jamesrichman.space
Index
Ideas, experiments, future projects
Base
Singapore
On the board
4 open questions
Updated
October 2026

Entry 00Status: open · Kill criteria: written

Some ideas deserve room to breathe.

James Richman in a light shirt, looking at the camera, under an arc of light on a pale wall.
November 2025. Black and white, with the light on the wall printed in lime.

A digital playground for concepts that are too early, too strange, or too experimental for a polished company site.

The earliest stage of an idea is where it is easiest to kill and easiest to distort. This space exists before the branding, before the pitch deck, and before anyone has decided whether the thing is sensible. It is for prototypes, mental models, odd interfaces, speculative systems, and questions that need room.

Kept by James Richman · CEO, OTLEN · Singapore

  1. 0Question
  2. 1Sketch
  3. 2Prototype
  4. 3Project
0 · Pixel blocks
The shape is a guess.
1 · Characters
It can be written down.
2 · Dither
It can be tested.
3 · Solid
It has graduated.

Rules of the space · 5

Five rules keep the room useful.

01Rule

Unfinished ideas

An unfinished idea can be more valuable than a polished opinion because its assumptions are visible. Concepts here are allowed to be incomplete. The objective is to make them tangible enough that reality can respond.

Example idea · stage 04 assumptions · 0 tested

A multi-location clinic group loses revenue when packages sold at the front desk never reach the treatment ledger.

  1. A1The leak grows with every location that opens.Untested
  2. A2At group scale the amount matters to the P&L, beyond rounding.Untested
  3. A3The sale and the treatment are recorded in different systems.Untested
  4. A4No one reconciles the two every day.Untested
Written this way, each boxed phrase is a test someone can run. Example idea, for illustration.

02Rule

Interfaces as thinking tools

Sometimes the fastest way to understand an idea is to build its interface. A clickable mockup can reveal a workflow problem that twenty paragraphs would hide. Prototyping is implementation, and it is also a method of reasoning.

Spec · 20 paragraphsPrior auth

  1. ¶01Scope: outpatient imaging orders that need payer approval.
  2. ¶02The ordering clinician signs the order in the EHR.
  3. ¶03Staff confirm the patient’s active coverage.
  4. ¶04Staff confirm whether the payer requires authorization for the procedure code.
  5. ¶05Requirements vary by payer and by plan.
  6. ¶06Where authorization is required, a request is created.
  7. ¶07The request goes through the payer’s preferred channel.
  8. ¶08Patient demographics are included.
  9. ¶09The member ID is included.
  10. ¶10The ordering provider’s NPI is included.
  11. ¶11Procedure and diagnosis codes are included.
  12. ¶12Site of service is included.
  13. ¶13The request is checked for completeness.
  14. ¶14Clinical documentation supporting medical necessity is attached.
  15. ¶15The request is submitted.
  16. ¶16Staff monitor the request for a decision.
  17. ¶17If more information is requested, staff respond.
  18. ¶18Approvals are recorded against the order.
  19. ¶19Denials are routed for appeal review.
  20. ¶20The patient is scheduled once approval is recorded.

Mockup · MRI prior authorizationStep 1 of 5

Screens switched
0
Fields typed twice
0

Example workflow; the counts are illustrative. In the written spec the problem hides in ¶14, a sentence that reads as routine. In the mockup it appears at step 4, where the notes have to leave the EHR.

03Rule

Strange combinations

The interesting ideas often sit between categories: AI plus operations, healthcare plus automation, content plus software, research plus product design. This space is intentionally permissive about crossing those boundaries.

Crossings between four fields and four practices. Four named crossings are marked, each linked to an open question on the board.
OperationsAutomationSoftwareProduct design
AISP-02On the board+++
Healthcare+SP-01On the board++
Content++SP-04On the board+
Research+++SP-03On the board
The four crossings named above fall on the diagonal. Each one carries an open question on the board below. The other twelve cells are open.

04Rule

Kill criteria

Every experiment should have a reason it might die. No demand. Too expensive. Too complex. Bad user behavior. Weak economics. If the evidence says stop, stopping is a successful experiment.

Tick any criterion the evidence meets

Experiment status

Running

No criterion has tripped. Keep collecting evidence.

