Handlet Intelligence & Benchmarking Policy
This policy explains how Handlet uses conversation, operational, usage, and outcome data to provide personalised intelligence, improve the service, and create privacy-protected benchmark and insight products.
It should be read alongside our Privacy Policy, Terms of Service, Data Processing Addendum, and Subprocessors.
1. Purpose
Handlet is designed to help service businesses understand and improve how they win, serve, retain, and follow up with customers.
To do this, Handlet analyses patterns such as:
- response speed;
- message intent;
- tone and style;
- objections and questions;
- quote values and outcomes;
- booking and cancellation events;
- follow-up timing;
- review and complaint handling;
- lead source, channel, and funnel stage;
- service type, broad area, and customer journey events; and
- call-agent transcripts, summaries, recordings, and outcomes where call-agent features are enabled.
This analysis powers account-specific features. Cross-customer benchmark or partner intelligence using customer communication data is disabled unless the customer has given a separate, explicit instruction or authorisation covering the source processing.
2. Data Use Categories
| Category | What it means | External sharing |
|---|---|---|
| Account-bounded intelligence | We use your workspace data to provide your own Handlet features, such as drafts, automation, lead scoring, tone learning, quote insights, call-agent summaries, and recommendations. | Not shared externally except through integrations, subprocessors, or actions you configure. |
| Platform learning and service improvement | We use account-bounded patterns and feedback to improve features for that account. Cross-customer learning from customer communication data requires the customer's separate explicit authorisation. | Not sold or shared as identifiable data. |
| Aggregated benchmark intelligence | Where separately authorised, we may create statistics, benchmarks, trend reports, and performance patterns across cohorts and then apply anonymisation and disclosure controls. | May be shown to users or commercial partners only after release review confirms the output is anonymous and sufficiently aggregated. |
| Message pattern examples | We may create privacy-protected examples that illustrate common objections, winning reply structures, follow-up patterns, or customer journey themes. | External examples must be anonymised, transformed, summarised, paraphrased, or synthetic so they are not designed to identify a person, business, account, or customer. |
| Partner and agency insight products | We may provide reports, dashboards, APIs, or other insight products to partners or agencies using aggregated or anonymised intelligence. | Partners must not receive raw messages, customer lists, call recordings, CRM records, or identifiable account-level behavioural profiles unless a customer has explicitly authorised a separate integration or disclosure. |
3. What Handlet Does Not Sell
Handlet does not sell:
- raw customer messages;
- raw call recordings or transcripts;
- customer contact details;
- CRM records;
- identifiable business profiles;
- identifiable end-customer profiles;
- account-level behavioural profiles; or
- pseudonymised datasets presented as anonymous data.
Where Handlet commercialises intelligence derived from customer communication data, the customer must first have separately and explicitly authorised the source processing. Handlet must then ensure that any externally released output is irreversibly anonymised or otherwise lawfully disclosed under that authorisation. Merely pseudonymising, paraphrasing, labelling, or removing obvious identifiers is not treated as anonymisation by itself.
4. Aggregated Insights and Cohort Controls
Aggregated insights may be grouped by categories such as:
- trade or industry;
- broad geographic area;
- lead source;
- communication channel;
- funnel stage;
- service type;
- quote band;
- customer intent;
- persuasion pattern;
- follow-up timing; or
- outcome type.
Handlet uses controls designed to reduce re-identification risk. These may include minimum cohort thresholds, suppression of small samples, rounding, generalisation, removal of unusual details, transformation of examples, and internal review before external publication or partner access.
The more specific a segment is, the more care is needed. Handlet will not intentionally publish or share insight outputs that are designed to identify a specific business, customer, account, message, or call.
Before an external benchmark release, Handlet must record the dataset scope, authorisation source, cohort and suppression rules, re-identification assessment, reviewer, approval date, and intended recipients. Special-category or criminal-offence information must not be used in external benchmark products.
5. Message Examples
Some insights are easier to understand with examples. Where examples are based on real communications, Handlet will remove, transform, generalise, summarise, paraphrase, or replace information that could reasonably identify people, businesses, accounts, or customers.
Handlet may also use synthetic examples generated from observed patterns rather than actual messages.
External message examples should not include names, contact details, exact addresses, exact timestamps, unique property details, rare job details, or other information that could reasonably reveal who was involved.
6. Partner and Agency Restrictions
Partners, agencies, and other commercial recipients of Handlet intelligence must not:
- attempt to identify a Handlet customer, end customer, staff member, caller, account, message, or call from insight outputs;
- combine Handlet insight outputs with other data to re-identify people or businesses;
- use Handlet insight outputs to target or profile a specific identifiable business unless that business has explicitly authorised the relevant integration or disclosure;
- resell underlying insight datasets unless expressly permitted by contract; or
- use insight outputs in a way that breaches applicable data protection, marketing, consumer, or sector-specific law.
7. Customer Controls
Some processing is required to provide Handlet features. For example, Handlet cannot provide intent recognition, enabled social post support, call summaries, lead scoring, or account-specific recommendations without analysing the relevant workspace data. Future or optional AI message-drafting features will apply only where those features are enabled.
Benchmark contribution, benchmark receipt, partner and agency insights (including cross-customer platform learning), and real-message example participation require separate, explicit customer authorisation. Anonymising processor data after collection is not used as a substitute for that source-processing authorisation. These activities remain off by default until you enable them in Settings → Intelligence Network for your workspace. You may receive benchmarks without contributing your own data (receive-only mode). A customer may also give or withdraw written authorisation by contacting privacy@handlet.ai. Withdrawal does not affect processing already carried out lawfully, but Handlet will stop future authorised contributions for that customer.
Customers remain responsible for telling their own customers how communications may be handled through Handlet, including any call recording, transcription, AI assistance, and privacy-protected benchmarking notices required for their business.
8. Legal Basis and Data Protection
Handlet's lawful basis depends on the processing activity. Core service delivery may be necessary for contract. Security, reliability, service improvement, and some platform learning may rely on legitimate interests. Some activities may require consent or explicit authorisation.
Where a business customer is the controller of its own customer communication data, Handlet processes that data as processor under the Data Processing Addendum. If Handlet independently determines a separate purpose and means for processing personal data, Handlet acts as controller for that separate processing and must identify its lawful basis, complete any required legitimate-interests or impact assessment, provide appropriate transparency, and respect objections or withdrawals that apply.
Where Handlet acts as a controller for separately-authorised, privacy-protected benchmarking derived from end-customer data, Handlet relies on the Article 14(5)(b) exemption (provision of information would be impossible or involve disproportionate effort), mitigated by: processing only de-identified, aggregated or transformed data; this public policy; and a contact route (privacy@handlet.ai) for end customers. Tradespeople are provided with a notice template to inform their own customers. This benchmarking remains disabled unless and until the customer has given separate explicit authorisation.
9. Updates
Handlet may update this policy as the product, benchmarking features, partner products, and legal requirements evolve. If changes are material, we will provide reasonable notice.
10. Contact
Questions about this policy should be sent to privacy@handlet.ai.