BUILD WITH US
We deploy your own automation.
An engineer from our bench builds the system inside your firm, on the same parts that run our own back office: the report renderer, the evidence standard and the stage method. Then we run it for you, run it with your team, or hand it over.
AGENT SPECIFICATION
Quarterly portfolio update
- Reads
- Investor updates from your portfolio companies, as they arrive by email.
- Decides
- Which metrics moved beyond the thresholds you set.
- Stops for
- A partner's approval before anything leaves the firm.
- Writes
- One quarterly view in your template, every figure linked to its source page.
Illustrative specification. Every build starts as one of these, written with your team.
Four things we build.
Company brain
Your own documents and data, searchable, with every answer citing the file and the page.
For example: every deal your fund has looked at, with the memo and the committee's reason; your portfolio companies' reports, so a question about runway returns the page.
You get: an index over the documents you choose, a way to ask it by email or chat, and access that follows your own permissions.
Agentic automation
Agents that run your recurring work on a trigger or a schedule, and stop for a person at every decision that needs one.
For example: every new application triaged against your thesis on arrival; investor updates collected and summarized each quarter; the committee pack assembled from the deal file; the month-end close checklist across several entities; the lender and board pack, assembled for the CFO to approve.
You get: the agent running in your environment, its written specification (what it reads, what it decides, who approves), and its output in your template.
Custom models
Small models trained on your data, in your environment and for you, for narrow tasks where a large general model is the wrong tool. We use such a model for anyone else only with your written consent.
For example: reading the tables in portfolio company reports; extracting the same fields from the same kind of document, thousands of times.
You get: the model in your environment, tested on your own held-out data, with the results reported to you.
Data unification
Your records from the systems that matter, matched in one graph database, with the evidence for each match.
For example: one customer, location or supplier, matched across the companies a platform has bought, each with its own ERP and CRM. Or the companies and founders your fund has met, matched across the CRM, the data rooms and the portfolio reports.
You get: a connector to each source, the matching rules written down for your review, and a graph database your reports run on, with every link traced to the record it came from.
HOW AN ENGAGEMENT RUNS
Find it, write it down, build it, run it.
01
Find the work that repeats.
An engineer from our bench sits with your team and learns which documents arrive, who decides and what the output has to look like.
02
Write it down.
Each workflow becomes a specification: what it reads, what it decides, and who approves.
03
Build and test it on your materials.
We build in your environment and test against your real documents and your own standards.
04
Run it, or hand it over.
We run it for you, run it with your team, or train your team to own it.
WHERE EVERY BUILD STARTS
First, a paid discovery.
Every build starts with a paid discovery. We document the workflow, the data and the systems it touches, and give you a fixed quote for the build. The first conversation is free.
How long the discovery takes.
One to two weeks, before anything is built.
How the fee is set.
The discovery has a fixed scope and fee, agreed before it starts. You keep the documentation whether or not you build with us.
Who does the build.
An engineer from our bench of forward-deployed engineers, financial analysts and subject experts, from our network and partner firms. We name each person to you before they join your engagement, including engineers added as the build grows. You can usually interview each person first.
AFTER THE BUILD
Three ways to run it.
We run it for you.
For a fixed monthly fee, we watch the runs, fix what breaks and improve the agent from your corrections.
We run it with your team.
Your people run it; we stay on call and review it each quarter.
Your team takes it over.
We train your team, hand over the specification and step back. The build stays in your environment.
WHAT EVERY BUILD STARTS FROM
Parts already at work in our own back office.
The report renderer.
Every output in your house style, as Word, PowerPoint, Excel or PDF, from one content specification.
The evidence standard.
A source behind every figure, a label on every claim we check, and a register of corrections.
The stage method.
Long work broken into stages, each checked before the next begins, with one file that later work resumes from.
See them at work in Lensential Venture and Lensential Consult. Our own firm runs on them too, and our own fees were set with the pricing method in Lensential Consult.
WHAT YOU CAN INSPECT
We run our own back office on the same method and reports. Ask us to open them on a call.
WHAT STAYS WITH YOU
Your environment, data and documents stay yours, and so do the outputs. Our tools stay ours, and your build carries a license to the parts it uses; if we part ways, the build stays in your environment.