.avif)
BOSTON, MA, September 24, 2026 - DealRoom today announced an efficiency optimization to DealRoom MCP that reduces the amount of deal data AI needs to process, lowering token consumption and cost without sacrificing capability.
As M&A teams expand their use of AI, costs can rise quickly, especially when the AI has to retrieve large datasets and filter them itself. This waste compounds as AI usage scales, drawing scrutiny from finance leaders.
DealRoom MCP connects the AI deal teams already use to live deal data across their pipeline, data rooms, requests, and findings. Teams can use their AI tools of choice to read and write information in DealRoom in real time, leveraging existing permissions and a complete audit trail.
Now, DealRoom is making that connection more efficient by filtering results before they are sent to the AI.
How it works: before and after
The way teams use their AI does not change. They can ask the same question, in the same way they do today.
For example, say a user asks their AI: “What is the value of all Q3 deals?”
Before this update, the AI retrieved each DealRoom record and filtered the results afterward. That meant tokens were consumed on all of the data returned, including records that did not match the request.
Now, the AI sends the criteria to DealRoom first. DealRoom directly filters the data and returns only the matching records, reducing what the AI needs to process.
The result for the user is the same: identical output, fewer tokens consumed, and lower AI spend.
In testing, server-side filtering reduced the amount of deal data returned by roughly 90%. Each request receives only the relevant records, delivering the same output with significantly fewer tokens and lower AI costs.
Lower AI cost, same results
Instead of sending all available data to the AI tool to process, DealRoom MCP now directly filters the response and returns only matching records.
This shifts the filtering work from the AI to DealRoom, reducing token usage and cost without sacrificing capability. It is similar to improving fuel efficiency without impacting performance.
Processing costs now apply only to the records the request actually needs, rather than every record in the pipeline.
Filtering out irrelevant records also keeps the AI’s context focused, reducing noise and producing more reliable, consistent answers.
Why it matters as AI usage scales
Lower cost removes the need to ration AI usage. Teams can use it across their deals without concern for budget limits, while finance sees lower overall spend that is easier to justify and approve.
As adoption grows, the savings compound, allowing M&A teams to capture more value from DealRoom and AI without costs becoming the constraint.
For teams, the change is intentionally simple: ask the same question, get the same output, while the AI has far less data to process.
Available to DealRoom MCP Beta users
Filtered retrieval is available for deals today, with findings, tasks, and other additional objects being added over time.
The optimization is available to all DealRoom MCP Beta users. DealRoom MCP is currently in private beta and is included at no additional charge for DealRoom customers.
Not a DealRoom customer yet? Request a demo to see how DealRoom MCP can bring AI directly into your M&A workflow.
.png){kind=small}
.avif)
.avif)
.avif)
.avif)
.avif)
