Skip to main content

This is a prototype vision of how a future government service could work. It's not a real service yet, but we're exploring what it could look like. Your feedback will help shape the real service.

Step 4: Compare Voices & Calculate ROI - Text-to-Speech Walkthrough

Listen to voice quality differences and calculate cost savings for your council

Walkthrough progress

Step 4 of 4 • 2 minutes

Step 4 2 minutes

Compare Voices & Calculate ROI

Understand which voice works best for different announcements and calculate the business value for your council.

Find your Text to Speech application URL in CloudFormation Outputs
The application is ready to convert text into natural speech

Expected outcome

  • Identified which voice suits your chosen announcement
  • Understood when to use each voice (Amy/Brian/Emma)
  • Calculated cost savings for your council's announcement volume
  • Recognized accessibility compliance value (WCAG, PSBAR)
  • Can articulate text-to-speech ROI to decision-makers

Voice style comparison

Based on the three audio previews you generated in Step 3, notice how each voice creates a different impression:

Amy - Conversational

Tone characteristics:

  • Warm and friendly
  • Approachable, non-threatening
  • Conversational inflection
  • Reassuring for routine information

Best use cases:

  • Service updates (bins, libraries)
  • Community events
  • General announcements
  • Public health information

Resident perception: "Sounds like a helpful council officer explaining something in person."

Brian - Authoritative

Tone characteristics:

  • Clear and commanding
  • Authoritative, serious
  • Firm, deliberate pace
  • Conveys urgency and importance

Best use cases:

  • Weather warnings
  • Emergency alerts
  • Road closures (urgent)
  • Safety announcements

Resident perception: "Sounds official and important - I need to pay attention."

Emma - Formal

Tone characteristics:

  • Professional and precise
  • Measured, formal delivery
  • Clear articulation
  • Suitable for legal/statutory content

Best use cases:

  • Committee announcements
  • Planning notices
  • Formal consultations
  • Legal statements

Resident perception: "Sounds like an official council document being read professionally."

Calculate your savings

Based on your council's announcement volume, estimate the annual cost savings from text-to-speech:

Include service updates, emergency alerts, committee notices, public health information
Voice talent, studio time, editing (typical range: £150-£300)
Typical council announcements: 300-600 characters

Beyond cost savings: Accessibility compliance

The financial ROI is compelling, but the accessibility compliance value is even more significant:

WCAG 2.1 AA Requirements

Web Content Accessibility Guidelines (WCAG) 2.1 Level AA requires that information and user interface components be "perceivable" to all users. For text-only content, this means providing an audio alternative.

  • Criterion 1.1.1 (Level A): Non-text content must have text alternative
  • Criterion 1.2.1 (Level A): Audio-only content needs text transcript
  • Inverse application: Text-only content should have audio version for visually impaired

PSBAR 2018 (Public Sector Accessibility Regulations)

UK law requires public sector bodies to make their websites and mobile apps accessible. This came into force in September 2020, with enforcement by the Equality and Human Rights Commission.

  • Requirement: Councils must publish accessibility statements
  • Standard: Meet WCAG 2.1 AA by law
  • Enforcement: Equality and Human Rights Commission can issue fines
  • Penalty range: Up to £20,000 per breach
  • Text-to-speech benefit: Demonstrates "reasonable adjustments" for disabled residents

Equality Act 2010

Broader legal duty to make reasonable adjustments for disabled people accessing council services.

  • Section 29: Service providers must not discriminate against disabled users
  • Reasonable adjustments: Provide information in accessible formats
  • Protected characteristic: Visual impairment, reading difficulties, cognitive disabilities
  • Enforcement: Residents can file complaints with EHRC or take legal action
  • Reputational risk: Non-compliance damages public trust in council

Bottom line: Text-to-speech is not just a nice-to-have feature - it's a legal requirement under PSBAR 2018 and a "reasonable adjustment" under the Equality Act 2010. The £0.40 annual cost for 50 announcements is negligible compared to potential £20K penalties per breach and reputational damage from non-compliance.

Committee language generator

Use this language when presenting text-to-speech to decision-makers:

For Finance Committee:

"Amazon Polly text-to-speech reduces audio production costs by [calculated]% - from £[current] to £[polly] annually for [count] announcements. This frees up £[savings] that can be reallocated to frontline services. Additionally, achieving WCAG 2.1 AA compliance avoids potential penalties up to £20,000 per accessibility breach under PSBAR 2018. Total annual value: £[total] in cost savings and compliance protection."

For Accessibility/Equality Committee:

"Text-to-speech makes council announcements accessible to 2% of UK residents with visual impairments (1.3M people), 16% with literacy below Level 1, and 19% aged 65+ who may have age-related vision decline. This meets our legal duties under PSBAR 2018 and Equality Act 2010, demonstrates 'reasonable adjustments' for disabled residents, and provides multi-channel information delivery (web, phone, social media). Accessibility is a right, not a privilege."

For IT/Digital Committee:

"Amazon Polly is a managed service requiring zero infrastructure. It integrates with existing council CMS (WordPress, Drupal, Joomla) via simple API calls. Supports 29 languages for multilingual communications. Neural voices generate broadcast-quality audio in 30 seconds, indistinguishable from professional recording studios. Pay-per-use pricing (£0.016 per 1000 characters) means no upfront investment. This is a low-risk, high-value technology adoption that enhances our digital service delivery."

Reflection: Beyond announcements

Consider broader applications:

  • Phone systems (IVR): Generate audio for automated phone menu updates instead of studio recording
  • Council podcasts: Convert written newsletters to audio podcasts for commuters
  • Meeting minutes: Provide audio versions of committee meeting minutes for visually impaired residents
  • Consultations: Make planning applications and consultation documents accessible via audio
  • Emergency alerts: Integrate with SMS and phone systems for urgent weather/safety announcements
  • Multilingual content: Use Polly's 29 language support for diverse communities (Polish, Punjabi, Urdu, etc.)

If your council publishes 50 announcements but also 12 monthly newsletters, 20 committee meeting summaries, and 10 consultation documents per year, text-to-speech scales to 92 audio files annually for £0.74 total cost vs £18,400 with studio recording. The ROI multiplies with volume.

Troubleshooting

Can't decide which voice to use

Voice selection tips:

  • Default to Amy for most announcements (80% use case coverage)
  • Use Brian only for urgent/emergency content requiring authority
  • Use Emma only for formal legal/committee content
  • Survey residents for preference (include audio samples in consultation)
  • A/B test different voices on website to see which gets more plays
Need to use text-to-speech for long documents (>3000 characters)

For longer content:

  • Split document into logical sections (max 3000 chars each for API)
  • Use StartSpeechSynthesisTask API for asynchronous long-form audio (no limit)
  • Consider creating chapter markers for navigation in audio player
  • Provide text transcript alongside audio for reference
  • For very long content (10+ pages), consider summarization before TTS conversion

You've now experienced the full text-to-speech workflow and understand both the technical capability and business value. Let's wrap up with key takeaways.