Model ML at Work: Across Surfaces

Work in financial services doesn't happen in one place. A request lands in your inbox, the analysis lives in Excel, the output goes into PowerPoint and the final version goes back out by email, often while you're between meetings or in transit. Most AI tools don't reflect that. They sit in a separate tab or app, which means stopping what you're doing, pasting in context and then moving the output back into the tools you actually work in.

At Model ML, your agent is surface-agnostic. It works where you already work: your inbox, Excel, PowerPoint, Word, your internal tools and your phone.

Use cases: Model ML across surfaces

Example 1: An MD's request becomes a briefing pack in 30 minutes

In the video above, Laxmi, Head of Engagement Management at Model ML and a former banker at Perella Weinberg Partners, is in transit when her MD asks for a briefing pack on the defense tech startup landscape. The meeting is in 30 minutes.

• Phone: She replies to the MD's email, CCs her agent and hits send. The agent starts pulling from precedent databases, S&P, Capital IQ and PitchBook.

• Model ML app: When she's back at her laptop, the results are waiting in the app, with a summary of each company, why it made the cut and its sources, with an Excel backup attached.

• Excel: She opens the backup, already in the firm's template with every number tied to its source. From the Excel plugin, she asks Model ML to turn that tab into a logo page in PowerPoint.

• PowerPoint: The deck opens with the whole coverage universe, including companies and fundraising stats, in the firm's template. Once she's happy with it, she asks Model ML to create a PDF and draft a reply to her MD.

• Outlook: The drafted email is waiting with the PDF attached. She gives it a quick once-over and sends it.

An afternoon of work is now done in 30 minutes with Model ML.

Example 2: A partner’s message becomes a client-ready proposal in an hour

A consultant is between client meetings when a partner messages her on Teams: they need a proposal for a cost-transformation mandate.

• Teams: She tags @ Model ML and asks it to draft the proposal using the engagement’s discovery notes, project materials and Notetaker records. The agent brings together the client context, objectives and team discussions, then develops the proposed workstreams, deliverables and open questions.

• Email: Before she’s back at her laptop, an email arrives with a full summary of the proposal, the Word document and a PowerPoint pitch deck in the firm’s template. Assumptions and questions needing her judgment are clearly flagged, with sources attached to the underlying claims.

• Model ML mobile app: She taps the link in the email and picks up the conversation with the agent. Using voice, she sharpens the scope, adjusts the recommendations and asks it to put more emphasis on the client’s immediate priorities. The agent updates the proposal and deck without her having to work through each document manually.

• PowerPoint: Back at her laptop, she opens the revised deck. The client’s problem, proposed approach and mandate scope are already laid out in the firm’s format. She reviews the story and makes the final refinements.

• Site: The agent suggests turning the proposal into an interactive Site. A couple of clicks later, she has a browsable version of the pitch, allowing the client to explore the workstreams, deliverables and supporting context rather than just read through slides.

• Outlook: A draft email to the partner is already waiting, summarising the proposal and highlighting the decisions still needed. She attaches the Word proposal and PowerPoint, includes the Site link, gives it a final read and hits “Send.”

What would previously have taken her and the team a week has come together in an hour.

Example 3: An inbound CIM becomes a partner briefing on the same morning

A PE associate's deal inbox fills up with CIMs and teasers from bankers every week, and the partner wants to know about any that are worth pursuing, whether as a new investment or as an add-on for one of the portfolio companies.

• Model ML app: She sets up a Schedule in Model ML that runs whenever a CIM or teaser arrives. Each time, it logs the deal in the CRM, summarises it, and screens it two ways: as a standalone investment against the fund's mandate, and as a bolt-on against the portfolio companies' value creation plans.

• Email: A banker sends over a CIM for a regional competitor of one of the portfolio companies. Within minutes, the Schedule updates the CRM with the new deal and an email from Model ML is waiting with a summary of the business, the key risks, and the verdict: too small to be a platform under the fund's mandate, but a strong bolt-on fit. She replies asking Model ML to build a combined case in that company's model.

• Excel: She opens the model with the combined case already worked in. From the Excel plugin, she asks Model ML to compare it with the standalone plan and show the synergies, the leverage impact, and the downside.

• Site: She asks Model ML to update the fund's tracker site with the deal. The Site now shows the pro forma impact on fund allocation, the portfolio company's reserves, and remaining dry powder, next to the rest of the portfolio, so the partner can see the fund-level picture without opening a spreadsheet.

• Outlook: A draft email to the partner is already waiting, covering the opportunity, what it adds to the plan, what it would cost the fund, and what to ask the banker. She attaches the model, includes the tracker link, gives it a final read, and hits "Send."

Instead of waiting days to be screened, the CIM reaches the partner the morning it arrives.