identification & Selection

Technologies:

Matching, Performance

Published:

Continuously updated

Accessed:

18 August 2026

Directly observable. Establishes a task-, place- and time-constrained candidate subset. Ranking logic is not disclosed.

Confirmation produces an appointment communicated to both the customer and the provider, which may subsequently be rescheduled or cancelled through the application.

Function supported:

Booking & Schedule Transmission

Technologies:

Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Establishes construction and transmission of an executable schedule record with amendment handling.

Provider guidance addresses managing one's own calendar and allowing for travel between engagements; no system-generated route or sequence optimisation is described.

Function supported:

Travel & Sequencing

Technologies:

Routing, Optimisation

Published:

Continuously updated

Accessed:

18 August 2026

Absence of public description is not evidence of absence internally. Recorded as an unresolved area requiring further investigation.

Ratings, written reviews, completion counts and reliability expectations are documented, and performance standards are described as affecting a provider's standing on the platform.

Function supported:

Historical Performance

Technologies:

Performance

Published:

Continuously updated

Accessed:

18 August 2026

Establishes a closed performance loop. Whether performance data is fed into duration estimation or schedule construction is not disclosed.

Corporate communications reference investment in AI-supported discovery, assisted booking journeys and platform automation across the service marketplace.

Function supported:

AI Architecture

Technologies:

AI, Corporate Announcement

Published:

Ongoing

Accessed:

18 August 2026

Establishes stated direction of travel toward interpretation-layer AI. Does not establish that AI is applied to duration, ranking or schedule construction.

Parent-group communications describe digital and AI investment across the services businesses, including assembly and home-services fulfilment.

Function supported:

Group Technology Direction

Technologies:

AI, Corporate Announcement

Published:

Ongoing

Accessed:

18 August 2026

Group-level statement. Relevance to the Taskrabbit scheduling stack specifically is an interpretation, not an observed fact.

Retail integration allows a task to be originated from a partner purchase journey and fulfilled through the same capture, matching and scheduling pipeline.

Function supported:

Partner & Retail Integration

Technologies:

API, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Indicates programmatic task origination. The integration interface itself is not publicly documented.

Customers may request same-day or next-available work, with the presented options constrained by published provider availability in the service area.

Function supported:

Same-Day and Time-Slot Presentation

Technologies:

Scheduling, Availability

Published:

Continuously updated

Accessed:

18 August 2026

Demonstrates reconciliation of customer timing preference against provider availability at request time.

Services are organised into a structured taxonomy of categories and sub-tasks, with size or scope options presented in several categories.

Function supported:

Task Taxonomy & Sizing

Technologies:

Task Interpretation

Published:

Continuously updated

Accessed:

18 August 2026

Guided structuring achieves part of the interpretation objective through input design rather than through disclosed inference.

The application describes a data communications network in which customer task information, provider capability, geographic region and availability data are processed to identify suitable providers and to generate and transmit task schedules to those providers.

Function supported:

Patent Record

Technologies:

Patent

Published:

Application record

Accessed:

18 August 2026

Primary patent record. All patent positions in this portal are stated at architecture level and should be read against the claims as filed.

The published international application corresponding to the United States application sets out the specification, drawings and description of the scheduling architecture, including embodiments concerning provider capability, geographic region, availability, assigned task duration and task schedule generation.

Function supported:

Published International Application

Technologies:

Patent

Published:

International publication

Accessed:

18 August 2026

Use this record when a statement rests on specification or embodiment disclosure rather than on a claimed element. The distinction is preserved throughout the portal.

The international application PCT/AU2024/050370 was filed on 17 April 2024 and published on 24 October 2024 as WO2024216337A1. The published record identifies Nicole Milatos as inventor and as original applicant.

Function supported:

International Application Record

Technologies:

Patent

Published:

International filing date 17 April 2024

Accessed:

18 August 2026

Primary patent-family source. Bibliographic data only; it does not determine claim scope in any jurisdiction.

The published international application claims priority from Australian application AU2023901123 with a priority date of 17 April 2023.

Function supported:

Australian Priority Application

Technologies:

Patent

Published:

Priority date 17 April 2023

Accessed:

18 August 2026

Priority record as identified on the published international application. No independent examination record is asserted here.

Partner integration material describes a sequence in which a partner obtains an estimate for a service at an address, retrieves availability, submits a bid for a selected window, books the appointment and then manages the resulting project.

Function supported:

Partner API Booking Workflow

Technologies:

API, Scheduling, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Analyst interpretation: the partner-facing sequence exposes the same capture, availability reconciliation and schedule-transmission stages observable in the consumer flow, in machine-callable form. Precise field-level behaviour is not published openly.

Work is generally billed by the hour against the provider's rate, with the customer shown an estimate before booking; the final amount reflects the time actually worked.

