Company overview

The data layer for the agentic era.

Supper connects to a company's entire data stack, builds a semantic model around how the business actually measures itself, and lets anyone (or any agent) get a verified answer in seconds.

The founders

Lowell Putnam & Andy Salamon · Co-founders

Lowell Putnam

Co-founder & CEO

Lowell started his career in investment banking before founding Quovo, a fintech data platform acquired by Plaid. He then helped build MIRROR, the connected fitness company acquired by Lululemon. He brings a pattern of finding infrastructure gaps in large markets early and building the category-defining company around them. Supper is his third act, and the one most directly born from watching companies struggle to get trustworthy answers out of their own data.

Quovo → Plaid MIRROR → Lululemon Investment banking Active investor & board member Harvard College

Andy Salamon

Co-founder & President

Andy has been an entrepreneur and early-stage investor for over a decade, with prior roles at Bridgewater Associates and in the US Navy. He has invested in more than 20 startups, with notable exits including Hims & Hers Health. He brings operator discipline, pattern recognition across dozens of high-growth companies, and a deep understanding of where the pain actually sits; not just the data team, but the operators, executives, and founders waiting on answers.

Bridgewater Associates US Navy Hims & Hers exit 20+ startup investments Harvard College Wharton MBA

Usage & traction

Total platform messages · Jan–May 2026

Total platform messages

Jan–May 2026

Total messages processed per month across all accounts. Trailing two-week daily average rose from ~4,300 in January to ~21,500 in May.

35.4% CMGR Compound monthly growth in platform messages, Jan–May 2026.

— Messages more than tripled in five months — from 22.6K in January to 76.3K in May, with the sharpest jumps in the last two months (+97% in April).
— Daily usage is compounding. The trailing two-week average climbed ~5× over the same window. We're seeing deepening engagement from existing users as well as new organizations.
— Real outcomes, recurring. Coldcart saved 8 hours a week. One customer cut board reporting from 15 hours to minutes. Customers are learning how to integrate Supper for real ROI in their companies.

Customer pipeline &
onboarding

Onboarding timeline — signed & near-term pipeline

Onboarding timeline

Phases begin from each account's source-linking date. Vertical marker indicates today. Dates reflect signed and near-term-closing pipeline.

The pipeline is full and growing. We're currently constrained by onboarding capacity rather than customer demand.

— Forward-deployed analyst onboarding is a category differentiator for Supper. Analysts promote usage and handle more complex requests, which deepens engagement early on as customers learn the system.
— Expanding analyst headcount unlocks parallel onboarding and will be necessary to continue accelerating our GTM motions. Over time, the platform will grow in autonomy to add scale to the analyst function.

Market positioning

Legacy BI

Looker · Tableau · Mode

Built for data teams, not business users. Dashboard-centric, SQL-dependent. Every new question requires a new build. Fast answers require analyst involvement.

Analyst-dependent

BI 2.0

ThoughtSpot · Omni · Sigma

An extension of the legacy model: friendlier interfaces, some natural language, but still anchored to predefined schemas and human-maintained toolkits. Not built for agents.

Dashboard-first

Agentic Data

Supper

AI-native from the ground up. Verified semantic model, cross-system neutrality, built for both human questions and agent workflows. The answer layer that didn't exist before.

Open category

The cost of insight

Legacy BI and BI 2.0 share the same cost model: every insight requires a human to build something. A dashboard gets built, then maintained, then rebuilt when the business changes. The analyst queue handles everything the dashboards can't.

The cost of inference has collapsed.

Supper capitalizes on a new cost structure for inference. The system promotes more questions, more exploration, and more analysis. Our semantic models increases in accuracy with usage, completely inverting the standard BI workflow.

Legacy BI

High

Analyst time to build and maintain every dashboard

BI 2.0

High

Friendlier interface, same underlying cost structure

Supper

Low + falling

AI-powered semantic model and analysis engine, where usage begets usage, at appealing margins

Why Supper wins

01 · Accuracy layer

The accuracy layer nobody else has.

Every answer runs through a customer-specific semantic model: its definition of ARR, its churn formula, its pipeline stages. Not generic SQL generation. Not a best guess. Verified before it runs, auditable after. That's why CFOs trust it when they won't trust a generic AI tool.

02 · MCP-native

MCP-native for the agentic stack.

Supper exposes it's core Q&A system via MCP. Any compatible agent (Claude, ChatGPT, Copilot, LangChain, etc) can query Supper with the same semantic model, permissions, and audit trail as the native product. As enterprise AI deployments proliferate, Supper becomes the verified data substrate every agent in the stack calls.

03 · Retention moat

Deployed Analysts: the trojan horse for retention.

Every customer gets an embedded data analyst who builds and maintains their semantic model. This is the delivery mechanism and the moat. analysts drive fast time-to-value, deepen product embedding, and generate institutional trust that's hard to displace. As the semantic model matures, analyst time per account decreases expanding our margin profile.

04 · Token economics

Token efficiency that changes the economics.

Most agentic data systems consume 100,000+ tokens per query (broad context windows, speculative retrieval, expensive chains, etc). Supper's semantic model does the pre-work: queries run against a precise, pre-validated structure. Typical Supper processes run in hundreds to low thousands of tokens.

05 · Neutrality

Neutrality as structural moat.

Snowflake governs Snowflake data. Databricks governs Databricks data. Neither can be the neutral layer for an enterprise running across multiple independent systems. Supper has no stake in which warehouse or SAAS stack our companies use. This neutrality posture creates a moat that the platform vendors can't replicate without undermining their own business model.

06 · Enterprise-grade

Enterprise-grade from day one.

SOC 2 Type II. GDPR compliant. Role-based access control enforced at query time, not just the UI. Schema-, table-, and column-level permissions. Full audit trail on every query. BYO LLM and BYO S3 for enterprises with data residency requirements. The trust infrastructure regulated industries need before they'll let an agent touch their data.

A new kind of data infrastructure is being built.

The bottleneck in enterprise AI isn't the models. It isn't the warehouses. It's in between — the context layer that knows what the data means, who can see it, and whether the answer is right.