Salesforce MVP Hall of Fame (Shivanath Devinarayanan) · Founded 2022
One AI use case. In production. With the controls on.
Most pilots fail at production because nobody planned data access, identity, or governance. We assemble the stack from the process, then ship one seat with those controls attached.
Founded 2022 · Founder in the Salesforce MVP Hall of Fame · Dubai HQ · Kochi studio
The work
What we do in the org.
You have a use case, or a pilot that never left the sandbox. You need a production seat, not another workshop. The runtime might be Agentforce. It might be a custom stack. We pick from the process.
Pilot or production
A pilot, and a seat.
A pilot
A production seat
- Demo data
- The records the desk really uses, with sharing applied
- A shared service key
- On-behalf-of identity per caller
- Nobody reviews the transcripts
- A person reviews the exceptions each week
- No limit on spend
- A consumption limit per desk
- Owned by the innovation team
- Owned by the person who lives with it
Sequence
Readiness, stack, then one seat.
Free
The assessment
A short read of the org and a written note. No charge.
Week 1
Readiness
What the agent may see and whether that data is clean enough to act on. AED/SAR 14,900, credited if you sign the build within 60 days.
After week 1
Stack
Model, runtime, memory, identity, connector. Picked from the use case.
Shadow
Run beside a person
The agent proposes, a person decides, and you compare.
Production
One seat
Writes behind a human gate. Then we stay for the exceptions.
The honest no
When we are the wrong partner.
The honest no: if the work is a software factory with a named delivery date, we are the wrong partner.
Pilot to seat
Pilots we take to production.
Example
Customer service director at a Dubai property developer
Today: The pilot answered owner questions on demo data. Nobody has signed off a live write.
After: One seat on the real records, with sharing applied and a person reviewing the exceptions each week.
AgentforceExample
COO at a Riyadh logistics company
Today: The job lives in WhatsApp and a warehouse system, not in Salesforce.
After: A custom runtime picked from the process, with on-behalf-of identity for each caller.
For COOsExample
CIO at a US regional insurer
Today: Three pilots run on one shared service key, and nobody set a spend limit.
After: One goes to production with a consumption limit per desk and an owner who lives with it.
Financial services
Questions
What we actually say.
- Which models and runtimes do you use?
- Claude, Gemini, OpenAI, AWS Bedrock, Google ADK, CrewAI, n8n, Retell, Neo4j, Cognee, IndyKite, MCP, MuleSoft Agent Fabric, and Agentforce. We pick from the use case. See /ai-tools.
- Is this the same as setting up an agent?
- Narrower. If you can already name the job a person is doing like a robot, start there and we set up that one agent. Come here when you need the production path around it: the readiness check, the runtime choice, the governance, and the seats that follow the first one.
Before you pick a partner
Read these first. Then call anyone.
Our buyer guides. They apply whether or not you hire us.
Choosing an AI partner in the UAE
Five kinds of partner compared, the PDPL, DIFC, Arabic and WhatsApp questions, and red flags.
Choosing an AI implementation partner
Evaluation criteria, deployment models, governance fluency, and red flags.
AI partner evaluation framework
A five-part scorecard: depth, governance, integration, security, delivery.
Questions to ask an AI partner
Architecture, security, governance, pricing, and who is accountable.
AI partner vs in-house team
Hiring timelines, real costs, and dependency risk, scenario by scenario.
AI pilot to production
Which of the four ways your pilot stalled, and the sequence that moves it.
The brief
Send the situation.
One email. The situation travels with it. The assessment is free: a short read of the org and a written note. If you go ahead, the first build week is a fixed AED/SAR 14,900 (about USD 4,050), credited in full when you sign the build within 60 days.