Function supported:

Estimated Price and Hourly Billing

Technologies:

Duration, Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Analyst interpretation: an estimate expressed in currency does not establish a system-determined estimate of the time required to complete the task. The two must be assessed separately.

A confirmed appointment may be rescheduled or cancelled through the application, with rescheduling described as agreed between the customer and the provider and cancellation policies applying within stated windows.

Function supported:

Rescheduling, Cancellation and Project Management

Technologies:

Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Analyst interpretation: post-booking change is presented as a negotiated human process. No automated schedule reconstruction or substitute-provider assignment is described in public material.

Provider-facing material indicates that the order in which Taskers appear may reflect factors including availability, work area, selected categories, hourly rate and business performance metrics.

Function supported:

Search Result Ordering Factors

Technologies:

Matching, Performance

Published:

Continuously updated

Accessed:

18 August 2026

Analyst interpretation: this evidences ranking inputs at a descriptive level. It does not disclose the ranking function, its weighting, or whether any predictive model is applied.

Providers are subject to identity verification and background screening before activation, and are required to hold any licence that applicable law requires for the work performed.

Function supported:

Identity, Background and Licensing

Technologies:

Skills, Performance

Published:

Continuously updated

Accessed:

18 August 2026

Analyst interpretation: verification operates as an activation gate. There is no public evidence of a structured, machine-readable credential record used as a matching constraint at task level.

Ingka Group reports group-level financial results covering IKEA retail operations and its wider portfolio, within which Taskrabbit sits following acquisition.

Function supported:

Ingka Group Ownership and Financial Reporting

Technologies:

Corporate Announcement

Published:

Annual

Accessed:

18 August 2026

Analyst interpretation: figures published here are Ingka Group financial information. They are not Taskrabbit revenue, and no third-party estimate of Taskrabbit revenue is treated as audited fact anywhere in this portal.

Reporting describes Taskrabbit services being surfaced or booked through third-party AI assistant interfaces, distinct from AI capability developed within the Taskrabbit platform itself.

Function supported:

Third-Party AI Interface

Technologies:

AI, Task Interpretation

Published:

Varies

Accessed:

18 August 2026

Tier 3 source. Analyst interpretation: an external interface to a booking capability is not evidence of native AI in the matching or scheduling path, and is recorded separately from Taskrabbit-native AI throughout.

Service availability is determined by the address or postal code supplied by the customer; where a location is outside a served metropolitan area, no provider options are returned.

Function supported:

Address Validation and Service Eligibility

Technologies:

Geography, API

Published:

Continuously updated

Accessed:

18 August 2026

Analyst interpretation: geography operates as a hard eligibility filter applied before any provider is presented, not merely as a display attribute.

Application listings describe browsing local Taskers, comparing rates and reviews, selecting a time and chatting with the Tasker before and after booking.

Function supported:

Application Distribution and Product Description

Technologies:

Corporate Announcement, AI

Published:

Continuously updated

Accessed:

18 August 2026

Tier 2 source. Analyst interpretation: corroborates the customer-selection model — the customer chooses from presented candidates rather than receiving a system-assigned provider.

Ingka Group describes Taskrabbit as an on-demand platform through which people can hire others to perform tasks, acquired in 2017 and operated as a stand-alone company within the group structure.

Function supported:

Corporate Ownership and Acquisition

Technologies:

Corporate Announcement

Published:

2017 (acquisition), reporting continuously updated

Accessed:

18 August 2026

Establishes ownership, acquisition year and stand-alone operating posture. Does not establish any technical integration between IKEA systems and Taskrabbit scheduling infrastructure.

Ingka's investment portfolio material identifies Taskrabbit within its home-furnishing and home-services related holdings.

Function supported:

Portfolio and Home-Services Positioning

Technologies:

Corporate Announcement

Published:

Continuously updated

Accessed:

18 August 2026

Establishes the current portfolio relationship and home-services positioning. Does not establish operational or technical dependency in either direction.

Developer documentation describes an Estimate call that accepts service and location information — postal code or full address — and returns whether the service is available for that address together with estimated pricing.

Function supported:

Partner API — Estimate

Technologies:

API, Geography, Duration

Published:

Continuously updated

Accessed:

18 August 2026

Tier 1 technical evidence. Establishes address-driven service eligibility and price estimation. Does not establish any duration computation, nor the use of AI techniques.

Availability guidance describes retrieval of real-time appointment windows for a service and address combination, with services, location, date-from and date-to identified as query information, returning available date/time windows and estimated prices.

Function supported:

Partner API — Availability

Technologies:

API, Availability, Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Tier 1 technical evidence. Establishes that bookable windows are exposed programmatically. Does not establish what generates the windows, whether Tasker work area or availability is applied, or whether task duration influences window length.

