Articles
Notes on AI integration.
Short writing on what we see across client engagements, the gap between enterprise AI pilots and production results, and the patterns that actually move organizations forward.
-
What private AI actually means — and when your business needs it
Private AI means running AI on infrastructure you control, so your data never leaves your custody. What that means, when it matters versus when public cloud is fine, and where a small business starts.
-
Private AI for property managers
Property managers and HOA firms hold tenant PII, financials, and governing documents. Private AI keeps that data in-house — and lets one manager serve more communities without the exposure of public-cloud AI.
-
Private AI for contractors
Contractors handle bids, plans, contracts, and client data — often with no IT staff and real confidentiality obligations. Private AI keeps it in-house, and Interchange lets you bill AI back to the job.
-
Private AI for utilities
Utilities can use AI for customer communications, outreach, and program content while keeping customer data in-house. California law fences in customer usage data — private AI sidesteps the disclosure problem.
-
Right-size your in-house AI server — measure before you buy
Don't guess how much GPU you need. Measure real usage first — concurrency, context length, time-to-first-token, tokens per second — then spec the server. The Interchange gateway is the meter.
-
AI as a pass-through cost — cost recovery for property management & HOAs
Management agreements already pass itemized operating costs through to owners and associations. AI used on a property's behalf is one more recoverable line item — with per-property attribution.
-
AI as a line item — cost recovery for marketing agencies
Agencies use AI heavily — and most are subsidizing it out of their own margin. SOWs already itemize stock licensing and third-party tool costs. AI tokens belong on the same line.
-
Project-billed AI for small engineering firms
AIA and ACEC standard contracts already provide for reimbursable expenses. AI used for code research, calc verification, and report drafting joins the same line.
-
AI as cost of goods sold — recovery in custom manufacturing
For custom manufacturers, AI absorbs into COGS per job — not SG&A. A cleaner accounting story than services: cost-per-unit absorbs the spend, gross margin reflects reality.
-
AI as a project cost code — cost recovery for construction firms
Cost-plus and T&M construction contracts already pass direct costs through with markup. AI used for bid prep, RFIs, submittals, and change orders is one more cost code.
-
AI cost recovery for small law firms — the modern Westlaw model
Law firms have billed back Westlaw and LexisNexis usage to clients for forty years. AI research, drafting, and document review fit the same soft-cost pattern.
-
Why 95% of enterprise AI pilots fail — and what MIT's research actually shows
MIT's State of AI in Business study made headlines with the 95% failure figure. The interesting detail isn't the failure rate — it's what separates the 5% that succeed.