The three tools
- A client portal and data room. A branded portal per engagement: a files tab, a chat tab, and up to four custom pages, an overview and forecast, a deals view, a data room, and a client-updates, billing or people view depending on the engagement. Pages are versioned and stay hidden from the client until deliberately published, so a page in progress never gets seen half-finished. On the firm's own side it gives one cockpit across every client at once: snapshot tiles, a forward look at recurring charges, and a client-health table, so a partner does not have to open five different engagements to know which ones need attention this week.
- A search, generative-engine and answer-engine visibility service. A scoring engine that audits how a site actually shows up in traditional search, in an AI answer engine's response, and in a language model's own retrieval, then drives the specific fixes that move the score. It did not start as a scoring engine; it started as a written standard for how the firm's own sites should be structured for this new kind of visibility, deployed across the firm's own portfolio first. The scoring engine that came after was checked against a hand audit of ten real sites across sixteen distinct check types before anyone trusted its output, and it landed better than the team expected once tested that way.
- A systems, accounts and automations audit. What is running, what it costs, what is exposed, and what should be consolidated, run as the first step before Common Ground installs anything new for a client, on the theory that a firm cannot tell someone else to clean up their stack from a position of never having looked hard at its own.
How the three fit together
The three tools sit behind the firm's engagements rather than in front of them. The portal carries an engagement day to day. The visibility service makes a client, or the firm itself, findable once the engagement's public work is live. The audit tells the team where to start before either of the other two gets involved. Other internal tooling beyond these three, the firm's own relationship database, its analytics, and the agent framework it trains on its own domains, stays inside the firm rather than becoming a fourth client-facing product; not everything built for internal use is meant to ship.
What we kept, replaced and installed
We kept the discipline of using each tool on our own operation before offering it outward, the same rule as every other system in this room. We replaced ad hoc client updates over email with the portal, replaced guessing at search and AI-answer visibility with a checked scoring engine, and replaced an unaudited, growing pile of tools and accounts with a standing first-step audit. We installed all three as a suite that can be white-labeled per client, credited to Common Ground as the firm that created and operates it.
What it cost, and what we would watch
Validating the visibility engine against a real hand audit before naming it cost real time that a faster, unchecked launch would not have spent, and it is the reason the engine is trusted rather than merely hoped to work. What we would watch going forward: the visibility service's own name is still under review internally, which is a small, deliberate signal that a product built for internal use first does not need to rush a public identity before it is ready for one.
What it produced
A branded client portal and data room stood up its first live client workspace in September 2026. The visibility engine, validated against a hand audit of ten sites across sixteen check types before it was named, is now built into every product the firm ships and into its first clients this year. The systems audit runs as the standing first step before any new client tool goes in. All three now run for clients under white-labeled deployments while continuing to run the firm's own portfolio.
A slice of the project list
A few related projects.
- The RFP outreach workflow: a daily bid sweep, a rubric-matched proposal method, and the two-question gate that should run first
- The voice engine: teaching a model to sound like the person it is speaking for, measured by a blind test
- Credibility and the story pipeline: a claims register with a status and a source on every row
- A CRM migrated twice in three weeks, once the first move turned out to be the wrong shape
- The data room pattern: one template and one access standard, now running more than ten rooms
- AI sales agents trained on our own record, held to a human gate before anything reaches a client
- The operating system itself: sessions built by type of work, a measured floor cost, and a review chain before anything ships unattended