Exploring NotebookLM for Enhanced Deal Data Insights

·

Hi folks. I have recently joined a start-up where we have lots of sources of deal data (Jiminny, HubSpot, Notion, GDrive, NotebookLM, ...). We are thinking

encouraging more use of NotebookLM for all info we gather
finding or building a tool that provides a natural language interface to query across all our data sources. (Eg I can ask "who is the most likely Champion for Deal X and why" and get a good answer.)

Couple of questions...

Are others using NotebookLM Like this?
Does the idea in general make sense?
Have folks built their own query tools for this?
Or would you recommend on off-the-shelf tool?

THANK YOU!!

👀1

8 comments

· Sorted by Oldest

Pat H.
·
·
Which of these systems if you single source of truth?
Ben M.
·
·
That's exactly the issue, none is. Data all over the place. I'd like to allow teams to keep using the tools they like and which they have configured for their purposes, and then provide a search interface that means I don't care where the data is, we can still find the answer.
Pat H.
·
·
While I appreciate the democratized approach, I would discourage you from it. Chose one, ideally HubSpot, and pull all data into that platform. Then, you can use the HubSpot Deep Research connector in ChatGPT OR you could use their own AskBreeze tool to gain the same level of insights. Spreading your data across too many systems turns into an untenable ocean.
Ben M.
·
·
Yep, we can push that way, and will if we have to. But with today's tools, I was hoping we could be more flexible.
Alex
·
·
Mind that Notebookllm has a limit on 20 sources and any dynamic stuff it won't update automatically
Alex
·
·
I agree that single source of truth (even if just a dump) and then post-processing is a good way to go 🙂
Isaac L.
·
·
Ben M. this is right up my alley -- as in I solve this for folks. Started in post acquisition integration for PE firms and have expanded more broadly. Centralized data + natural language exploration is in my wheelhouse.
Ben M.
·
·
In case anyone is interested, we have largely solved this as I suggested above. We have made NotebookLM the single source of truth. We are using Zappier, Tray and other tools to get data from the call recording tool, the CRM, Slack channels, Google Docs, Notion pages, ... It can be updated every time you open it, and then we can query the deal, anything from specific questions about specific meetings, to high-level strategy analysis.

Pat H.
·
·
Which of these systems if you single source of truth?
Ben M.
·
·
That's exactly the issue, none is. Data all over the place. I'd like to allow teams to keep using the tools they like and which they have configured for their purposes, and then provide a search interface that means I don't care where the data is, we can still find the answer.
Pat H.
·
·
While I appreciate the democratized approach, I would discourage you from it. Chose one, ideally HubSpot, and pull all data into that platform. Then, you can use the HubSpot Deep Research connector in ChatGPT OR you could use their own AskBreeze tool to gain the same level of insights. Spreading your data across too many systems turns into an untenable ocean.
Ben M.
·
·
Yep, we can push that way, and will if we have to. But with today's tools, I was hoping we could be more flexible.
Alex
·
·
Mind that Notebookllm has a limit on 20 sources and any dynamic stuff it won't update automatically
Alex
·
·
I agree that single source of truth (even if just a dump) and then post-processing is a good way to go 🙂
Isaac L.
·
·
Ben M. this is right up my alley -- as in I solve this for folks. Started in post acquisition integration for PE firms and have expanded more broadly. Centralized data + natural language exploration is in my wheelhouse.
Ben M.
·
·
In case anyone is interested, we have largely solved this as I suggested above. We have made NotebookLM the single source of truth. We are using Zappier, Tray and other tools to get data from the call recording tool, the CRM, Slack channels, Google Docs, Notion pages, ... It can be updated every time you open it, and then we can query the deal, anything from specific questions about specific meetings, to high-level strategy analysis.