01AI agents and assistants06 items
Inside the system, not beside it.
Intake and qualification, ticket triage, internal question answering over your own records, scheduling, escalation routing, drafting with a human approval step.
AISTORSServices
ServicesTen groups, named plainly, so you can tell what is included.
If a group is not on this list, we do not do it.
Applied AI and cloud engineering, delivered inside your own accounts. Every group below starts with the same diagnostic, records a baseline before anything is built, and hands over documentation your team can rebuild from. Each page says who the work is for, what it fixes, and who it is not for.
Book a 30-minute call
Where every engagement starts
One process, measured, before anything is designed. The diagnostic can end with a written recommendation not to build, and when that is the honest answer it is the cheapest one.
01
Every engagement starts here
Before anyone writes code, we establish whether AI is the right answer and what it is worth.
The AI Readiness and Baseline Diagnostic.
It exists because the most expensive AI project is the one that should never have started. Three to five days, or a single day for a solo operator or small business. You leave with a measured baseline of the target process, three prioritised opportunities, a data readiness verdict with named integration points, an implementation plan, risk and compliance flags, and a written go or no-go recommendation.
If the recommendation is no, you keep the report, the baseline method and the plan, and there is no further commitment. That happens, and we would rather it happened at this stage.
02
The ten groups
If a group is not on this list, we do not do it. That is the point of publishing the list.
02Document and process automation06 items
The procedure already exists. It just isn't running itself.
Document processing, CRM hygiene, quoting, reconciliation, ticketing, month-end close.
03Data and retrieval for AI04 items
Answers from your own records.
Enterprise search, retrieval grounded in your documents, pipelines, private and self-hosted models where data cannot leave your estate.
04Forecasting and prediction07 items
Only where the history is long enough to test against.
Demand forecasting, inventory and replenishment, pricing support, churn, fraud and anomaly detection, predictive maintenance, computer vision.
05Cloud architecture and migration07 items
Built once, documented, handed over.
Landing zones, migration waves, Kubernetes, infrastructure as code, observability, site reliability engineering, cloud security.
06Cloud cost and FinOps04 items
Rightsize first. Commit second.
Rightsizing, commitment coverage, storage lifecycle, AI token and GPU attribution so inference spend reaches the team that caused it.
07Workload placement and repatriation05 items
Move the few that are cheaper. Leave the rest alone.
Placement analysis, selective repatriation, hybrid design, sovereignty architecture, exit cost and lock-in analysis.
08AI governance and compliance06 items
Written down before anyone asks for it.
EU AI Act readiness, ISO 42001 readiness, HIPAA, SOC 2, GDPR and DPDP programme support, red teaming of deployed systems.
09Managed AI and cloud operations04 items
The same four measures, every month.
Drift detection, model upgrade handling, escalation, compliance reporting against the baseline recorded in the Diagnostic.
10Enablement and handover05 items
You should not need us in year two.
Runbooks, infrastructure as code handover, evaluation practice for your team, internal AI usage policy, working sessions with the eventual owners.
03
What is true of all ten
These are commitments rather than preferences, and they are the same on every engagement on this page.
On every engagement
- Built in your accounts and your tenancy. Nothing on this list requires a platform of ours to sit in the middle, and nothing on it is billed as a licence.
- You own what we build. Code, configuration, prompts, evaluation suites and documentation, assigned to you on final payment.
- Measured before and after. A baseline is recorded before the build so the result can be argued with rather than asserted.
- Read-only until you approve otherwise. No production change without your agreement and a recorded rollback path.
- The person who scopes it builds it. Founder-led delivery, no handoff to a team you have not met.
- Exit designed in from the start. Runbooks and infrastructure as code are written during delivery, so leaving is an administrative act.
What we will not do
- No reseller margin. We hold no reseller agreement or commission with any cloud or model provider, so there is nothing behind a recommendation except the reasoning.
- No proof we cannot show you. Every figure on this site is attributed to a named, dated source, and where we hold no case study we say so rather than imply one.
- No accuracy figure before seeing your data. Any number quoted earlier than that is marketing.
- No autonomous action by default. What a system may decide alone is agreed in writing and becomes a contract term.
- No work we cannot staff properly. We will tell you what we can absorb, and refuse what we cannot.
- No silent scope change. Fixed scope means fixed scope, and a change is re-scoped in writing before any work proceeds.
04
Working with us from another country
Where you are is not the constraint. The hours we can overlap with you are, and we state them rather than imply them.
Nothing of ours sits in the middle.
Work is built and run inside your own tenancy, in your own region, on your own accounts and your own model provider contracts. Where your data cannot leave a jurisdiction, open-weight models run fully self-hosted in your estate. That is what makes distance workable: there is no platform of ours you have to send data to, and no region of ours you have to accept.
It also means the exit is the same wherever you are. Code, configuration and runbooks live in your repository from the first week, so ending the engagement is an administrative act rather than a migration.
01
India, the Middle East, the United Kingdom and Europe
A working-day overlap without anyone keeping unsocial hours, which means live design sessions, same-day decisions and a normal escalation path. This is the widest overlap we have and the easiest to commit to in writing.
02
United States, eastern and central time
An afternoon and evening overlap from our side, which is sustainable and is stated in the contract rather than left to goodwill. Design and review sessions are scheduled inside it; build work continues outside it.
03
United States western time, Australia and New Zealand
There is no honest natural overlap here. These engagements are run asynchronously with a named associate for cover, and we will tell you plainly if we cannot staff that rather than promise a window we would miss.
What makes this work
- Your tenancy, your region. Everything is built in your accounts. We do not move your data to infrastructure of ours, because there is none.
- Residency treated as a requirement. Not a preference. Where rules say data cannot leave a jurisdiction, the architecture satisfies that before cost is discussed.
- Your contracts with model providers. Builds run on your agreements and your enterprise tiers. We hold no reseller relationship with any provider.
- Sub-processors named before work starts. Every third party in the path, with a standing right to refuse any of them.
- Written in your language of record. Runbooks, decision records and reports in English, structured for an assessor or an auditor rather than for us.
- One accountable person. Founder-led. The person who scopes the work builds it, so nothing is lost across a timezone in a handover.
What we will not claim
- No round-the-clock cover by default. A founder-led practice cannot honestly offer it as standard, and we would rather say so than discover it during an incident.
- No follow-the-sun team. There is no second shift. Where cover beyond our stated hours is genuinely needed, it is scoped with vetted associates who are disclosed to you individually.
- No overlap we cannot sustain. We will not agree to a window that depends on one person never taking leave.
- No on-site presence as standard. Travel is scoped separately where it is genuinely required, and most of this work does not require it.
- No legal or tax advice in your jurisdiction. We describe what the system does and where it runs. Whether an obligation attaches to you is a question for your own advisers.
- No engagement we cannot staff. If taking your work would mean doing it badly, we will tell you before you commit rather than after.
Contracting entity ‹FILL: legal entity name and jurisdiction›. Invoicing currency ‹FILL: USD for international engagements, INR for India›. Coverage hours, response commitments and escalation paths are stated per engagement rather than published as a single figure, because they differ by tier.
05
Related work
Disclosure
AI capability and data handling
Which model providers we use, on what terms, with training on your data disabled and every sub-processor named before you sign.
Service 08AI governance and compliance
Where the audit trail, model documentation and incident playbook are built as evidence rather than as a policy pack.
Service 10Enablement and handover
How the work leaves our hands and stays working in yours, tested by your team operating it while we watch.