1. How to Upload Files as Context in the AI Content Builder
The AI Content Builder is embedded directly into the editing experience so you can work without switching tools.
- Open the Spekit Web App
- Click + New
- Select Create a Spek or Create a Topic Page.
- Click on the AI Content Builder button
- Click the Plus icon to open the source library
- Under the Speks & Files tab, type at least 2 characters to find your source materials. You can select multiple records to provide a comprehensive knowledge base.
- Click Accept to "prime" the AI with your selected data.
- Issue your request. The AI will now prioritize your attached records over general knowledge to draft your content.
Prompt Examples:
- "Using the attached Product Roadmap and Pricing Sheet, create a summary of how our upcoming features justify the Q3 price increase."
- "Based on the attached High-Performer Call Transcript, draft a 'Discovery Question' list that mimics the top rep's style."
- "Compare this draft to the attached Brand Voice Guide and rewrite the intro to match our official tone."
- "Review the three attached Competitor Battlecards and create a single 'Quick Comparison' table for a prospect."
- "Using the attached Security Whitepaper, update the 'Data Privacy' section of this Spek to ensure 100% technical accuracy."
2. How to Add Transcripts (Including Gong Calls) as Context in the AI Content Builder
You can use transcripts to transform real conversations into structured, actionable content. This includes both uploaded transcript files and Gong call data (if connected).
- Open the Spekit Web App
- Click + New
- Select Create a Spek or Create a Topic Page.
- Click on the AI Content Builder button
- Click the paperclip icon to open the source library
- Navigate to the Meetings tab and search for a synced recording (type at least 2 characters). OR click Attach Transcript to manually upload a
.txtfile from tools like Zoom, Otter, or Chorus. - Select multiple transcripts to synthesize themes across an entire discovery process or deal cycle.
- Click Accept to ground the AI in that specific conversation.
- Issue your request to generate content based solely on what was actually said.
Prompt Examples:
- "Based on the transcript, extract the top 3 pain points mentioned by the prospect and draft a 'Discovery Recap' Spek for this account."
- "Search the transcript for budget or timeline mentions. Draft a 'Deal Risk' summary highlighting any potential red flags."
- "Create a 'Next Steps' Spek that lists every commitment made by both the rep and the prospect during this conversation."
- "The prospect mentioned [Competitor Name]. Based on their comments, draft a 'Competitive Landmine' section focusing on why our UI/UX won them over."
- "Find the moment the rep was asked about pricing. Rewrite their response into a polished 'Pricing Objection Handler' for the rest of the team."
- "Identify any questions the rep couldn't answer. Create a 'Knowledge Gap' Spek that provides those answers for the next call."
- "Find the section where the rep explained our AI differentiation. Turn that talk track into a 'Best Practice' Spek titled 'Pitching Spekit AI vs. General LLMs'."
- "Analyze the discovery questions asked in this transcript. Create a 'Stage 1 Playbook' list of the 5 most effective questions that got the customer talking."
- "The customer reacted very positively to the ROI story at the 20-minute mark. Turn that story into a 'Customer Success Snippet' that other reps can use."
- "The Sales Engineer gave a detailed technical explanation of our Slack Integration. Simplify that transcript into a 5-step 'How-to' guide for a non-technical audience."
- "Extract any product feature requests or 'wish list' items mentioned by the customer and format them into a 'Product Feedback' Spek."
- "Identify the specific security standards discussed. Create a 'Security FAQ' Spek so reps don't have to wait for an SE to answer these in the future."
- "Draft a 'Next Steps' email for the prospect that references the specific budget concerns they mentioned at the 15-minute mark."