Sam Davis PhD data scientist · AI builder · new to fractional B2B
B2B Build · 2026 Coogee, Sydney
AI for climate & ESG businesses
Help your team adopt AI without breaking your product.
If you run a climate or sustainability business, you've already started doing AI — or you've been told to. Most of it isn't landing the way it should. The audit feature your accountants wanted is half-deployed. The estimation pipeline still takes a week of manual work per client. Your engineers are stretched between three roadmaps.
I'm a PhD data scientist and AI builder. I come from research, prototyping, and teaching — not enterprise consulting. What I bring is technical depth, eight AI builds in twelve months, and the translator instinct from years of explaining complex science to non-specialists. I work in the codebase, not in slide decks. The fractional B2B practice is one I'm building, and the pricing reflects that.
Who this is for
Climate & ESG
Scaleups, consultancies, advisory firms with a product or delivery process where AI should compound but currently doesn't.
Service businesses
Coaches, professional services, community-builders where AI inside the operations would change the unit economics.
Need a champion
A founder, head of product or eng, ops lead — someone who can carry the work after I'm out. I work alongside, not instead of.
What I keep seeing
01
The AI isn't the bottleneck — the fifteen integration points around it are. The model works in a notebook. Then it needs to read your customer's data format, write to your audit-trail schema, route to your CRM, and explain itself when it's wrong. That's the actual project, and it's where most AI initiatives stall.
02
Adoption is built before the feature ships, not after. The team that has to use the AI — sales, customer success, the channel partner — needs to be in the room while it's being built, not in the demo afterward.
03
Hand-tuning is where margin goes to die. If every new client requires a week of taxonomy mapping, supplier categorisation, or chart-of-accounts wrangling, growth caps at the rate you can hire analysts. AI applied right makes that a one-hour configuration. AI applied wrong adds a chatbot nobody uses.
04
Where you put the AI matters more than which model you pick. AI sitting inside a deterministic pipeline tends to fail silently and break the audit trail. AI sitting above the pipeline — designing and adapting the deterministic steps the engine then runs — keeps the audit trail clean and the failure modes legible.
05
Translation between layers is the real job. Between technical and non-technical, between platform team and customer team, between what the LLM can do and what your product needs it to do. That's where I sit.
How we can work together — two rungs
Entry · pay-for-clarity
AI Leverage Read
$2,500 fixed · 1 week
I sit with your operations, product, and customer flow. I name the three places AI compounds in your business, and the three where it would add noise. You get a written read (4–6 pages), a recommended 90-day move, and a pricing-anchored plan if you want to continue. Three structured calls; I do the rest async.
1 day/week embedded + async between days. I work alongside your team — picking up AI problems as they surface, building or coaching your engineers to, translating between technical and non-technical, and making sure what ships actually lands with the customer. Quarterly review of what compounded vs what didn't.
Read credits: every dollar you spend on the AI Leverage Read counts toward your first month of Fractional AI Lead if you start within 90 days.
What this isn't
Not
a vendor pitch. I'm not selling carbon software or a wrapper around Claude.
Not
a Sam-solo sprint. I work alongside your team because that's what makes the work stick.
Not
strategy-only. I write code. I open PRs.
Not
enterprise-agency priced. Small roster, real work, no SOC-2 sales team.
Proof — what I've built
Carbon Tracker
11-agent prototype for a UK carbon consultancy: PostgreSQL + Climatiq + RAG + audit trail. The most architecturally sophisticated build I've shipped.
Derwen
Live AI biodiversity assistant. n8n + Perplexity + GBIF + OpenAI. Live at derwenai.replit.app.
SEAF data pipeline
Three-tier data warehouse for environmental impact assessments at SEAF / UWA, built with Claude Code as a development collaborator. Systems-thinking + facilitating scientists rather than replacing them.
Turing DSG '22
10-person AI sprint at the Alan Turing Institute, delivering in a week for the UK government's marine science agency.
Background. PhD environmental & ecological data science (USYD, DARE ARC). MPhys Physics & Mathematics, First Class (Manchester). Seven years across the European Space Agency, the Alan Turing Institute, UK universities, and a UK carbon consultancy — always as the translator between technical depth and the people who need to use it.