GuideA guide for firms adopting AI, in three steps
General AI assistant or vertical AI: where each belongs
Start with an enterprise assistant for everyone, learn where it stops, then buy specialised tools only for the workflows that decide the outcome.
The argumentWhy not begin with a specialised product
Do not start with a vertical product.
Give everyone a general assistant first, find out where it stops, and only then buy specialised tools for the few workflows where accuracy and integration matter.
The goal of the first step is not to overhaul your due diligence overnight. It is to build an AI-ready culture, use the simple cases at low cost, and learn their limits.
Step 1Try an enterprise assistant on simple, frequent tasks
Step 1. Experience the assistant.
Subscribe the team to an enterprise-grade general assistant. These are inexpensive, secure and connect to the tools the firm already uses: shared documents, e-mail and cloud servers. Choose cases that are simple, frequent and limited to writing in plain language:
- Drafting e-mails and investor communications
- Preparing for meetings by summarizing earlier notes and call transcripts
- Generating first ideas for thematic write-ups
- Cleaning up and formatting basic data sets
Step 2Where a general assistant reaches its limits
Step 2. Understand the limits.
On complex private-investment work, a general-purpose assistant meets three boundaries.
Quality and nuance
Long reports, such as a full diligence analysis, lack depth and accuracy. The tools are not made for exact calculations, industry-specific logic or complicated reporting.
No integration
They do not connect to the firm’s CRM or to the industry databases it relies on.
No workflow design
You cannot define reusable templates or multi-step processes for tasks such as deal screening or portfolio analysis.
For the most critical, value-adding workflows you then need agents trained on the firm’s own data, methods and objectives.
Step 3Five places specialised tools earn their cost
Step 3. Add specialised tools where they pay.
- Legal
Document review
Analyze NDAs, credit agreements and purchase agreements; extract key clauses, flag risks, compare terms with the firm’s precedent positions.
- Finance
Modelling and financial analysis
Build models from simple prompts, automate sensitivity analysis, run scenarios quickly and audit formulas for errors.
- Deals
Deal screening
Review information memoranda, summarize each in a standard one-pager, score them on the investment committee’s criteria, and connect to the CRM.
- Deals
Deal sourcing and execution
From screening through data-room analysis to financial models, inside the CRM and workflow, to judge opportunities faster and with less diligence risk.
- Portfolio
Portfolio management and operations
Performance monitoring, answers from the internal knowledge base, and faster reports and board materials.
For operating companiesThe same order applies outside funds
The order is the same for a company.
The guide was written for investment firms. The order holds for an operating company too: start broad, learn where the general tool stops, then specialise where the workflow decides the result.
OriginalThe guide as first published on aix2.ai
The original post
Written for private investment firms, with a link to the aix2.ai use cases page.