Optimize with AI

Uniform's AI capabilities help you optimize conversions by making it faster to create content variations for testing and personalization. Use Scout and AI quick edits to generate alternative copy, set up experiments, and create personalized experiences for different audience segments.

Building experiences with AI

For general AI workflows including creating compositions, localization, SEO, and social media optimization, see Building experiences with AI.

AI-powered conversion optimization in Uniform helps you:

  • A/B testing - Generate and test content variations to find what drives better engagement
  • Personalization - Create tailored experiences for different audience segments
  • Rapid iteration - Quickly generate multiple variations to test different approaches

note

Before starting optimization work, ensure your team has AI credits allocated and consider setting up AI guidance for more consistent results.

A/B tests in Uniform are always configured at the component level. You create test variants of a component, then customize the parameters within each variant.

Use Scout to set up component-level A/B tests. Scout can create the test structure and generate complete content for all parameters across all variants:

  1. Ask Scout to create an A/B test on a specific component: "Set up an A/B test for the hero component with 3 variants"
  2. Scout creates the component test variants with appropriate traffic splits
  3. Scout can generate different content for all parameters in each variant in a single conversation
  4. Review the test configuration
  5. After collecting sufficient data, ask Scout to declare a winner

Example prompts:

  • "Set up an A/B test for the hero component with 3 variants with equal traffic split. Make variant A benefit-focused, variant B urgency-driven, and variant C social proof-focused"
  • "Create an A/B test on the CTA component with 2 variants - one with a free trial offer and one with a demo request"
  • "Add an A/B test to the pricing component with 3 variants - 50% control, 25% variant B with annual billing emphasis, 25% variant C with monthly pricing emphasis"

Use AI quick edits when you need to tweak specific parameters within a variant without changing the entire variation:

For a headline parameter in variant B:

  • "Rewrite this headline to emphasize urgency"
  • "Make this headline more action-oriented"
  • "Focus this headline on the primary benefit"

For a CTA button parameter in variant C:

  • "Rewrite this CTA to reduce friction"
  • "Make this CTA more benefit-focused"
  • "Use a stronger action verb"

For a description parameter:

  • "Rewrite emphasizing ROI and business value"
  • "Make this more feature-focused"
  • "Emphasize the outcome rather than the features"

Test one component at a time: Focus on a single component to clearly identify what drives improvement. Within that component, you can test variations of multiple parameters (headline, CTA text, image, etc.), but testing one component at a time makes results clearer.

Create diverse component variations: Use Scout to create complete variations testing different psychological approaches. Scout can generate all the content for each variant in a single conversation:

  • Benefit-focused vs. feature-focused
  • Urgency-driven vs. trust-building
  • Social proof vs. expert authority
  • Problem-focused vs. solution-focused

Example: "Set up test with 3 variants - variant A benefit-focused messaging across all fields, variant B urgency-driven with time-sensitive language, variant C social proof with customer success stories."

Use AI guidance for consistency: Set up AI guidance on component parameters to ensure all variations maintain brand voice while testing different approaches.

Personalization in Uniform is always configured at the component level. You create personalized variants of a component for different audience segments, then customize the parameters within each variant.

Scout can update descriptions for your audiences, intents, and signals. These descriptions act as guidance when Scout generates personalized content for each segment.

Example prompts:

  • "Update the 'Enterprise Prospects' audience description to 'B2B visitors from companies with 500+ employees seeking enterprise features and compliance'"
  • "Update the 'pricing-tier' signal description to 'Indicates visitor interest in basic, pro, or enterprise pricing plans'"

When you create personalization targeting these dimensions, Scout uses their descriptions to generate contextually appropriate content.

Use Scout to set up component-level personalization for different audience segments. Scout can create the personalization structure and generate complete content tailored to each audience across all parameters:

Example workflow:

  1. Ask Scout: "Add personalization to the hero component targeting the 'Returning Customers' audience and 'First-Time Visitors' audience. Show upgrade messaging for returning customers and getting started content for new visitors."
  2. Scout creates audience-based component variants with targeting rules
  3. Scout generates different content for all parameters in each variant based on the audience
  4. Review and adjust targeting rules
  5. Optionally use AI quick edits to tweak individual parameters

Example prompts:

  • "Add personalization to the hero component for the 'Returning Customers' audience showing advanced features and for the 'New Visitors' audience emphasizing ease of use and quick wins"
  • "Set up personalization on the CTA component targeting the 'Enterprise' audience with 'Request Enterprise Demo' and 'Small Business' audience with 'Start Free Trial'"
  • "Personalize the testimonials component based on the 'Industry' enrichment - show healthcare customer quotes for healthcare visitors and retail quotes for retail visitors"
  • "Add personalization to the pricing component using the 'customer-tier' signal - show annual plans for premium-tier and monthly plans for basic-tier"

Use AI quick edits when you need to adjust specific parameters without regenerating the entire personalized variant:

For the 'First-Time Visitors' audience variant - headline parameter:

  • "Rewrite this to address someone unfamiliar with our product"
  • "Emphasize ease of getting started and quick wins"
  • "Add trust signals and social proof"

For the 'Returning Customers' audience variant - headline parameter:

  • "Rewrite this for existing customers who know our product"
  • "Emphasize advanced features and new capabilities"
  • "Focus on upgrade benefits and exclusive features"

For the 'Enterprise' audience variant - description parameter:

  • "Make this more premium and exclusive"
  • "Emphasize ROI and business impact"
  • "Focus on enterprise-grade capabilities and compliance"

Start with high-impact segments: Focus personalization on segments with clear behavioral or demographic differences.

