Boston, MA — 2026-05-18 — Scale Growth Advisors LLC (scale-growth.ai), the Global Exclusive Go-to-Market Partner for IDESKS ONLINE AI, today announced the commercial availability of SARAH AI Suite, a sovereign, on-premises voice and reasoning AI platform engineered for regulated enterprises and small-practice operators. The platform spans two product tiers — SARAH Spark 2 and SARAH Enterprise on NVIDIA DGX B300 — and is now available to qualified customers in banking, healthcare, aviation, hospitality, government, travel, contact-centre operations, restaurants, dental practices, veterinary practices, and multi-specialty clinics.
The platform combines SOPHIA, a C-Level reasoning brain operating across an 34,792,085-feature connector surface, with SARAH, a voice conversational AI runtime, super-smart reasoning engine, orchestrator, and operational doer. Both roles operate on a single appliance located within the customer’s controlled environment. The architecture removes the multi-tenant cloud-LLM dependency that has, until now, blocked production-grade conversational AI deployment in tier-1 banks, federal and state agencies, healthcare systems, airline operations centres, and multi-property hospitality groups.
The market for production conversational AI has, for the prior three years, been dominated by three multi-tenant cloud-LLM providers operating per-token, per-minute, and per-egress meters against customer data hosted in vendor-controlled jurisdictions. That commercial and legal posture has prevented a regulated enterprise from owning the model weights, controlling the data plane, or producing a regulator-grade audit log for the conversations conducted on its behalf.
SARAH AI Suite is positioned as the sovereign alternative. The platform is delivered as a capital appliance — depreciable on the customer balance sheet, located within the customer’s controlled environment, with customer-held encryption keys, no third-party model API calls, and a complete audit trail covering every call, prompt, and output. The thesis is straightforward: enterprises operating in regulated industries should own their conversational AI infrastructure, not rent it.
“The conversation we keep having with chief risk officers, general counsel, and audit committees is the same one. Multi-tenant cloud AI cannot give a tier-1 bank or a federal agency the ownership posture the regulator expects. SARAH AI Suite was engineered to be the answer to that conversation — sovereign compute, owned model weights, customer-held keys, and a 30-day implementation schedule. The bank, the hospital, the agency owns the asset; we provide the platform.”
Chris Ismail · Founder, IDESKS ONLINE AI
SARAH AI Suite is delivered in three commercial tiers, each running the identical SOPHIA reasoning brain and SARAH voice runtime, scaled to the customer’s concurrency requirement and budget envelope.
Both tiers include the full SOPHIA + SARAH software stack, the 34,792,085-feature connector surface, the 36-month Global Replacement Warranty, maintenance and support, and onboarding for the customer’s engineering team. Commercial terms are confirmed during the discovery-call process and reflected in the customer’s engagement contract.
The Enterprise tier operates on the NVIDIA DGX B300 (Grace Blackwell Ultra) compute substrate. Each appliance contains eight NVIDIA B300 superchips connected via fifth-generation NVLink at 130 TB/s aggregate, 2,304 GB of unified HBM3e memory at 64 TB/s aggregate bandwidth, 10,000 TFLOPS FP16 / 2,000 TFLOPS FP8 / 1,400 TFLOPS FP4 per superchip, and direct-to-chip liquid cooling at 2,700-2,800 W TDP. The appliance is rack-mounted, occupies approximately 13 RU, and is shipped with a built-in carrier-grade router (2 × 200 Gbps WAN backhaul, 16 × 100 Gbps LAN) for federation across multi-site deployments via the Private Enterprise IP Network (PEIPN).
The Spark 2 tier is a desktop-class voice AI appliance designed for solo practitioners and single-location operators. The same SOPHIA brain and SARAH voice runtime that power the Enterprise NVIDIA DGX B300, sized for two concurrent voice calls. Form factor fits under a front-desk counter or on a standard 19″ rack. Standard 120 V or 240 V outlet, single network drop, acoustic profile suitable for a clinical or office environment.