Bid guidance describes creation of an agreement for a selected date/time window following customer selection, returning more accurate pricing and a Project ID used to complete the booking.

Function supported:

Partner API — Bid

Technologies:

API, Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Tier 1 technical evidence. Establishes that selection precedes agreement creation. Does not establish construction of a provider task schedule.

Book guidance describes completing the project using the Project ID and customer/project information, states that the call creates the live project and schedules a Tasker, and describes support for automated checkout experiences and synchronisation between Taskrabbit and partner systems.

Function supported:

Partner API — Book and Booking Guide

Technologies:

API, Scheduling, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Tier 1 technical evidence. Establishes programmatic booking and partner synchronisation. Does not establish how a Tasker is associated with the booked window, nor whether any broader schedule is generated.

Announcement describes a customer telling Claude what needs to be done and where help is required; Claude connects to Taskrabbit, checks whether the service is available in that area and provides a route into Taskrabbit to discover and book the service.

Function supported:

AI-Mediated Service Discovery

Technologies:

AI, Corporate Announcement, Task Interpretation

Published:

23 April 2026

Accessed:

18 August 2026

Establishes an AI-mediated discovery interface between natural-language need, location and Taskrabbit service availability. Does not establish AI-derived task duration, AI provider selection, AI schedule construction or agentic orchestration.

The homeowner interface asks what the customer needs help with and for a ZIP code, then presents relevant projects and local professionals.

Function supported:

Customer Task Capture & Provider Matching

Technologies:

Task Interpretation, Geography, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Establishes task requirement capture and location capture as request inputs. Does not establish how the eligible professional population is computed internally.

Angi describes an AI Helper that uses advanced large language models and targeted job-assessment questions to translate a homeowner's own description of a problem into a clearer, more detailed service request for professionals.

Function supported:

AI Task Interpretation

Technologies:

AI, Task Interpretation

Published:

2026

Accessed:

18 August 2026

Strong evidence of semantic task interpretation and AI-assisted project scoping. Does not establish AI assignment of a recommended amount of time required to complete the job.

Announcement describes an end-to-end AI-guided hiring journey from an initial homeowner question, through project scoping, to matching with a skilled professional.

Function supported:

AI-Guided Hiring Journey

Technologies:

AI, Matching, Corporate Announcement

Published:

March 2026

Accessed:

18 August 2026

Establishes AI-guided scoping and matching. Does not establish AI-derived duration or AI-generated provider schedules.

Angi tells service professionals that they can receive leads matching their trade, their schedule and their region, and allows professionals to present skills, ratings and reviews.

Function supported:

Trade, Schedule & Region Matching

Technologies:

Skills, Availability, Geography, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Establishes capability, availability and geography as matching variables. Does not expose how schedule is represented internally, or whether it means calendar availability, preferred job timing or lead preference.

Angi publishes extensive project cost information and AI-guided cost expectations to help homeowners understand what a project might cost.

Function supported:

Cost Estimation

Technologies:

Duration, Task Interpretation

Published:

Continuously updated

Accessed:

18 August 2026

Establishes cost intelligence. Does not establish duration intelligence: price is not a computed task length.

Handy presents pre-priced services, captures the customer's location, and allows selection of a date and a time before payment, with a professional supplied for the booked service.

Function supported:

Handy Direct Booking

Technologies:

Scheduling, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Strong scheduling relevance and evidence of platform-side professional assignment in some workflows. Does not establish how the offered windows are generated or whether an expected duration sizes them.

Help material describes appointment availability such as 7 AM to 11 PM subject to local professional availability, and allows customers to manage, reschedule and cancel bookings online.

Function supported:

Appointment Availability & Rescheduling

Technologies:

Scheduling, Availability

Published:

Continuously updated

Accessed:

18 August 2026

Establishes digital schedule management and professional-dependent availability. Does not establish autonomous schedule reconstruction or automatic replacement-provider identification.

Handy describes connecting customers with pre-screened independent service professionals and has described instantly matching large numbers of customers with highly rated professionals.

Function supported:

Professional Screening & Assignment

Technologies:

Matching, Performance

Published:

Continuously updated

Accessed:

18 August 2026

Indicates platform-side assignment for some service types. Does not establish the inputs or timing of the assignment decision.

Angi maintains extensive ratings, reviews and professional reputation data presented to homeowners.

Function supported:

Professional Performance Data

Technologies:

Performance

Published:

Continuously updated

Accessed:

18 August 2026

Establishes historical performance data capture. Does not establish that performance data influences scheduling rather than ranking or customer choice.

Shareholder material describes a hard shift from incremental AI adoption to AI-first product development.

