Signet
Signet · A division of Circle Square Consulting
We build and run AI agents inside your business — every one of them reviewed, logged, and owned by a named human. So when a client, an auditor, or your own board asks who decided what, and on what basis — the answer is already written down.
See what you can put to workWe build your Claude setup — connected, loaded with your knowledge, and used.
Security and AI policies, signed, and a platform to keep them current.
Agents for a workflow. Custom software when it needs to be software.
A named engineer inside your business. That's where this goes.
The gap
Your team already uses AI. They opened a chat window, pasted something in, and got something useful back. That part is finished — the awareness gap closed a year ago.
What almost no business has is an agent doing a job. Something that reads the inbox every morning, drafts the response, files it in the right system, and knows when to stop and ask a person. 74% of leaders say AI has improved productivity — yet no single function has moved past pilot for more than about a quarter of them.
The distance between those two numbers is not enthusiasm. It is the last mile: connecting to the systems you actually run, working with data that is actually messy, and having someone accountable when it gets something wrong at 2am.
"Couldn't we just have AI do it?"
You could — and that is the trap. An agent that nobody reviews, nobody logs, and nobody signs for is a liability dressed up as a solution. The moment a client, an insurer, an auditor, or opposing counsel asks "who decided this, and on what basis?" — the gap becomes the story.
Nobody signs their name to a chatbot's output. We sign ours.
The method
One process runs underneath everything we sell. We built it delivering attested policy and governance programs where being wrong had consequences, and it is the reason our output holds up when someone pushes on it. AI does the typing. A named human makes the judgment calls. The harness is everything that keeps that sentence true.
We encode how your business actually works — your systems, your rules, your obligations, your house style. This is the asset. It is yours, it is written down, and it compounds.
AI does the volume work: reading, classifying, routing, drafting, reconciling, summarizing. The part that is repetitive, rule-bound, and eats your team's week.
Every output that carries consequence passes a named human, against a written checklist — not a spot check, and not a vibe. Low-stakes work runs unattended; we tell you which is which.
Attestation, logging, and a defined escalation path. You can always answer who decided what, on what basis, and where the record lives.
Anything a reviewer corrects twice gets promoted into the context pack permanently. The agent gets more accurate every month — and cheaper to run.
The context pack, the policies, the governance record. If you ever bring this in-house, you are not starting over. That is the point.
Why Circle Square
Grounded in your reality
We manage technology environments for a living — the credentials, the security stack, the ticket history. Your agents describe what your business actually does, not what a template imagines it does.
We run it on ourselves first
Every agent we sell, we operate internally before we offer it. Our own ticket queue is triaged by one. Our own operating brief is written by one. We are the first reference.
Compliance is home field
We deliver attested security policy programs mapped to NIST CSF 2.0 and CIS Controls v8. Governance is not bolted on afterward — it is the delivery method, whatever we are building.
We ship real software
Beyond services, we design, build, and operate commercial products of our own. When a workflow needs an application rather than an agent, that is not a subcontract for us.
Who we build for
The method is the same everywhere. The evidence, the vocabulary, and the first thirty minutes are built separately for the three industries we know best.
Law firms · 15–75 attorneys
The renewal checklist, the readiness assessment, and the Renewal Dry Run — thirty minutes on your last liability application. You keep the findings.
Registered advisers · $250M–$5B AUM
The pre-exam checklist, the evidence assessment, and the Evidence File — thirty minutes on the artifacts an examiner, a carrier and an acquirer each want.
Specialty manufacturers · $10M–$150M
The Quote Clock checklist, the assessment, and thirty minutes on your own RFQ numbers. We say honestly whether an agent helps.
Free tool
Every carrier, examiner and enterprise customer has started asking the same question: what is your policy on AI? Most firms do not have one. This builds you a first draft in about ten minutes — specific to the tools you actually use, not a template with your name pasted on top.
AI Use Policy Starter · ten minutes · no cost
Answer five short sets of questions about the tools your people use, the data you handle, and the uses you want off the table. You get a draft you can circulate the same day, plus an honest read on where a generic policy stops being enough for your obligations.
Offerings
Every one stands alone and is priced alone. Nothing obligates the next thing. They connect because each engagement makes the next one cheaper — your context pack carries forward — not because we bundle them.
Two ways we bill, and only two
Fixed fee for work with a known shape — setup, policies, agent builds. You know the number before we start.
A monthly block for custom applications and longer engagements. $6,000 a month buys 30 hours. We agree at the start of each month what we're working towards; when we hit it, we tell you and keep going. It's more affordable, more predictable, and it lets us change direction as you watch the thing run in production — instead of arguing with a statement of work written before anyone knew what it would need.
