Private AI for mid-market companies

“What did we do last time?” Your company finally has the answer.

Every project, proposal, decision, and hard-won fix your team has ever produced — turned into a private AI your people can ask, compare against, and build on. Live in 30 days. And when it's done, you own it.

Book a 30-minute call

Built by the architect behind a company-wide AI knowledge platform serving a 1,500-employee organization.

Your company keeps re-solving problems it already solved.

What did we quote for a job like this in 2023? How did we fix this the last time it broke? Why did we stop working with that vendor? The answers exist — in a document only the person who left knew about. So your best people re-derive them from scratch, on your most expensive hours. The SaaS fixes for this cost six figures a year, forever — and your data lives on someone else's servers.

The status quo

Memory that walks out the door

Institutional knowledge lives in people's heads and in files only they can find. When a veteran leaves — or is just on vacation — "how did we handle this last time?" has no answer, and the work starts over.

The SaaS trap

Renting your own brain

Enterprise AI search platforms run $50–75 per user per month with 100-seat minimums. For a 500-person company, that's $300k+ per year — forever — for access to your own documents.

The shrinking middle

Fewer managed options

AWS is closing Amazon Q Business to new customers in July 2026. Mid-market teams that wanted a managed option now face a choice: expensive SaaS, or build. We're the build — done for you.

Not just finding files. Recalling what you learned — and working with it.

Ask "what did we do last time this came up?" and get the answer, with sources. Then go further: compare today's situation against it, summarize how your approach evolved, draft the new deliverable from the old one. A fixed-scope, fixed-price engagement — live in 30 days.

Discovery & access

Map your data sources, security requirements, and highest-value questions. Stand up the deployment pipeline in your AWS account.

Ingestion & index

Connectors to your file stores and systems. Documents parsed, chunked, embedded, and indexed with permission-aware access controls.

Chat, tuned

The assistant goes live for a pilot group behind your SSO. We tune retrieval and answers against your real questions with your pilot users.

Rollout & handoff

Company-wide launch, admin training, and full handoff: architecture docs, runbooks, and every line of infrastructure-as-code. It's yours.

Your precedent × your industry's rules. One answer.

Generic AI doesn't know your industry's current rules. Your file server doesn't either. So neither can answer the question that actually matters:

"How did we handle this in 2023 — and is that approach still valid under today's rules?"

Your memory

Company precedent

Every project, decision, and document your company has produced — permission-aware and cited, so you can see exactly where an answer came from.

Authority Pack

Your industry's rules, built in

The regulations, codes, and standards your industry runs on — tax authority for accountants, building codes for design firms, carrier guidelines for insurance — curated into the same assistant and kept continuously current. Answers cite your precedent and the live rules together.

Plays nice

Keep your research tools — for now

Already paying per-seat for an industry research tool? Keep it — this is the system that knows what your company decided, which no research tool can. Many teams find that once the Authority Pack is in place, they can consolidate subscriptions and keep the savings.

Own it once vs. rent it forever.

Three-year cost of AI search for a 500-employee company.

Hadlock Systems Enterprise SaaS (e.g. Glean) Copilot-style add-ons
Pricing model One-time build + optional support $50–75 / user / month, 100-seat min. $30 / user / month per suite
3-year cost (500 users) ~$50–75k build + AWS usage costs $900k–$1.35M+ licensing $540k+, limited to one vendor's silo
Where your data lives Your AWS account Vendor's cloud Vendor's cloud
Covers all your sources Yes — custom connectors included Catalog connectors only Mostly its own ecosystem
Who owns it at the end You. All code and infrastructure. Nobody — cancel and it's gone Nobody — cancel and it's gone
Case study

1,500 employees. Years of institutional knowledge, on demand.

For a 1,500-person organization, we built a company-wide AI index and private chat service over the full document corpus — so any employee can ask "how did we handle this before?" in plain English and get the answer with sources, in seconds, instead of hunting for files or interrupting a colleague. The same team then shipped an internal AI app-builder platform that lets non-developers create working web apps in minutes.

1,500employees served
Secondsto surface prior work, with sources
$0per-seat licensing fees

Two ways to run it. No per-seat meters on either.

Have a cloud team? Own it outright in your AWS account for a one-time build. Don't? We run a dedicated environment for you at a flat monthly rate. Either way, price never scales with headcount.

Hosted — companies under ~200 people

from $2,500 / month
  • Same system, fully managed — no cloud team needed
  • Dedicated single-tenant environment (never shared)
  • Hosting, monitoring & upgrades included
  • SSO & permission-aware answers
  • Live in 30 days
Book a call

Authority Pack — add-on

from $1,500 / month
  • Your industry's regulations, codes & standards in the same assistant
  • Continuously updated as the rules change
  • Answers cite your precedent and current rules together
  • Flags where your past approach may be affected by new rules
Book a call

Support Retainer — optional

$3–8k / month, by company size
  • Monitoring, upgrades & model improvements
  • New connectors and features over time
  • Priority response SLA
  • Quarterly roadmap & usage review
  • Cancel anytime — the system still runs. It's yours.
Book a call

Questions a careful buyer asks

Does our data ever leave our environment?

On the Deployment plan, no — ingestion, indexing, and chat all run inside your AWS account, with model inference via AWS Bedrock in your account by default. On the Hosted plan, your company gets a dedicated single-tenant environment that is never shared with or visible to any other client. Either way, we architect to your confidentiality obligations.

What happens if we stop working with you?

Nothing breaks. You receive the full infrastructure-as-code, architecture documentation, and runbooks at handoff. The support retainer is optional and cancelable — the system is yours either way. That's the point.

How do you handle document permissions?

Answers are permission-aware: the assistant only retrieves from documents the asking employee can already access, enforced at query time against your SSO groups and source-system ACLs.

Why not just buy Glean or wait for Microsoft?

If a $300k+/year line item and vendor-hosted data are acceptable, those are fine products. This offer exists for companies that want the capability at a one-time cost, across all their systems, on their own infrastructure — with no ongoing licensing exposure.

We don't have a cloud team — or an AWS account. Can you still help?

Yes — that's exactly what the Hosted plan is for. We run a dedicated environment for your company, fully managed, at a flat monthly rate. If you later build cloud capability, we can migrate the whole system into your own AWS account; it's the same architecture.

How do we know it will actually get used?

It's built into the deal: on the Deployment plan, the final milestone payment isn't due unless your pilot group is using the system weekly by the end of tuning. We tune against your team's real questions — "how did we handle this last time?" — until the answers earn daily use.

Bring us a real “what did we do last time?”

Book a 30-minute call. Bring a question your team wasted hours re-answering this quarter — we'll show you exactly how the system would have answered it in seconds.

Book a call