Connect the signals
Pull together the data that actually bears on the decision, across the systems you already run.
AI decision systems
Signal systems that move the needle.
Better decisions from the systems you already run. Mantis builds focused AI decision systems that turn fragmented business signals into explainable recommendations, one recurring decision at a time.
Every system we build is another step toward an AI Operating Partner.
Operating signals · example
Customer Renewal risk rising Usage -18% · 41 days to renewal
Revenue Expansion window Seat utilization at 92%
Operations Location divergence Margin -6% vs peer stores
Mantis OS Context + lenses + reasoning What changed? Why now? What is missing?
Recommended action High confidence
Prioritize a targeted renewal intervention for the enterprise cohort before the broad reactivation campaign.
An illustration of the operating layer: signals in, one recommended action out.
20+ yrs
building operating systems in data-rich businesses
3 domains
SaaS · restaurants · investing,
one decision architecture
one decision architecture
Strategy → build
diagnosis through working systems
Built around recurring decisions, not an empty chat box.
Signal systems that move the needle§ 01 The real problem
Your business has more
data than judgment.
Data lives across CRM, product, billing, finance, customer, operational, and external systems. Dashboards describe pieces of the business. AI tools answer isolated questions.
Leaders still have to connect the signals, recover the context, judge what matters, and decide where to act. That gap between information and decision is where the work is.
Mantis builds the intelligence layer between business systems and business decisions.
Business decisions
Intelligence layer
Business systems
§ 02 What we build
Focused AI decision systems,
built around one decision at a time.
A decision system is built around a recurring business decision. It connects the signals that bear on it, applies the reasoning frameworks that fit the problem, and produces an explainable recommendation you can act on.
Today that means one of three ways to work together: a focused decision diagnostic, a first decision system built on your data, or an ongoing operating advisor.
Apply the lenses
Reason through the frameworks that fit the problem, with logic you can follow, not a black box.
Recommend the call
Produce an explainable recommendation and a next step, pointed at the decision itself.
Built around your operating model
Each system is shaped to how your business actually runs, its data, metrics, and workflows, not a generic template.
Problem-led reasoning
Lenses are selected by the problem in front of you, not by the org chart.
Explainable judgment
Every recommendation shows the signals and lenses behind it, so you can follow the reasoning and push back on it.
Persistent intelligence
The system remembers the business over time and gets sharper as decisions play out.
§ 03 Proof
One reasoning architecture,
already running in three domains.
Each system below is a different business, the same architecture underneath. They demonstrate the approach rather than describe it.
B2B SaaS growth Advisory + architecture
The whole growth lifecycle, self-serve and sales-led.
A decision system for SaaS growth across both motions, product-led self-serve and enterprise sales-led. It reasons over acquisition, activation, and then retention, adoption, and expansion, the part that quietly decides net revenue and that most tooling ignores.
Signals
Product usage, activation, pipeline, engagement, renewals, expansion
Lenses
Activation & Adoption, Self-serve Conversion, Pipeline Health, Churn & Retention, Expansion, Forecasting
Judgment Where net revenue is growing or leaking across both motions, and the highest-leverage place to act.
§ 04 The architecture
One reasoning model under
every decision system.
Every system we build runs the same model. It is the working architecture underneath them, not a boxed platform. This is the repeatable part.
01 Problem What decision is at stake?
02 Signals What changed?
03 Lenses How should it be viewed?
04 Judgment What matters most?
05 Recommendations What should we do?
06 Actions Where does it land?
07 Learning What did we get right?
On lenses
A lens is a reusable reasoning framework, not a department. The system activates the right combination for the problem in front of it. Lenses are selected by the problem, not by the org chart.
Customer LifecycleUnit EconomicsPricingRiskForecastingProduct AdoptionCustomer BehaviorCompetitive PositionOperational Bottlenecks
Persistent context. The system remembers the business, not just the last prompt.
Transparent reasoning. It shows which signals and lenses shaped the judgment.
Operational connection. Recommendations reach the people and workflows that can act.
§ 05 Why Mantis exists
The systems were always the
same thing underneath.
For more than twenty years I built systems that help businesses make better decisions. Revenue systems. Lifecycle systems. Restaurant systems. Investment systems.
They ran on the same architecture underneath: one disciplined way of turning business signals into judgment. Mantis makes that architecture repeatable, one decision system at a time.
Abhay Taiwade, founder
§ 06 Where this is going
Every decision system is a step
toward an AI Operating Partner.
Enterprise software has climbed toward judgment for decades, from recording what happened, to reporting it, to automating tasks, to assisting with prompts. The next category makes the decision.
01 Systems of Record
02 Systems of Engagement
03 Analytics & BI
04 Workflow Automation
05 AI Assistants
06 AI Operating Partners Where this is going
An AI Operating Partner is a persistent system that reasons across a business's decisions over time. We are not claiming to have built a universal one. We are building toward it, one focused decision system at a time, on a single reasoning architecture.
§ 07 How to begin
Start with one decision.
Build from evidence.
Every engagement starts with one recurring decision and one system to improve it. Prove it, then expand. Not every problem needs AI, and when the honest answer is cleaner data or a call someone just has to make, we say so. No transformation theater.
Why are renewals falling? Which locations need intervention? Which opportunities deserve attention? Which customers are at risk?
01 Focused entry
Decision Diagnostic
A focused read on the recurring decisions, workflows, signals, and data where a decision system can create real operating leverage.
- Decision and workflow diagnosis
- Data and system readiness
- Prioritized system roadmap
02 Build + prove
First Decision System
Design and build a focused decision system connected to real data, explicit reasoning, and a recurring operating workflow.
- Signals and decision architecture
- Working intelligence layer
- Briefs, recommendations, and action path
03 Ongoing partner
Operating Advisor
Ongoing strategic and operating support for leaders evaluating, building, and governing practical decision systems. The deeper end of the Twin.
- Use-case and architecture guidance
- Decision reviews and system evolution
- An experienced operator in the room
§ 08 Meet the thinking
Start with the Twin before
engaging the operator.
Abhay's AI Twin is an interactive introduction to how he frames a business problem, hunts the missing signal, challenges assumptions, and moves toward action. Bring it a real operating question.
The diagnosis can begin with the Twin. The system behind it is the work.
01
Meet Abhay.
Explore the experience, frameworks, and philosophy behind the work.
02
Work the problem.
Bring a real operating question and clarify the decision before solutioning.
§ 09 The operator behind Mantis
AI credibility built on twenty
years of operating reality.
Mantis Growth Labs was founded by Abhay Taiwade, an AI-native operator and builder with more than 20 years connecting data, business systems, customer behavior, product engagement, GTM execution, and financial performance.
He has built operating and intelligence systems inside scaling SaaS companies, led data and analytics at two companies with a combined $680M in exits, and now builds focused AI decision systems for data-rich businesses, each a step toward the AI Operating Partner category.
Useful AI requires more than a model. It requires the right signals, business context,
decision architecture, operating cadence, and connection to action.
Founder
Abhay Taiwade
§ 09.1 Who Mantis is for
Data-rich businesses where decisions still run on fragmented information.
Self-serve (PLG) and sales-led SaaS
Founders & CEOs
CFOs & Finance leaders
Revenue & GTM leaders
Multi-location restaurant operators
Operators in data-rich businesses
Investors evaluating signal systems
Teams whose AI pilots stalled at chatbots
§ 10 Bring us one recurring decision
Where could an AI operating partner
create leverage in your business?
Start with one fragmented workflow, recurring decision, or operating problem. We will help determine whether it should become an AI system, a better data system, or neither. No transformation theater, operator to operator.