SARAH AI Suite is engineered around the SOPHIA Layer-3 API HUB — a unified data and connector fabric currently progressing through the 100,000-API Marathon. As of 2026-05-18, the platform operates against an 34,792,085-feature connector surface comprising 512,653 categories (cats) and 11.74 million live connector instances spanning every major enterprise SaaS, payments, communications, EHR, GDS, PMS, POS, contact-centre, banking-core, and government-services platform. Post-100K-API Marathon, the platform will consolidate into a single unified Postgres database governed by one schema and a single source of truth — the architectural commitment Scale Growth Advisors LLC characterises as “one DB runs the entire world.”
“The institutional buyers we are working with — banks, federal agencies, large health systems, airlines, hotel groups — have been waiting for an AI platform they can actually authorise to deploy. The answer to the question ‘who controls the data’ cannot be ‘a third-party AI vendor.’ It has to be ‘we do.’ SARAH AI Suite is the platform that makes that answer possible. We are taking it to market across thirteen industries on day one.”
Mike Farrell · Partner, Scale Growth Advisors LLC
SARAH AI Suite is engineered to satisfy the compliance frameworks that govern its target industries. Sovereign on-premises deployment removes the multi-tenant data-residency disclosure obligation across every framework below, and the platform ships with the documentation package required for the relevant Authorising Official, Regulator, or Accreditation Body.
SARAH AI Suite is available across twelve industry-specific landing pages, structured into two commercial groups by buyer profile.
Sovereign on-premises voice + reasoning AI for tier-1 / tier-2 institutional buyers, deployed on NVIDIA DGX B300. Available at scale-growth.ai/edms/:
An additional Australian Banks page is available at idesksonline.ai/edms/banks/, covering CBA, Westpac, NAB, ANZ, Macquarie, and 17 other AU institutions with APRA / ASIC / AUSTRAC alignment.
Sovereign on-premises voice AI for solo practitioners and single-location operators, deployed on the compact SARAH Spark 2 appliance. Available at scale-growth.ai/spark2-edms/:
SARAH AI Suite is delivered with a 30-day implementation schedule from contract execution to first live voice agent, supported by Scale Growth Advisors’ engineering team. The 36-month Global Replacement Warranty operates from staging depots in Boston, Frankfurt, Sydney, and Singapore, with a 48-hour appliance-swap commitment in the event of hardware or software failure. Maintenance and support are included for the full 36-month financing term across all three tiers. The customer’s operations team assumes day-to-day responsibility from Day 90, with 24 × 7 on-call engineering support from Scale Growth Advisors LLC for the duration of the engagement.
SARAH AI Suite is commercially available worldwide as of 2026-05-18. Commercial terms, technical architecture review, and customer-specific compliance documentation are confirmed through a 60-minute discovery call scheduled directly with Scale Growth Advisors LLC.
Forward-Looking Statements. This press release contains forward-looking statements regarding Scale Growth Advisors LLC, IDESKS ONLINE AI, and the SARAH AI Suite platform, including statements regarding product availability, the SOPHIA Layer-3 API HUB consolidation, the 100,000-API Marathon, and implementation schedules. These statements are based on management’s current expectations and assumptions and involve risks and uncertainties, including hardware supply availability, regulatory approval timelines, and customer deployment schedules. Actual results may differ materially. The 34,792,085 feature count reflects the platform’s connector surface as of the date of this release and is subject to change as the platform evolves.
Trademarks. NVIDIA, DGX, Blackwell, Grace Blackwell, NVLink, and HBM3e are trademarks of NVIDIA Corporation. SARAH AI Suite, SOPHIA, SARAH, SARAH Spark 2, SARAH Enterprise, and SOPHIA Layer-3 API HUB are trademarks of IDESKS ONLINE AI. All other product, vendor, and platform names referenced in this release are the property of their respective owners and are used here for descriptive identification of integration capability only; no endorsement or affiliation is implied.
Document control. Press Release · 2026-05-18 · Boston, Massachusetts · Scale Growth Advisors LLC.