Version: 2.0
Last Updated: November 30, 2025
Target Audience: Customer Support Team
Table of Contents
- Overview of New Features
- Campaign-Centric Workflow
- AI Configuration Management
- A/B Testing Explained
- Analytics Dashboard
- Troubleshooting Common Issues
- Customer FAQ
Overview of New Features
What Changed?
The Campaign Architecture v2.0 release introduces a fundamental shift in how TendSocial organizes content creation:
Before (Blog-Centric):
- Everything started with a blog post
- Social posts were derived from blogs
- Limited flexibility for non-blog businesses
After (Campaign-Centric):
- Campaigns are the central organizing concept
- Blogs, social posts, and videos all belong to campaigns
- Businesses without blogs can use TendSocial effectively
Key Benefits for Customers
- Better Organization - Group related content by campaign/initiative
- Improved AI Quality - AI learns from user editing patterns
- Transparent Costs - Detailed AI usage and cost tracking
- Smarter AI - Multiple AI models optimized for different tasks
- Performance Insights - Compare AI-generated vs human-created content
Campaign-Centric Workflow
What is a Campaign?
A Campaign is a coordinated content initiative with a specific goal. Examples:
- "Q1 Cybersecurity Awareness Push"
- "Spring Sale Promotion"
- "Thought Leadership Series"
- "Product Launch - New Widget"
Campaign Types
- One-Time - Single campaign with defined start/end (product launch, event)
- Recurring - Repeating campaign (weekly tips, monthly newsletter)
- Evergreen - Always-on content themes (general tips, FAQ series)
Creating Content with Campaigns
Old Workflow:
- Write blog post
- Generate social posts from blog
New Workflow:
- (Optional) Create a campaign
- Generate any content type (blog, social, video)
- Associate content with campaign
- AI uses campaign context for better results
Important: Campaigns are optional. Users can still create content without a campaign.
AI Configuration Management
What Customers Should Know
- Different AI models for different tasks (blog writing, social posts, images)
- Automatic selection - Users don't need to choose
- Cost-optimized - We use the best model for each task
- Admin-controlled - We can change models without code deployment
Current AI Models (as of Nov 2025)
| Task | Model | Cost per Request |
|---|---|---|
| Campaign Planning | Claude Opus | ~$0.50 |
| Blog Writing | Claude Opus | ~$1.00 |
| Social Posts | Claude Sonnet | ~$0.10 |
| Video Scripts | Gemini 1.5 Pro | ~$0.20 |
| Image Generation | DALL-E 3 | ~$0.04 |
Note: These costs are approximate and vary based on content length.
What Changed?
- Before: Hard-coded model selection
- After: Database-driven configuration with A/B testing
Customer Impact: Better AI quality and lower costs through optimization.
A/B Testing Explained
What is A/B Testing?
We test different AI models to find the best balance of cost vs quality.
Example: For blog writing, we're testing:
- Variant A (Sonnet): Cheaper, faster
- Variant B (Opus): More expensive, potentially higher quality
Users are randomly assigned to one variant and consistently use that model for the duration of the test.
How to Explain to Customers
"We're always testing to give you the best AI results at the best price. You might see slight differences in AI output quality as we optimize our models. This is normal and helps us improve over time."
If a Customer Asks About Test Participation
"All users are automatically included in our optimization tests. You'll always get a high-quality AI model - we're just testing which specific model works best. You can opt out by contacting us, but we recommend staying in the test to help improve the platform."
Analytics Dashboard
New Analytics Features
Customers can now see:
AI vs Human Performance
- Engagement rates for AI-generated vs manually created posts
- Edit ratios (how much they change AI content)
- Regeneration rates (how often they request rewrites)
AI Usage Tracking
- Number of AI requests per day/week/month
- Cost per post/blog/image
- Most expensive content types
Campaign Performance
- Posts per campaign
- Total impressions/engagement by campaign
- Top performing campaigns
How to Access
Customer View:
- Dashboard → Analytics Tab
- Shows their company's data only
Admin View:
- Admin Panel → AI Analytics
- Shows all companies, can drill down
Common Questions
Q: "Why do I see AI costs?"
A: We're transparent about AI usage. This helps you understand the value you're getting and track usage across your team.
Q: "Can I reduce my AI costs?"
A: AI costs are included in your subscription. We show them for transparency, but there's no additional charge. If you notice unusually high usage, we can help optimize your workflow.
Troubleshooting Common Issues
Issue: "My posts don't have campaign context"
Cause: Post was created before campaign feature was deployed
Solution: This is normal for older posts. New posts will automatically support campaigns.