Keep variations manageable: Start with 2-3 key audience segments before expanding to more granular personalization.

Test personalized vs. control: Use A/B testing to validate that personalized experiences actually improve conversions compared to a single generic experience.

Tools: Scout for full variation editing + optionally AI quick edits for parameter tweaks

  1. Ask Scout to help you identify the component to test (hero, CTA, testimonials, pricing, etc.)
  2. Ask Scout: "Set up an A/B test for the [component name] with [number] variants. Make variant A [approach], variant B [approach], variant C [approach]"
  3. Scout creates component test variants with traffic splits and generates complete content for all parameters in each variant
  4. Review the variations Scout created
  5. Optionally, use AI quick edits to fine-tune specific parameters in any variant
  6. Monitor test performance
  7. After statistical significance, ask Scout to declare a winner
  8. Apply winning component variation to similar pages

Example: "Set up an A/B test for the hero component with 3 variants - 50% control (current), 25% variant B emphasizing cost savings and ROI, 25% variant C emphasizing speed and ease of implementation. Update headlines, descriptions, and CTA text for each approach."

Time investment: 5-10 minutes to set up with Scout, then monitor until statistical significance.

Tools: Scout for full variation editing + optionally AI quick edits for parameter tweaks

  1. If needed, ask Scout to update descriptions for your audiences, intents, or signals (see Using descriptions for better personalization)
  2. Define 2-3 key audiences, signals, or enrichments to target
  3. Identify which component to personalize (hero, CTA, product recommendations, etc.)
  4. Ask Scout: "Add personalization to the [component name] targeting [audience/signal/enrichment name] with [specific content approach for each]"
  5. Scout creates audience-based component variants with targeting rules and generates complete content for all parameters tailored to each audience (using the descriptions as guidance)
  6. Review targeting rules and the content Scout created for each audience
  7. Optionally, use AI quick edits to refine specific parameters in any variant
  8. Monitor performance by segment

Examples:

  • "Add personalization to the hero component for the 'New Visitors' audience emphasizing getting started quickly and the 'Returning Customers' audience highlighting new features and upgrade benefits"
  • "Personalize the CTA component using the 'user-role' signal - show 'Manage Team' for admin role and 'View Dashboard' for user role"
  • "Add personalization to the testimonials component based on the 'Industry' enrichment - show healthcare testimonials for healthcare industry and financial services testimonials for finance industry"

Tools: Scout for complete variation editing

  1. Identify underperforming component based on analytics
  2. Ask Scout: "Set up an A/B test for the [component name] with 2 variants. Keep control as-is. Make variant B test [specific different approach with details]"
  3. Scout creates component test variants and generates complete content for the new variant
  4. Review and adjust if needed (use AI quick edits for minor parameter tweaks)
  5. Monitor performance
  6. Implement the winner and repeat for other components

Example: "The pricing component has low conversions. Set up an A/B test with 2 variants - keep control with monthly pricing emphasis, create variant B emphasizing annual savings and ROI with updated headline, description, and pricing display."

Best for: Quick wins on existing pages with known conversion issues.

Before creating personalization, ask Scout to update descriptions for your audiences, intents, and signals. Scout uses these descriptions as guidance to generate more contextually appropriate personalized content.

Example: An audience description like "B2B visitors seeking enterprise features and compliance" helps Scout generate enterprise-focused messaging for that audience's personalized variant.

Set up AI guidance before running optimization campaigns. This ensures all AI-generated variations maintain your brand voice while testing different approaches.

Example AI guidance for A/B test variations:

"Maintain the core message while varying the approach. Test options include: benefit-focused, urgency-driven, social proof, or feature-highlight. Keep tone consistent with brand voice."

  • Start with high-impact components: Hero components, CTA components, and pricing components typically have the biggest conversion impact
  • Test one component at a time: Isolate what's driving improvement by focusing on a single component per test
  • Let Scout create complete variations: Use Scout to generate entire variations with different approaches. For example: "Create 3 variants testing benefit-focused, urgency-driven, and social proof messaging across all parameters." Use AI quick edits only for minor parameter tweaks.
  • Document winners: Build a library of successful component variations to scale across similar pages
  • Generate variations quickly with AI
  • Let real user data determine the winners
  • Use Uniform Insights or your analytics platform to track conversion rates
  • Apply insights from winning tests to inform future AI-generated variations

Always review AI-generated variations before launching tests:

  • Accuracy: Ensure all variations are factually correct
  • Brand consistency: Verify variations maintain your brand voice
  • Clarity: Check that variations are clear and compelling
  • Technical correctness: Ensure no broken links or formatting issues

tip

AI accelerates the variation creation process, but conversion optimization still requires human judgment and data validation. Use AI to generate ideas quickly, then let data tell you what actually works with your audience.

  • Building experiences with AI - Learn about creating compositions, localization, SEO, and social media optimization
  • Scout - Detailed guide on Scout's capabilities for A/B testing and personalization
  • AI quick edits - Learn how to refine content within test and personalization variants
  • AI guidance - Configure guidance to maintain brand voice in all variations
  • AI credits - Manage your team's AI usage
Last modified