We come in and build your Claude setup — connected to your real systems, loaded with your real knowledge, and actually used by your team.
| What | Included | Fee |
|---|---|---|
| Deploy | Tenant, seats, admin controls, SSO, data-handling settings, admin runbook | $2,900 |
| Connect | Claude wired into email, files, CRM, PSA, accounting — scoped and permission-bounded | $3,900 |
| Build | Projects, named skills, house style, and your context pack | $4,900 |
| Enable | Working sessions by role, a champion per team, usage instrumented, 30-day check-in | $3,900 |
| All four · 5–6 weeks | $12,900 | |
Security and AI policies written around how your business actually operates — and signed by a named person who stands behind them.
| What | Included | Fee |
|---|---|---|
| AI Policies | Acceptable use, tool and model inventory, vendor risk, human-review procedures, employee briefing | $4,900 |
| Security Policies | The four clients and carriers ask for first, plus a NIST CSF 2.0 / CIS v8 crosswalk | $4,900 |
| Full Program | The complete ten-policy framework. Sold as Compliance Ready for firms starting from nothing | $8,900 |
Somewhere for your policies to actually live — current, shareable, and ready when someone asks.
| What | Included | Fee |
|---|---|---|
| The portal | Policies, evidence and assessments in one place your clients and carrier can trust | included |
| Annual review | Full review and refreshed attestation, timed to your insurance renewal | included |
| Staying current | Regulatory-change monitoring and questionnaire support, all year | included |
One agent does one job, unattended, inside a system you already use.
| What | Included | Fee |
|---|---|---|
| Blueprint | The spec: what it does, what it never does, where it escalates — plus a fixed quote and a real off-ramp | $4,900 |
| Build | Integrated with your system of record, piloted on live work, launched, tuned 30 days | $10,000 |
| Care optional | Monitoring and tuning after launch, with a named engineer | $750/agent/mo |
| Blueprint + Build · one agent end to end | $14,900 | |
When the answer isn't an agent — it's software you don't have yet. Often with agents inside it.
| What | Included | Fee |
|---|---|---|
| How a month works | We agree what we're working towards, we hit it, we tell you, we keep going. No rigid SOW to fight with when reality changes. | $6,000 |
| Month 1 | Requirements, architecture, working prototype. Either party can stop here. | $6,000 |
| Month 2 | Environments, identity, schema, integration access. Scope freezes at the end. | $6,000 |
| Months 3–6 | Build, demoed monthly. User testing, pilot, cutover, 30 days hypercare inside the last month | $6,000/mo |
| Typical engagement · live at the end of it | $36,000 | |
An AI engineer inside your business — in your standups, your chat, and your backlog.
| What | Included | Fee |
|---|---|---|
| One day a week | ~30 productive hours. A named engineer, not a rotating bench | $5,900/mo |
| Two days a week | ~60 hours. Enough continuity to own a system end to end | $11,400/mo |
| Four days a week | ~120 hours. The version that replaces a hire you were considering | $21,900/mo |
The five agent templates
Frontdesk inbound triage · Scribe document production · Books finance back office · Recruit hiring and onboarding operations · Brief your operating rhythm. You choose one and we adapt it to your systems, or we build custom. Recruit deliberately does not screen, score or rank candidates.
What you are buying
An agent is not a chat window somebody has to remember to open. It is a process that runs on its own inside a system you already use, and hands work to a person only when it should. Six parts, every time.
Something happens in your stack. An email lands, a ticket opens, an invoice ages past thirty days, a form gets submitted. No one has to start it.
The written rules for this workflow — your categories, your thresholds, your tone, your obligations. This is what makes it yours instead of generic, and it is the asset you keep.
Classifies, extracts, matches against your history, and writes the response or the record. The repetitive judgment-light work that currently eats somebody's morning.
A note on the ticket, a draft in the mailbox, a row in the ledger, a status change. This is the part most stalled AI pilots skip — and the reason they stall.
A written list of what it must never decide alone. Those are drafted and handed to a named person with the reasoning attached. Never sent.
Every action recorded. Review rate and override rate tracked. When a client or a carrier asks how it is governed, you have the answer in a report.
What the managing partner actually notices
The inbox is already triaged before the first person arrives, and nothing went out unreviewed that shouldn't have.
| When | What happens |
|---|---|
| Weeks 1–2 Blueprint | Workflow spec and data-readiness check. What it does, what it never does, what good looks like, and a fixed quote for the build. You approve this before anything is written. |
| Week 3 | Context pack written, and read/write access wired into your system of record |
| Weeks 4–5 | Build and test against your real historical data — not synthetic examples |
| Week 6 | Pilot on a live subset with every output reviewed by a person. Tune against what actually comes back |
| Week 7 | Launch, runbook, and training for the internal owner |
| After | Thirty days of tuning included. Then Care at $750/month, or you run it yourself |
What you own at the end
The workflow spec, the context pack, the runbook, and the governance record. If you ever take this in-house or move on, you leave with all of it.
