How to Research an Existing Client Before an Upsell Call With NotebookLM
Turn recent client website updates, announcements, and your own project notes into a source-grounded upsell call brief with NotebookLM.
The problem and who this is for
The worst upsell call is the one where you show up with a generic idea that ignores what changed inside the client business. A better call starts with a short research brief built from the client's recent public moves and your existing project context.
This is for consultants and freelancers who already have an active client relationship and want a sharper, more relevant phase-two conversation.
Prerequisites
- A NotebookLM account.
- Two to five public client sources such as website pages, press releases, blog posts, job listings, or product updates.
- Your own project notes, scope summary, or latest status document.
How to capture or gather the source material
- Choose public sources that reflect current direction, not background noise. Good options are the homepage, services page, pricing page, recent blog posts, product changelog, careers page, or press release page.
- Copy the URLs directly into NotebookLM if they are public pages. If the information sits behind a login or dynamic interface, capture it as a PDF or screenshot outside the tool first.
- Add one short internal context file that explains what you already do for this client. Without that, the AI may suggest ideas you already delivered.
- Keep the source set tight. Five good sources beat fifteen mediocre ones. The goal is to detect change, not dump the internet into a notebook.
Step-by-step workflow
- Create a notebook and add the public client URLs plus your internal context note.
- Ask for a briefing document that answers three questions: what appears to have changed recently, what business priorities those changes suggest, and which of those priorities connect to your current work.
- Review the brief and cut anything speculative. Stay with moves the client actually made, pages they actually updated, or roles they are actively hiring for.
- Ask NotebookLM to convert the brief into three expansion hypotheses. Each hypothesis should tie one business signal to one practical next service you could offer.
- Ask for a final call-prep page with talking points, proof points, and one discovery question per hypothesis. This keeps the call consultative instead of pitch-heavy.
- Before the meeting, sanity check the brief yourself. Visit the client site again and make sure the key pages or announcements are still live and current.
Tool-specific instructions
NotebookLM
- NotebookLM is a strong primary tool here because it can combine public URLs with your own files and keep the resulting brief grounded in those sources.
- Use a briefing document first. Then ask for hypotheses and call talking points in separate passes.
- Keep the notebook focused on recent pages and recent changes.
Gemini
- Gemini is a good fallback if you want to upload screenshots or copied text and then export the resulting brief to Docs.
- It is especially useful when the client site has visual changes that are easier to capture as screenshots.
ChatGPT
- ChatGPT is a practical fallback when the research packet is small and you mainly need help converting notes into a call brief.
- Upload screenshots, PDFs, or notes instead of relying on memory.
Claude
- Claude is useful if you want to maintain a persistent Project for a client account and keep updating the research brief as new sources appear.
Copy/paste prompt blocks
NotebookLM client research prompt
{
"role": "account_research_analyst",
"goal": "Create an upsell call brief from the uploaded public client sources and internal project context.",
"instructions": [
"Use only the uploaded or linked sources.",
"Identify recent changes, likely priorities, and signals that matter to the existing engagement.",
"Do not suggest services the internal context file shows are already being delivered.",
"Create 3 expansion hypotheses with one discovery question each."
],
"output_format": {
"sections": [
"recent_business_signals",
"likely_priorities",
"expansion_hypotheses",
"call_talking_points"
]
}
}
Fallback prompt
{
"role": "client_account_planner",
"goal": "Turn the uploaded client research materials into a concise pre-call brief.",
"instructions": [
"Stay grounded in the source material.",
"Favor specific next-step ideas over broad strategic advice."
]
}
Quality checks
- Every major claim ties back to a page, announcement, or internal note.
- The brief does not recommend work you already do for the client.
- Each hypothesis sounds like a natural next step, not a forced add-on.
- The final talking points are short enough to use in a real call.
Common failure modes and fixes
The brief sounds generic
Use fewer but higher-signal sources and add your current scope note.
The brief makes speculative claims
Ask the model to separate observed changes from inferred implications.
The ideas duplicate current work
Rewrite the internal context note so it clearly defines what is already included.
The research packet is too large
Cut older pages and low-signal content.
Sources Checked
- Google NotebookLM Help - Add or discover new sources for your notebook: https://support.google.com/notebooklm/answer/16215270 (accessed 2026-03-24)
- Google NotebookLM Help - Create a notebook in NotebookLM: https://support.google.com/notebooklm/answer/16206563 (accessed 2026-03-24)
- Google Gemini Apps Help - Upload & analyze files in Gemini Apps: https://support.google.com/gemini/answer/14903178 (accessed 2026-03-24)
- Google Gemini Apps Help - Export responses from Gemini Apps: https://support.google.com/gemini/answer/14184041 (accessed 2026-03-24)
- OpenAI Help Center - File Uploads FAQ: https://help.openai.com/en/articles/8555545-file-uploads-faq (accessed 2026-03-24)
- OpenAI Help Center - ChatGPT Image Inputs FAQ: https://help.openai.com/en/articles/8400551-chatgpt-image-inputs-faq (accessed 2026-03-24)
- Claude Help Center - What are projects?: https://support.claude.com/en/articles/9517075-what-are-projects (accessed 2026-03-24)
Quarterly Refresh Flag
Review this article by 2026-06-22. Re-check tool capabilities, source limits, mobile support, export behavior, and any changes to file upload or output features before republishing unchanged.
Related Workflows
How to Turn Client Emails and Revision Notes Into an Upsell Opportunity Map With NotebookLM
Turn client emails, revision notes, and meeting recaps into a source-grounded map of upsell opportunities that actually match what the client needs.
How to Turn a Client KPI Export Into a Quarterly Business Review Deck With NotebookLM
Turn a client KPI export, deliverables list, and win log into a source-grounded quarterly business review deck you can present or send.
How to Turn a Client Debrief Recording Into a Renewal-Risk Brief With NotebookLM
Turn a client debrief recording into a source-grounded renewal-risk brief that shows concerns, wins, blockers, and the next conversation to have.