Disclosure
Where this comes from
Everything a reader needs in order to judge how much weight to put on what I write here. The short version: I work for a platform vendor, I write about the layer their product sits in, and you should read me with that in front of you.
Employment and capacity
I am employed full time as a technical architect working on Agentforce and Data 360. My career has been in Salesforce platform and business architecture, including ten years at Accenture.
Everything on this site is written in a personal capacity, on my own time and equipment. The views are mine alone and are not those of my employer or of any client. Nothing here has been reviewed, approved, or requested by my employer, and nothing is written on anyone's behalf.
The obvious conflict, stated plainly
My professional experience, and my living, come from one vendor's platform. That creates a bias in both directions and it would be dishonest to claim otherwise.
Anything positive I write about Salesforce products should be discounted accordingly. Anything critical is written by someone with an obvious interest in not being critical, which is a different problem and arguably the more dangerous one — the pressure runs towards silence, not towards praise.
The same caution applies more widely now that this publication covers the whole field. Salesforce partners with some model providers and platform vendors and competes with others, so read what I write about any of them with that in mind.
My response to all of it is the same: keep the arguments at the level of architecture rather than product. A claim about how an agent should be governed, or what a new model capability changes, should hold whether you are on Salesforce, ServiceNow, Microsoft or a stack you built yourself. Where it stops holding, that is worth knowing, and I would rather a reader tell me than a vendor.
How claims are produced
Every technical claim is either rebuilt in a personal development environment, outside work, or rests on a public source that a reader can check and that the piece cites. If it can be neither, it does not publish. Diagrams are drawn fresh, never from work artefacts.
Analysis is marked as analysis. Where a piece argues a position rather than reports a finding, it says so, and it sets out the strongest reading against it.
No client is named. No employer system is described. No engagement metrics, revenue figures or customer outcomes appear anywhere on this site — they cannot be published without pointing at specific work. Where a scenario is assembled from several situations, the piece says so in the piece.
Independence
No sponsorship, no affiliate links, no paid placement, no vendor-reviewed content. No vendor has been given sight of a piece before publication and none will be. If that ever changes it will be disclosed here first and on the piece itself.
There is nothing for sale on this site and no advisory or consulting work on offer. That is a sequencing decision rather than a permanent one, and if it changes, this page changes with it.
Corrections
Errors are corrected in public, dated, with the original claim left visible and struck through. Pieces carry a revision log. Nothing is silently edited.
A technical argument you cannot watch the author revise is not an argument. If you think something here is wrong, the fastest route is to say so — corrections are published with attribution unless you ask otherwise.
How AI is used here
Drafting assistance and editing, yes. Diagrams, architecture claims and conclusions, no. Nothing publishes that I have not verified myself, and nothing publishes that I could not defend without the tool that helped write it.