Go.AI Company Spotlight: Private AI for Regulated Work

Go.AI builds private AI infrastructure for regulated and compliance-sensitive organizations. The Chicago company combines Go1 hardware with Go.OS software so banks, healthcare organizations, insurers, manufacturers, and defense teams can run models, index data, build applications, and retain audit records inside a customer-controlled environment.

Co-founders David Moscatelli, CEO, and Lisa Gillespie, COO, lead Go.AI. The company's current history traces the business to Abacus Analytics LLP in 2018, followed by a corporate conversion and product refocus in 2022 and the adoption of the Go.AI name in September 2026.

Go.AI matters now because enterprise AI is moving from experiments into workflows that carry legal, operational, and regulatory consequences. The market will not be decided only by model quality. It will also be decided by where a model runs, what data it touches, how usage is governed, and whether an institution can reconstruct the answer later.

About Go.AI

Go.AI began as the predecessor partnership Abacus Analytics LLP, founded by Moscatelli and Gillespie in 2018. The business became Go Abacus Corporation in 2022 as the team shifted toward enterprise software and AI infrastructure. It rebranded as Go.AI on September 1, 2026.

That timeline explains why public descriptions sometimes cite different founding dates. An older seed announcement described the company as founded in 2022, while the current company history identifies the 2018 predecessor. The clean reading is 2018 for the original partnership and 2022 for the corporate conversion and product refocus.

The company is headquartered at 111 South Wacker Drive in Chicago. Go.AI opened the office in July 2026 with initial capacity for 50 employees and now says its team has grown beyond 50 people.

How Go.AI's Private AI Stack Works

The Go1 appliance supplies local compute, while Go.OS manages models, data indexing, connected sources, applications, and audit records. Go.AI says a standard Go1 can support as many as 2,000 concurrent users and move from a network connection to production in about 15 minutes. Those performance specifications are vendor claims and have not been independently benchmarked in the public record reviewed for this Spotlight.

The architecture is local-first rather than absolute about every workload. On-premises and air-gapped deployments can keep sensitive data and inference inside an institution's boundary. Go.OS documentation also says customers may authorize selected requests to cloud models while maintaining the same audit chain. Updata Partners describes another option: operating the stack in a private-cloud instance.

That flexibility is the product thesis. A regulated buyer should be able to choose the model, deployment boundary, data path, and audit policy for each use case instead of accepting one public-cloud rule for every task.

Why the Deployment Boundary Matters

Enterprise AI becomes harder when a useful answer must survive scrutiny from security, compliance, model-risk, legal, IT, and an external examiner. Sensitive documents cannot wander casually through third-party systems. A valid model response can still create a governance problem when nobody can explain which data, model, prompt, or policy produced it.

Go.AI is turning that constraint into infrastructure. The company emphasizes local inference, customer-controlled data, fixed pricing, and auditability rather than a per-token public-cloud meter. Its site displays SOC 2 Type II, ISO 27001, and HIPAA labels. Those are company representations about the platform and controls, not proof that every customer deployment automatically satisfies every regulatory obligation.

The tradeoff is operational. Hardware must be installed, secured, updated, monitored, and supported across customer environments. Private infrastructure can reduce one class of cloud risk while creating supply-chain, lifecycle, integration, and field-support work that a software-only vendor does not carry.

Leadership, Customers, and Traction

Moscatelli and Gillespie remain the only co-founders identified on Go.AI's official leadership materials. The current About page lists Moscatelli as CEO and Gillespie as COO; it does not identify a CTO. That matters because executive lists drift quickly, and a credible Company Spotlight should not turn an old profile into a current title.

Go.AI reports more than 200 customers, 900,000 monthly users, and more than 12.5M queries processed each day across customer deployments. The company also reports annual recurring revenue growth above 8x year over year and continued profitability. Updata Partners repeated the ARR, profitability, and query figures in its investment announcement.

Those figures show meaningful operating momentum, but they remain company and investor claims rather than audited financial results. Go.AI has not disclosed revenue, the prior-year ARR base, retention, margins, customer concentration, contract values, or the number of installed appliances. Query volume measures activity, not the economic value or risk profile of each workload.

Funding and Market Position

Go.AI announced a $5M Seed round in November 2025 and an $85M Series A in September 2026, bringing disclosed funding to $90M. Updata Partners VII led the Series A, with GFT Ventures and LAUNCH returning. Updata General Partner Carter Griffin is joining the board.

The financing gives Go.AI room to expand engineering, develop Go.OS and the Go1 hardware family, grow client advisory and field capacity, and sell beyond its original financial-services base. It does not settle the competitive question. The company must prove that private AI can scale across institutions without turning each implementation into a long custom infrastructure project.

Go.AI sits at the intersection of enterprise AI, hardware, model operations, security, and regulatory technology. Its durable advantage will depend less on owning one model than on making many models governable inside environments where accountability is part of the buying decision.

Why Go.AI's Hiring Momentum Matters

Go.AI's current careers page shows openings across client services, field deployment, sales, marketing and design, and strategy and operations. That mix is more revealing than a generic growth claim. The company needs people who can turn hardware, software, controls, training, and customer policy into working production systems.

Field and client roles reflect the burden of local infrastructure. Sales and advisory roles reflect a buying committee that includes technology, risk, compliance, and business leaders. Strategy and operations roles suggest the company is building the coordination layer required to scale a hardware-enabled business, not only adding model researchers.

Hiring is therefore a demand signal, but not proof of future growth. Candidates and partners can review active roles directly on the Go.AI careers page. The more important market signal is what the roles reveal: private AI is becoming an operating discipline that needs deployment, governance, education, and support alongside software.

What Go.AI Signals for Enterprise AI

The first enterprise AI race rewarded access to powerful models. The next race will reward control over the full decision path. Organizations will ask where inference happens, which information is available, how permissions change, what a workload costs, and who can produce evidence after an output enters a regulated process.

Go.AI is betting that local and private infrastructure will become a permanent layer of that market rather than a temporary reaction to cloud anxiety. The thesis is credible because regulated organizations cannot outsource accountability, even when they outsource compute.

The proof now moves from funding and reported usage to repeatable deployment. If Go.AI can make customer-controlled AI easier to install, govern, and operate across hundreds of institutions, the company will have done more than sell an appliance. It will have made the deployment boundary a first-class product decision for enterprise AI.