From Assessment to Scale: ChatGPT Activation at a $10B PE Firm

Client:

$10B AUM private equity firm

Time to First Value:

6 Weeks (flagship use case live)

Users Enabled:

150-person firm — deal teams, operations, and executive leadership

Our Problem

Problem

A $10 billion PE firm had deployed ChatGPT Enterprise but activation had stalled. Leadership — COO, Head of AI, Head of Due Diligence, CFO, and deal partners — were being pulled in different directions, evaluating competing vertical AI tools with no shared framework for deciding where to start. A previously failed CIM analysis project with another vendor had eroded trust. And usage was shallow: people were rewriting emails, not transforming deal workflows.

The firm didn't need another tool demo. They needed stakeholder alignment, a high-ROI starting point, and a path to firm-wide scale.

Our Solution

Solution

We ran our full Assess → Build & Activate → Scale framework end-to-end.

Phase 1: Assess (Weeks 1–3)

Structured discovery sessions with senior leadership evaluated use cases on ROI (time saved × frequency × people impacted), complexity, and risk — producing a prioritized roadmap separating quick wins from strategic bets. The assessment also surfaced unexpected wins: a DISC Profile GPT built during evaluation gave managers tactical conversation guidance and became an early proof point with the CEO.

Flagship selection: The CIM write-up process — roughly 1,000 CIMs per year, each consuming hours of analyst time — emerged as the highest-ROI workflow. We chose it for the momentum it would create.

Phase 2: Build & Activate (Weeks 3–8)

We built a CIM Write-Up GPT that transforms 100+ page CIMs into IC-ready draft write-ups — financial summaries, thesis framing, risk flags, comparable context — in under two minutes, matched to the firm's tone and formatting standards.

Quality was validated through a rigorous eval loop: a scoring framework for factual accuracy and tone, human evals on real CIMs with deal leads, LLM-based evals for automated regression catching, and in-flight usage monitoring. The result: ~95% factual accuracy, with review time dropping from four hours to one hour per CIM.

In parallel, enabling Outlook and SharePoint connectors drove a 2,156% increase in connector usage — signaling the shift from occasional tool to daily infrastructure.

Phase 3: Scale (Ongoing)

With the flagship proven, the firm moved into expansion: Board Deck Summary GPT, NDA Review GPT, Financial Summary GPT, DealCloud connector integration, and compliance workflows via API. The proven success also gave us standing to consolidate the firm around ChatGPT as the platform — displacing vertical point solutions that had been competing for budget.

Our Results

Results

3,000+ hours saved per year from the CIM Write-Up GPT alone (1,000 CIMs × ~3 hours saved each). < 2 minutes to first draft, down from 4+ hours. ~95% factual accuracy validated through human and LLM evals. 92% weekly active user rate across the 150-person firm. 2,156% increase in connector usage after enabling enterprise integrations.

The CIM write-up became the proof point that unlocked everything else — deal teams went from skeptical to requesting new GPTs, and leadership consolidated around a single AI platform instead of layering point solutions.

"Eliza gave us a clear roadmap to real, firm-wide value from ChatGPT, far beyond simple tasks. Without them, we wouldn't have hit the ROI we needed."— Head of AI

What this means for PE firms

Start with high ROI, then build momentum. The CIM write-up was chosen because it was high-frequency, high-effort, and could deliver measurable value fast — not because it was the hardest problem.
Assessment is the unlock. Getting the COO, Head of AI, and deal partners aligned on one prioritized roadmap prevented the fragmentation that kills most enterprise AI programs.
Measure everything, or it doesn't stick. Human evals, LLM evals, and usage monitoring gave leadership the confidence to expand — because they could see the ROI in their own data.