05Rule

From space to reality

When an idea becomes useful enough, it graduates. It may become an OTLEN project, a personal tool, an article, or something completely different. The point of the space is to make that transition easier.

SpaceAn idea that passed its tests
  • An OTLEN projectBuilt and run by the company, inside revenue operations for healthcare.
  • A personal toolKept for one job, used until it stops earning its place.
  • An articleArgued in public. The peptides evidence panel on jamesrichman.health began as SP-03.
  • Something completely differentThe idea decides.
Four exits, one door. An idea leaves the space only through one of them, or through a kill criterion.
James Richman seated by a window, holding a tablet in one hand and explaining a point with the other.
In conversation, 2026. Photograph via London Daily News.

Who keeps the board

Kept from Singapore, beside the work at OTLEN.

James Richman leads OTLEN, which builds AI agents for revenue operations in healthcare. Its team and clients work across London, New York, Singapore, San Francisco and Middle East capitals. This board runs beside that work. An idea starts here when it is too early for the company, and some ideas end up inside it.

Role
CEO, OTLEN
Base
Singapore
OTLEN founded
2008
Full profile
jamesrichman.name

On the board · 4

Open questions, each with a reason it might die.

Each question sits on one of the four named crossings. Each carries the public fact that makes it timely, the criterion that would end it, and where it could go if it survives. Three are still questions. One has a first written pass.

SP-01Healthcare × AutomationStage 0

72hours for an expedited prior authorization decision

What does prior authorization look like when both sides answer by API?

Why now
Under CMS rule CMS-0057-F, from 1 January 2026 impacted payers send decisions within 72 hours for expedited requests and seven calendar days for standard ones. A Prior Authorization API is required from 1 January 2027. [1]
Kill if
Bad user behavior: the APIs ship and offices keep faxing.
Could become
An OTLEN project

SP-02AI × OperationsStage 0

2022negotiated rates became public files

Can every payment be checked against the rate the payer agreed to, automatically?

Why now
CMS has enforced the Transparency in Coverage rule since 1 July 2022. Most group health plans and insurers now post their pricing, including negotiated in-network rates, as public machine-readable files. [2]
Kill if
Too complex: the files cannot be matched to a contract line reliably.
Could become
An OTLEN project

SP-03Research × Product designStage 1

20%fewer major cardiovascular events, semaglutide vs placebo

What would a plain evidence ledger for peptides look like: trial stage, endpoint, regulator status?

Why now
The SELECT trial gave a peptide drug cardiovascular outcome data in adults with obesity and no diabetes (hazard ratio 0.80). [3] The FDA lists BPC-157 among compounding substances that may present significant safety risks. [4] The spread of evidence is wide, and headlines flatten it.
Kill if
Too expensive: it cannot stay current without a full-time editor.
Could become
An article. A first pass is the peptides panel on jamesrichman.health.

SP-04Content × SoftwareStage 0

1,614AI-enabled medical devices on the FDA list

Can an article about clinical AI be operated as well as read?

Why now
The FDA’s public list of AI-enabled medical devices holds 1,614 entries, latest decision 29 June 2026. [5] A list that long is a dataset, and a dataset can carry an interface: filter by specialty, by year, by clearance path.
Kill if
No demand: readers want the summary and skip the tool.
Could become
An article with a tool inside it

Before an idea becomes a business, it deserves enough room to become a real idea.

This space is a starting point. It grows with original articles, prototypes, research notes, tools and case studies as each idea earns them.

Sources · 5

Every fact on the board, with its source.

  1. CMS. CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F), fact sheet, 17 January 2024. cms.gov
  2. CMS. Health plan price transparency (Transparency in Coverage), enforcement from 1 July 2022. cms.gov
  3. Lincoff AM et al. Semaglutide and cardiovascular outcomes in obesity without diabetes (SELECT). N Engl J Med. 2023;389:2221–2232. doi.org/10.1056/NEJMoa2307563
  4. FDA. Certain bulk drug substances for use in compounding that may present significant safety risks (BPC-157 entry). fda.gov
  5. FDA. Artificial Intelligence-Enabled Medical Devices. 1,614 entries, latest decision 29 June 2026. fda.gov