Function supported:

AI-First Product Direction

Technologies:

AI, Corporate Announcement

Published:

Q1 2026

Accessed:

18 August 2026

Establishes declared strategic direction. Does not establish any currently deployed duration or scheduling capability.

Angi has described a planned AI front desk for professionals able to make calls, receive calls and book appointments.

Function supported:

AI Appointment Booking

Technologies:

AI, Scheduling, Corporate Announcement

Published:

2026

Accessed:

18 August 2026

Highly relevant to agentic booking trajectory. Announced rather than independently observed; does not establish schedule construction or optimisation.

Thumbtack describes a nationwide network of approximately 300,000 local service businesses serving homeowners across the United States.

Function supported:

Company Technology Model

Technologies:

Corporate Announcement, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Establishes scale and national marketplace operation. Does not establish backend matching mechanics.

The public journey asks the customer to answer a few questions describing what they need, after which the platform identifies relevant local professionals.

Function supported:

Customer Task Capture

Technologies:

Task Interpretation, Geography

Published:

Continuously updated

Accessed:

18 August 2026

Establishes structured task and location capture preceding provider presentation. Does not establish the internal filtering methodology.

Thumbtack has publicly described an Instant Book e-commerce engine connecting homeowners with service professionals for direct hiring.

Function supported:

Instant Book

Technologies:

Scheduling, Matching

Published:

Continuously updated

Accessed:

18 August 2026

Indicates a direct booking pathway. Does not establish slot generation, calendar sourcing, service duration or platform assignment mechanics.

Describes receiving a service request including a location and request type, determining a candidate group of providers matching location and category, computing variables for candidate providers including likelihood that the customer will select the provider and provider capacity (including the number of requests a provider can fulfil in a given period), selecting a smaller group of providers, and generating and transmitting responses. Provider statistical information such as selection frequency, fulfilment, response rates, reviews and ratings is described as stored and used.

Function supported:

Multi-Stage Provider Matching & Capacity

Technologies:

Patent, Matching, Availability, Performance

Published:

13 December 2022

Accessed:

18 August 2026

Thumbtack Patent-Disclosed Architecture. Directly documents multi-stage subset generation, capacity and performance statistics. Disclosure in a patent does not establish current commercial implementation, and it does not disclose AI-derived task duration or provider schedule construction.

Patent documentation describes provider selection that may use machine-learning models and heuristics, with factors including request relevance, provider interest, customer selection likelihood, supply and demand, provider qualification, provider fit and business value.

Function supported:

Machine-Learning Provider Selection

Technologies:

Patent, AI, Matching

Published:

2017–2022

Accessed:

18 August 2026

Thumbtack Patent-Disclosed Architecture, recorded separately from Current Product Evidence. Does not establish which models operate in production today.

Thumbtack describes a new AI experience allowing homeowners to describe problems in natural language, by photo or by voice, and matching them to a curated list of local professionals tailored to the project, explaining why each professional is a strong fit based on expertise, project needs, customer ratings and preferred availability.

Function supported:

AI Curated Provider Recommendation

Technologies:

AI, Task Interpretation, Availability, Performance, Matching

Published:

2026

Accessed:

18 August 2026

Current Product Evidence. Establishes AI task interpretation and multi-variable curated subset generation including preferred availability. Does not establish AI-derived task duration or provider schedule generation.

Thumbtack has announced integrations and collaborations involving OpenAI Operator, ChatGPT and Claude, exposing local-professional recommendation and hiring pathways to AI systems.

Function supported:

Agentic AI Access

Technologies:

AI, API, Corporate Announcement

Published:

2025–2026

Accessed:

18 August 2026

Establishes a strategy of exposing marketplace infrastructure to AI agents. Does not establish autonomous end-to-end scheduling by an agent.

Public material describes AI assistants reaching Thumbtack services to surface local professional recommendations in response to home questions.

Function supported:

AI Partner Interfaces

Technologies:

AI, API

Published:

2026

Accessed:

18 August 2026

Interface detail is not fully public. Does not establish the data exchanged, nor whether booking can be completed by the agent.

The Home Profile records the home address, home features, systems and goals, and Thumbtack provides AI-driven suggestions covering what to do, when to do it and who to hire.

Function supported:

Home Profile & Timing Intelligence

Technologies:

AI, Task Interpretation, Scheduling

Published:

Continuously updated

Accessed:

18 August 2026

Establishes timing intelligence at the task level. Does not establish generation of a service provider's task schedule.

Thumbtack surfaces seasonal and weather-informed maintenance recommendations to homeowners.

Function supported:

Seasonal & Weather-Based Recommendations

Technologies:

AI, Optimisation

Published:

Continuously updated

Accessed:

18 August 2026

Relevant to predictive task initiation. Does not establish weather-aware provider scheduling or routing.