AI workflows
that keep working
without me.
I help small teams build, measure, and maintain AI workflows for leads, proposals, reporting, and operations. Whether your team is brand-new to AI or already experimenting with automation, the goal is the same: useful systems people can run without me in the room.
Built to run without me.
Scattered AI use is not a workflow.
I'm Andrew Gooding. I help small teams turn repeated work into AI-assisted systems with clear inputs, useful outputs, review rules, handoffs, and a way to tell whether the workflow saved time.
Some teams need setup and safe usage. Others already know the tools and need deeper workflow design, integrations, and measurement. I meet the team at its current AI level, then build toward workflows people can use without me in the room.
- 4+ years building with AI systems
- Workflow design, automation, and handoff
- Background in computer science
- Based in Seattle, working across timezones
Three places AI usually pays for itself first.
I start where the work repeats, the handoffs are messy, and the result can be measured. Also useful for support triage, internal knowledge lookup, CRM notes, and document summaries when the review rules are clear.
Lead intake and follow-up
A lead comes in. Nobody should have to chase the inbox. Capture it, draft the reply, log the details, and create the follow-up while a human stays in control.
Proposal, quote, and client email drafting
A proposal needs to go out. Nobody should start from a blank page. Turn structured intake into a reviewed first draft with your examples, voice, and approval rules built in.
Weekly reporting and operations rollups
A report is due every week. Nobody should rebuild it from scratch. Pull scattered updates into a consistent rollup with wins, risks, blockers, metrics, and next actions.
Enervest, onsite for three days.
An Australian battery and energy company working on government contracts. A 31-person team, unevenly familiar with Claude — brought in across three onsite days to turn curiosity into practical, day-to-day use.
A team of 31, practical in three days.
Battery and energy, working with the Australian Government.
What I delivered
- 01Onsite session and a Claude Usage Guide tailored to how the team actually worked.
- 02Ran about 10 one-on-one setup and training sessions across different roles and comfort levels.
- 03Ready-to-use prompt patterns for recurring work, written in plain language rather than AI jargon.
- 04A custom prompt-builder that made stronger prompts faster and more consistent.
- 05A clear lesson for future work: training helps, but workflows are what teams keep using after the session ends.
Outcome
The team left with better AI fluency and a clearer sense of where Claude fit. The next level is turning that fluency into workflows that run without me in the room.
Two ways to work together.
Smaller teams can fix one painful workflow and keep optional light aftercare. Larger teams can roll out, measure, and maintain workflows over time.
Door 1 · One-time workflow build
Fix one painful workflow, get it handed off, and keep optional light aftercare. Best for smaller teams.
Map one workflow, define the business metric, and decide whether it is worth building.
- Short intake before the call so I can see the workflow
- Current-state map: trigger, inputs, handoffs, and review point
- One or two workflow opportunities ranked by usefulness
- Baseline metric to measure before anything gets built
- Full fee credited toward a build if you go ahead
Build one repeatable workflow, hand it off, and support it lightly after launch.
- Workflow map and tool choice before build
- AI-assisted workflow configured in the tools you already use
- Test examples and a short SOP for the team
- Walkthrough recording or live handoff
- 30 days of bug-fix support, with larger changes scoped separately
Door 2 · Ongoing workflow partner
Roll workflows out across a team, then measure and maintain them over time. Best for larger or growing teams.
Roll out workflows across a team, with adoption, review rules, and measurement.
- Discovery across one or two business functions
- Workflow build plus team rollout plan
- Role-specific usage guidance where the team needs it
- Owner, fallback, and review rules for sensitive outputs
- First adoption and ROI baseline
Maintain current workflows and improve them as tools, models, and team needs change.
- Maintain: up to two active workflows, monthly health check and fixes
- Partner: up to three active workflows, plus a monthly improvement cycle
- Model and tool compatibility updates as things change
- Adoption or ROI signals tracked on Partner
- New full workflow builds scoped separately
Built to run without me. Maintained to keep getting better.
Maintain keeps up to two active workflows healthy with monthly checks, light fixes, and model or tool compatibility updates. Partner covers up to three active workflows, adds a monthly improvement cycle, and tracks adoption or ROI signals. New full workflow builds are scoped separately.
One-time build
Best for small teams that want one painful workflow fixed, handed off, and supported lightly after launch.
Ongoing partner
Best for teams that want workflow rollout, measurement, and maintenance as tools and team needs change.
Embedded option
For complex teams, I can work as an embedded workflow partner for a defined period so I can learn the process from inside the business.
The scanner shows the what. I build the how.
In about a minute, the free scanner names one workflow worth a closer look: the likely bottleneck, the business metric it affects, and the unknowns to check first. It is a starting point, not a finished consulting diagnosis — paid work turns that idea into a usable system with owners, review rules, measurement, and maintenance.
What it can point to
A likely workflow opportunity, the business metric it may affect, and the unknowns to verify before anyone spends money building.
What paid work adds
I map edge cases, connect approved tools, define review rules, measure a baseline, and make the workflow maintainable.
What it will not fake
No invented ROI, payback period, or dollar impact without your real baseline and follow-up measurement.
Data, review, and common sense.
AI workflow projects often touch real business information. The goal is useful automation without pretending every task should be fully automated.
Approved tools first
I work inside approved tools where possible instead of forcing a new platform before the workflow proves itself.
Narrow, scoped access
I only ask for the access a workflow actually needs. Sensitive data stays out of AI systems unless there is a clear reason, explicit approval, and a safer handling rule.
The right kind of review
Client-facing, high-risk, or judgment-heavy outputs keep human review. Legal, financial, tax, medical, and compliance-relevant outputs require qualified review before anyone relies on them.
Built to be owned
Every workflow gets an owner, fallback, and maintenance plan before handoff so the system can keep working without me.
Words from the people who hired me.
Three voices from the Enervest engagement — leadership, finance, and the front line.
Andrew delivered a compelling presentation that positioned AI as a strategic enabler for the organisation. He clearly articulated both the opportunity and the responsibility of adoption, encouraging the team to think critically about how AI can support future business performance.
Andrew approached AI from a practical business perspective, focusing on efficiency, consistency and governance rather than hype. His work helped identify where AI can create measurable value across internal processes while still maintaining appropriate structure and control.
Through his guidance, education and practical consultation, I estimate that I have saved approximately 4–6 hours per week by streamlining drafting, document preparation, research, process design and day-to-day administrative workflows. I would happily recommend Andrew to any individual or business looking to better understand and apply AI in a practical, productive and commercially useful way.
Book a workflow audit.
Pick the closest engagement type, then a time. You will get a calendar invite and a short intake so we can start with the workflow, not small talk.
Every session ends with concrete steps.
Selected: Workflow audit / $1,500
- Video call via Google Meet or Zoom, your pick.
- Recording and written summary delivered same day.
- Reschedule free up to 24 hours before the session.
- If AI is not worth building for the workflow, I will say so.
Five questions I will ask before the audit.
- What AI tools you currently use
- Your biggest time-drain or weak spot right now
- What success looks like after the session
- Your role at the company
- Whether you are the decision-maker for AI investment
Three minutes to fill out. It means we skip the small talk and get straight to the workflow on the call.
Not ready to book? Send a note.
For quick questions, partnership ideas, or longer engagement conversations. I read every message myself and reply within two business days.
- Emailandrew@andrewgooding.com
- Based inSeattle, PST
- AvailableWeekdays, 9am to 5pm PST