Who is Periscope a fit for?
Any team that has to get AI into production, from a fifty-person company automating its back office to an enterprise standing up an AI lab. What they share: real systems, security obligations, and no appetite for a two-year program. Software companies and traditional businesses alike, across healthcare, financial services, manufacturing and automotive, retail, energy and professional services. If you need one freelancer or a body shop by the hour, we are the wrong shape, and we will say so.
We are a VC or PE firm. How do you work with portfolio companies?
Company by company, starting with a fixed-fee Value-Creation Diagnostic, typically two weeks, that returns the replatform cost and benefit, an AI product roadmap, a go-to-market plan built around forward-deployed engineers, a security gap list and a 12-month ROI model, ending in a go / no-go. Execution runs as milestone-priced pods against that model, with a board-ready report each month. Across a portfolio we work on one rate card, share the AI Foundation platform, security monitoring and pods between companies, and can tie a portion of fees to agreed outcomes where that aligns everyone. The point is to give a company with a real niche the senior AI, platform and security capacity it cannot hire in time, instead of a write-off.
How do you price work?
Every engagement starts fixed-scope, priced before it begins. The starting prices on this page cover typical scope; we confirm a fixed quote after a short scoping call, usually within two business days. Phased builds are quoted phase by phase, with a go / no-go decision at each boundary, so you are never committed further than the phase you are in.
Who actually does the work?
A senior engineer based in the US leads every engagement and stays on it through production. Delivery is shared with our Mumbai team, full-time Periscope employees on the same tooling, standards and security controls, not contractors found for the project. You always know who is on your team, and the people who scoped the work are the people who build it.
How do you handle security and compliance?
Periscope operates under SOC 2 Type II controls, audited annually, and our teams have delivered under HIPAA for years. Client data stays inside agreed boundaries. For sensitive workloads we design private deployments, VPC or on-premises, including NVIDIA DGX systems, so data never leaves your environment, and every agent we ship carries approval gates and an audit log by default.
Can you run AI on-prem, in a private cloud, or hybrid?
Yes, that is the point of the infrastructure specialization. We design and operate AI stacks on NVIDIA DGX-class systems on-premises, in private clouds, on Azure, Google Cloud and AWS, and in hybrid arrangements where training, inference and data each live where they should. Sovereignty is a routing decision on one platform, not a separate project, and the same eval, monitoring and cost controls apply wherever the model runs.
Our data isn't ready. Is that a blocker?
Almost never, it is the normal starting condition. We scope readiness to the workload in front of us: which sources it needs, how clean and how current they must be, who may access them and how the results are retained. That pass is the first week of every Sprint Zero, and the pipelines, retrieval stores and evaluation sets we build for the first workload are designed to be reused by the next ones. You do not need a data lake, a catalogue program or a new platform before the first agent ships; you need the data for one workload to be trustworthy, and a plan for the rest that grows with demand.
Can you migrate us off Adobe Customer Journey Analytics?
Yes. We have moved a digital health platform's Adobe Experience Platform event data off CJA into a HIPAA-compliant, Postgres-based stack feeding Power BI, and we maintain a four-phase playbook, inventory of the reports, segments and data views you actually use; a parallel run; cutover; license retirement, with Piwik PRO as an alternative target where a packaged, privacy-first analytics product fits better. The free CJA exit assessment gives you the inventory, a target architecture, a plan and a cost comparison before you decide anything.
Do you train our teams as well as build?
Yes. DevSecOps Enablement and AI enablement engagements pair implementation with hands-on training, so the pipelines, controls and agent tooling we put in place are run by your engineers afterwards. The goal is capability that stays, not a dependency.
Can you get our pipelines and cloud ready for agents?
Yes, it is one of the most common asks now. Agents multiply the number of things that ship code and call APIs: internal teams, agent runtimes, third-party systems, SaaS tools, open-source components and whatever shadow IT has adopted. We review and audit the code (human- and agent-written), build pipelines that both humans and agents move through across every environment, size the cloud workloads for the traffic that creates, and put an inventory and a policy around the endpoints so growth is governed rather than discovered. It is scoped per environment, in fixed-fee phases, and it is the same work we run for platforms we operate.
How do you protect API keys, tokens and agent credentials at scale?
Treat them as what they are: non-human identities, and soon the majority of identities you have. An enterprise running thousands of agents across hundreds of models holds far more tokens, keys, service accounts and agent credentials than people, many long-lived, over-scoped and never rotated. We start with a read-only exposure scan across repos, CI, cloud and SaaS (age, scope, last use, where each leaks), then rotate and scope down, move workloads to short-lived credentials and workload identity federation, put secret scanning and supply-chain controls, signed commits, protected branches, artifact attestation, into the pipelines, and add detection for abused tokens. We have run this after a live GitHub supply-chain attack, not only as a checklist.
Do you offer managed security monitoring?
Yes. As a WatchGuard Gold Partner we run 24×7 security monitoring, threat detection and response, and compliance reporting for SOC 2, HIPAA and PCI, either as a standalone managed service or as part of operating the platforms and agents we build. It starts with a free, read-only security posture review of your environment; the service is then priced per environment and reported monthly, alongside the same evidence report our operated systems carry.
Which clouds and AI models do you work with?
Azure, Google Cloud, AWS, private cloud and on-prem. On models we are deliberately agnostic: OpenAI, Anthropic, Google and open-weight models on private infrastructure, routed per workload for accuracy, cost and data-boundary requirements, never for a partner rebate. Token spend is measured and managed from day one, and we will tell you when a simpler, non-AI fix is the better answer.
How quickly can we start?
Free reviews and audits are usually scheduled within a week. Sprint Zero typically starts within two weeks of a signed scope, depending on team availability. We run a small number of sprints and pilots at a time and will confirm the next available start when we send the quote.
