Command Palette
Type to search. Press Enter for the top match.
runpoint.
← All case studies
Field-guide line illustration of a partner at a wood conference table reviewing diligence reports and a laptop dashboard, mountain horizon through the window
Tech due diligence advisory

A global tech due-diligence advisor

We started by building two new tools the firm did not have. Once those proved out, the client trusted us inside its main product, and the engagement expanded toward a platform the firm owns outright.

3 months → 1 month
Time to ship a production feature, compared with the prior internal baseline
1 operator-engineer
One thread from analysis and product definition through implementation

Before

The client is a global advisory firm whose consultants do the technology due-diligence work behind hundreds of private-equity deals each year. Over more than a decade they had built an enormous archive of past assessments that, on paper, was the firm's competitive moat. The data belonged to the firm, but it sat in systems that could not put it to work in real time.

The first bet

Two tools they didn't have yet

We started with two new builds because they were the safest places to prove ourselves. The first was a staffing tool that ingested a candidate's resume, scraped their LinkedIn, tagged their experience, and matched them to the right kind of project, producing a defensible answer to the question every services firm asks every week. The second was a workforce classifier that took a client's organizational data and classified every employee by role and seniority, with the reasoning shown for each one. Slow, manual analysis turned into fast analysis with the work visible.

What we noticed

Both tools worked, so the client let us into their main product. We started shipping fully-vetted features into their core codebases at roughly a third of the time their internal teams took, by collapsing what is normally a four-role handoff (analyst, product manager, designer, engineer) into a single operator-engineer who could move from sketch to merged PR without losing a week to coordination.

How we work in their codebase

We work on the what. The agent works on the how.

What we focus on

  1. 01Which features are actually worth building
  2. 02What good looks like for the business
  3. 03Talking to the people who'll use it
  4. 04Reviewing what the agent ships before it lands
  5. 05Picking what to do next

What the agent does

  1. 01Reads the ticket
  2. 02Walks the codebase
  3. 03Writes the code
  4. 04Opens the PR
  5. 05Runs the tests
  6. 06Waits for review

Most of our PRs come back overnight. We come in the next morning, review, merge what's good, and pick the next batch.

What we built next

01

A shared starting point for everything

We built a production-ready scaffold every new product could start from, which meant each new build got cheaper than the last and the firm stopped re-solving the same plumbing problems on every project.

02

The archive, made useful in real time

We built a tool that puts thousands of past assessments in front of a partner during a live call with a private-equity firm, turning the archive from something the firm referenced after the call into a tool they use during it.

What was hard

Earning access to the core product

The first tools were separate from the firm's main platform. Shipping them well earned the right to contribute inside an established codebase, where every change had to fit the client's authentication, deployment, review, and product standards.

Keeping the business problem through the handoff

The 3:1 speed difference did not come from typing code faster. One operator-engineer carried the analysis, product decisions, interaction design, and implementation as a single thread, with fewer rounds of translation between roles.

How they use it now

Working sessions twice a week. Both teams are getting faster, and patterns we use are starting to show up inside their organization.

What's now possible

The firm's archive answers questions during a deal call, not after one. Each new product they want to build is cheaper than the last, because the platform underneath does more of the work. And inside their own engineering organization, the way Runpoint people ship has become a reference point.

What's next

The firm is now replacing its legacy data-collection platform with a system built around how the work runs today. That is the largest and most sensitive part of the engagement, and it remains in flight.

I don't think I've ever heard the term operator engineer, but I think it fits really well for what you guys do… I've found with you guys, I can just let you go to these meetings by yourselves. I don't even need to be there.

Client lead

Real quotes from real clients. We anonymize on the site until each one says yes to being named. Most are in the queue. References available now.

What we built

SalesStaffing matchmaker. A tool that takes a candidate's resume, scrapes their LinkedIn, tags their experience, and matches them to the right kind of project. The output is a defensible answer to the question every services firm asks every week: who do we put on this.
Built with
ClaudeLinkedIn scrapingFastAPIReact
OperationsWorkforce classifier. A tool that takes a client's organizational data and classifies every employee by role and seniority, with the work shown. Turns a slow, manual analysis into a fast one.
Built with
OpenAI GPT-5.1ReactExpressMSAL (Azure AD)AWS App RunnerPptxGenJS
OperationsMain platform rebuild. We're replacing the firm's legacy data-collection platform with a new system built around how they actually work today.
Built with
Next.jsFastAPISQLAlchemyMSAL (Azure AD)PostgreSQL
OperationsLive archive lookup. A tool that puts thousands of past assessments in front of a partner during a live call with a private-equity firm. The archive went from something they referenced after the call to something they use during it.
Built with
FastAPIAWS LambdaReactMSAL (Azure AD)

Want this kind of work for your business?