Tan Gravam

Tan Gravam

Enterprise AI practitioner, systems architect and product builder.

I work at the intersection of AI transformation, enterprise systems and workflow design. I spent 18 years inside large organizations building the systems that run cash, treasury and the financial close — the kind of complex, controlled, exception-heavy work that AI is now supposed to change. Gravam is where I turn that experience into practical thinking, tools and software for the AI-native enterprise.

The through-line is one idea: complex enterprise work, redesigned with AI. Not AI bolted onto a process whose shape was set by paper and ERP screens, but the harder question of what the work should look like now — what stays deterministic, what a model can genuinely judge, where a named human still decides, and what the architecture around all of it has to guarantee. Treasury, SAP and finance systems are where I learned to answer that, and they remain the deepest evidence on this site rather than the boundary of it.

What I work on

  • AI transformation — where AI creates real business value, how a workflow gets redesigned rather than accelerated, and how you tell whether anything actually moved.
  • Workflow design — taking a complex process apart: the deterministic core, what AI does well, the decisions that need a name, and the exception path where enterprise work really lives.
  • Enterprise AI systems — agents, integration with the systems of record, data, controls, observability and human oversight.

If that's a problem you're facing, here's how we could work together.

Where the expertise comes from

The edge isn't that I can build apps — plenty of people can. It's the 18 years behind them:

  • SAP Treasury & Cash Management — TRM, One Exposure from Operations, liquidity items, bank statement integration and BCM.
  • Corporate treasury architecture — how ERP, TMS, banks and market data fit together, and which system should own which data.
  • Finance-systems delivery — turning vague demands into clear delivery decisions: intake, scope, UAT, cutover and governance.

Writing

I publish practical, first-hand notes in two bands. The enterprise AI band is the thinking: transformation, workflow design and systems. The field-expertise band is the evidence underneath it: treasury management systems, SAP treasury, architecture and finance-systems delivery. If that's your world, the newsletter sends one practical idea at a time, when there is one worth sending.

What I'm building

Delivery Sheet is this method as software — turning vague leadership asks into clear, reviewable delivery decisions before a team commits. Workomap maps the SAP world I come from — a structured, city-by-city map of SAP professionals. Alongside them I build outside the enterprise too: Yollardayız (live border-crossing waits for the sıla yolu drive) and Listemizde (local businesses ranked by community recommendation, not payment). Each carries an honest status on the products page.

How I build

Small, focused products from real operational problems — shipped fast with a solo stack (Nuxt, Supabase) and AI in the workflow, but never letting AI make the product decisions. I build in public and write down what works, what I killed, and why.

How this site's content is made

The writing here is expert-led and first-hand — but I'm transparent about the workflow behind it:

  • AI-assisted, human-reviewed. I draft with AI in the loop, but every published page is reviewed by me before it goes live. A review gate keeps anything unreviewed out of the sitemap, topic hubs and search index — draft work stays draft.
  • Sourced where it's technical. SAP and standards-based articles cite verified primary sources (SAP Help, SWIFT, the Basel Committee) and carry an explicit release-scope note, because behaviour differs by version, edition and channel — so verify against the docs for your system.
  • No invented specifics. I don't fabricate client names, first-hand anecdotes or metrics. Where a number is used to illustrate a method, it's labelled as illustrative, not a real figure.

What I write about — and what I don't

I write from first-hand experience in SAP, treasury, finance systems, enterprise architecture and complex systems delivery. The Enterprise AI work applies those lessons to workflow redesign, controls and architecture — it is reasoning from that experience, not client AI-transformation case studies I have not run. I don't give personalized investment or financial advice, I don't endorse specific vendors, and I try hard to stay inside what I've actually done rather than dressing up second-hand knowledge as expertise.

Start here

The reading path, in order — the thinking first, then the domain depth it rests on:

Elsewhere

If the work above is the work you have, here is how we could work together — or just tell me about the workflow.