Issue: "AI quality seems inconsistent"
Possible Causes:
- User is in an A/B test (different models)
- Input context varies (brand profile not complete, no examples provided)
- Different tasks use different models
Solution:
- Check if brand profile is complete
- Ensure campaign brief is detailed
- Provide example posts for AI to learn from
Issue: "Content generation is slow"
Possible Causes:
- First generation of the day (cold start)
- Long content (blogs are slower than social posts)
- High platform load
Solution:
- Normal for first request: 5-10 seconds
- Blogs: 10-30 seconds is normal
- If consistently > 30 seconds, escalate to engineering
Issue: "AI costs showing as very high"
Investigation Steps:
- Check which content types (blogs are more expensive)
- Check regeneration count (many rewrites increase cost)
- Check if multiple team members generating content
Normal Ranges:
- Social post: $0.05 - $0.15
- Blog post: $0.50 - $2.00
- Video script: $0.10 - $0.30
- Image: $0.04 - $0.08
If significantly higher, escalate to engineering.
Customer FAQ
Q: "What's new in this release?"
A: We've made campaigns the center of content planning, improved AI quality through smarter learning, and added transparency around AI costs and performance. The biggest change you'll notice is the optional campaign selector when creating content.
Q: "Do I have to use campaigns?"
A: No! Campaigns are completely optional. They help organize related content and give AI better context, but you can continue creating posts the same way you always have.
Q: "Will my old content still work?"
A: Yes! All your existing posts, blogs, and videos work exactly as before. New features are additive - nothing was removed.
Q: "Why do I see different AI results than my colleague?"
A: You might be in different groups for our A/B testing program. This is temporary (30 days) and helps us optimize AI quality for everyone.
Q: "Can I see which AI model generated my content?"
A: Yes! In the post/blog editor, there's a small info icon that shows generation details including the model used, tokens consumed, and cost.
Q: "What if I don't like what AI generated?"
A: You have several options:
- Edit the content directly
- Click "Regenerate" for a new version
- Provide more context in the campaign brief
- Add example posts for AI to learn from your style
Q: "How do I improve AI quality for my brand?"
Best Practices:
- Complete your brand profile with voice, tone, and keywords
- Create detailed campaign briefs with key messages
- Mark good posts as examples for AI to learn from
- Provide feedback by rating AI-generated content
- Edit consistently - AI learns your editing patterns over time
Q: "Is there an extra charge for AI usage?"
A: No! AI generation is included in your subscription. We show AI costs for transparency, but you're not billed separately.
When to Escalate to Engineering
Escalate if:
- ❌ AI generation completely fails (not just bad quality)
- ❌ Campaign CRUD operations error out
- ❌ Analytics dashboard not loading
- ❌ AI costs > 10x normal range
- ❌ All AI requests timing out
- ❌ Database errors visible to users
Do NOT escalate:
- ✅ User doesn't like AI output (this is subjective - guide them on improvement)
- ✅ AI costs seem high but within normal range (educate on what's normal)
- ✅ User prefers old workflow (campaigns are optional)
- ✅ A/B test variance (this is expected)
Training Exercises
Exercise 1: Create a Campaign
- Log into demo account
- Navigate to Campaigns
- Create a new one-time campaign: "Holiday Sale 2025"
- Add brief: "Promote our 25% off holiday sale"
- Add key messages
- Save campaign
Exercise 2: Generate Content with Campaign Context
- Navigate to Promote (social posts)
- Select the campaign you created
- Generate a post
- Note how AI uses campaign context
- Edit the post and save
- Check analytics to see edit metrics
Exercise 3: Review Analytics
- Navigate to Analytics dashboard
- Find AI vs Human performance comparison
- Identify your top-performing campaign
- Review cost breakdown by task
Quick Reference Card
Campaign Status Values:
- 📝 Draft - Being planned
- ✅ Active - Currently running
- ⏸️ Paused - Temporarily stopped
- ✓ Completed - Finished successfully
- 📦 Archived - Historical reference
AI Model Tasks:
- Campaign Planning
- Social Posts
- Blog Writing
- Video Scripts
- Image Generation
- Profile Analysis (background job)
Normal Response Times:
- Social post: 3-8 seconds
- Blog post: 15-30 seconds
- Video script: 10-20 seconds
- Image: 5-15 seconds
Escalation Email: engineering@tendsocial.com
Slack Channel: #customer-support
Documentation: docs/monitoring-setup.md
Training Complete!
Next Step: Schedule 1-on-1 sessions for Q&A and practice.