AI Workspace Consultant — Claude Cowork Setup + Ongoing Advisory for Small Manufacturer

WHO WE ARE InCon Process Systems designs, fabricates, pilots, and supports wiped-film and short-path distillation systems. We are a specialty manufacturer in the western Chicago suburbs with about seven employees. The company is engineering-heavy and we have no dedicated IT department. Our team already uses Claude in the browser and is comfortable with it. We are not starting from scratch, and we are not looking for someone to explain what AI is. WHAT WE WANT Two things, in this order. First, expand what our team can do — we want our people using AI across more of their daily work and using it well. Second, move reliable work off people's plates. Where the tool can handle something dependably, it should. Where the work takes judgment, our people keep it. One measure we'd use six months out: fewer customer inquiries sitting too long before someone responds. PHASE 1 (about one month) 1. Configure Claude connectors for Gmail, Google Calendar, and Slack. Slack scoped by explicit allow-list of channels, not a deny-list. One channel excluded. 2. Build a recurring Project Status and Action Summary. Open items and completed items over a trailing one to two weeks, plus upcoming meetings with context on what each concerns. It should track what we owe others and what others owe us, identify likely owners, and flag unanswered customer inquiries, stalled work, and emerging risks. Uncertain conclusions should be separated from established ones. Important: we have no project management system, no CRM, and no tracking software. Project status currently lives spread across email, Slack, and people's heads. The digest is not summarizing a list that already exists — it has to build the project picture out of ordinary communication. That is the point of it, and it is the difficult part. 3. Role-based visibility. Each user gets a personalized view of the work they're involved with, delivered by Slack DM. A single company-wide summary for everyone is not sufficient. CEO and President get company-wide visibility and should be able to ask cross-company questions on demand — what's behind, what's due, what we're waiting on, what looks at risk. 4. Where the system identifies work needing attention and has enough context, it should prepare a next step for human review — a follow-up, an acknowledgment, a request for missing information. Nothing sends or changes anything externally without our review and approval. If information is insufficient, say what's missing rather than inventing an answer. 5. Discovery interviews with our users, one at a time. Not as a group, not through a single point of contact. We want you to understand what we do well enough to suggest things we haven't thought of. This is paid work, not part of your pitch. 6. Training and a written runbook. Our people are technically capable and learn quickly — teach the architecture, not button-clicking. Our internal IT staffer needs enough to handle first-line questions, but is not taking ownership of this system. 7. A written recommendations document for what we should do next. This carries real weight in how we evaluate the engagement. OUT OF SCOPE FOR PHASE 1 Standalone quotation drafting. Meeting capture and transcription (note: meeting outcomes later documented in Gmail or Slack are in scope). Dropbox and local file integration. Automation platform work — n8n, Make, Zapier. Chatbots, voice agents, customer-facing deployments, software development. Everyday writing help — someone asking Claude to tighten an email — is ordinary use of the tool. That needs training, not development. Please don't quote build hours for it. THE ONGOING ROLE — READ THIS This matters more than the build in deciding fit. We want someone who checks in regularly with our users to hear what they're doing, what they want to do, and what they could be doing better, and turns that into improvements. Monthly to start, with a strong expectation of at least six months following Phase 1. We don't think an internal non-specialist can do this. The technology keeps changing and staying current is somebody's job. If your strength is implementation and your interest ends at handoff, we're probably not a good match. The build is the way in. The ongoing cycle of listening and improving is the work. OUR ENVIRONMENT Windows. Gmail is our system of record. Slack with 10 to 20 active channels. Four-seat license pack, likely growing to five or six users. Documents in Dropbox with some local storage (outside Phase 1). No project tracking in place. WORKING ARRANGEMENTS Remote work on our systems with real content, under a named account we provision and can revoke. NDA signed before any data access — our agreement, not the platform default. All material returned or deleted at close unless a continuing agreement is in place. Someone local to Chicago is a mild preference, not a requirement. HOW TO PROPOSE Give us an hourly rate and an estimate of hours broken out across the deliverables above. We'll work not-to-exceed against that estimate. Include your proposed structure and rate for ongoing monthly support. Please do not send a single lump sum. We want to see where the hours go.