How it fits together
This is not a menu of unrelated services that happen to share a logo. Every offering feeds the one after it, and the whole thing gets less expensive to run the longer you are in it.
Deploys and connects it, and produces the context pack everything else builds on.
The rules, signed — and somewhere they stay current instead of going stale.
Automate a workflow. Templates and context mean agent three costs far less than agent one.
When it needs to be software. Monthly block, so it can change as you watch it run.
The destination. When you want capacity instead of quotes.
Your context pack starts on day one of Claude Setup and never resets. Every correction a reviewer makes twice becomes a permanent rule. Every agent we build inherits everything the last one learned about your business.
By year two, an agent that took six weeks to build in year one takes two — and the accuracy starts higher than the first one ended.
The vCISO signs the policies. The named engineer signs the agents. Same principle either way: a person whose name is attached, whose judgment is documented, and who can be asked to explain it.
That is what makes this work in a law firm, a bank, a clinic, a factory, a contractor, or a franchise group — and it is the part that generic AI vendors structurally cannot offer.
See one work
Six things arrive the way they arrive on any Monday. Step through them one at a time — the agent explains each move as it makes it. Watch the two it refuses to touch.
The two it refused are the whole point
A contract termination and a site outage never got an automatic reply. They were read, classified, summarized, and put in front of a named person with the reasoning attached. Any vendor can show you the four it handled. Ask them to show you the two it wouldn't.
What you'd need
Rather than hand you a rate card, answer these and we will show you the one to four things that actually apply to your business — and what they cost.
Nothing is sent anywhere. This runs entirely in your browser.
Be honest. Most businesses are further back than they think.
Pick up to two.
Clients, insurers, auditors, a board, a big customer's procurement team.
The honest answer here decides whether this lasts past month three.
Fixed fees. Model and licence costs pass through at cost — we don't mark up tokens. Every engagement carries a written exclusions list and change control.
| Offering | What it is | Timeline | Investment |
|---|---|---|---|
| Claude Setup — Deploy | Tenant, seats, admin controls, SSO, runbook | 1 week | $2,900 |
| Claude Setup — Connect | Connectors into your real systems, scoped and bounded | 1–2 weeks | $3,900 |
| Claude Setup — Build | Projects, skills, house style, and your context pack | 2 weeks | $4,900 |
| Claude Setup — Enable | Sessions by role, champions, usage instrumented | 2 weeks | $3,900 |
| Claude Setup — all four | Saves $2,700 against buying them separately | 5–6 weeks | $12,900 |
| AI Policies | Acceptable use, tool inventory, vendor risk, attestation | 3 weeks | $4,900 |
| Security Policies | Four policies plus NIST CSF 2.0 / CIS v8 crosswalk | 4–6 weeks | $4,900 |
| Full Program | Ten-policy framework. Compliance Ready for firms starting from nothing | 6–8 weeks | $8,900 |
| Agent — Blueprint | Workflow spec, escalation rules, fixed quote for the build | 2 weeks | $4,900 |
| Agent — Build | Integrated, piloted, launched, tuned 30 days | 4–5 weeks | $10,000 |
| Agent — both | One agent end to end. Each additional agent $9,900 | 6–7 weeks | $14,900 |
| Compliance advisory | Assessments, audit support, governance consulting | Scoped | $225/hr |
| Offering | What it covers | Monthly |
|---|---|---|
| GRC Platform | Portal, annual review with refreshed attestation, regulatory monitoring, questionnaire support | $495 |
| Agent Care | Monitoring, tuning, and a named engineer — per agent, optional | $750 |
| Custom Applications | A 30-hour block at $200/hr. Typically six months to live | $6,000 |
| Embedded Engineer | Named engineer — one, two, or four days a week | $5,900 / $11,400 / $21,900 |
What is never included
Third-party licensing, AI model and API consumption, and cloud costs are billed to you at cost or held in your own accounts. We do not mark up tokens.
How fixed fee stays fixed
Every engagement carries a written exclusions list and a change-control process. Requests outside scope pause at the boundary and proceed under a written change order — never a surprise invoice.
How to start
30 minutes · no obligation
Bring one process that eats your week. We will tell you honestly whether an agent can do it, what it would take, and what it would cost — or that it cannot, and why. You keep the assessment either way.
30 minutes · no obligation
We compare your last insurance application against your actual environment and show you exactly where you are exposed. No obligation — you keep the findings.
In legal, wealth management, or manufacturing? Your industry has its own free front door.
Then, if it makes sense
Claude Setup starts at $2,900 and takes a week. A Blueprint is $4,900 and takes two. Either one is a real commitment with a real deliverable at the end, and neither obligates the next thing.