Tejas AI

What do they actually do

Tejas AI is building an AI‑powered decisioning tool for banks and lenders that helps risk teams change, test, and deploy credit/risk policies faster than manual processes. Banks upload historical loan data, existing rules, and business goals; the platform runs what‑if simulations to estimate approvals, default rates, and profit impact, with explainable recommendations and a plain‑English interface YC launch post / YC profile.

When a policy change is approved, Tejas AI can convert it into executable rule code for the bank’s rule engine to reduce hand‑coding and deployment time, and then monitor performance to retrain/update recommendations as the portfolio evolves YC launch post.

Public signals suggest the product is early: the company launched with YC (W25) via a post and short video; the listed team is two founders; and the live website currently shows a different product (people/company intelligence), implying the public site is out of sync with the YC banking product description YC profile / YouTube video / trytejas.ai homepage / Terms/contact. A promotional example mentions identifying a locality with ~25% higher defaults and rapidly updating policy, but this is not an independently audited case study LinkedIn.

Who are their target customer(s)

• Risk managers and policy owners at banks/NBFCs: They update credit rules regularly but face slow, manual processes and limited ability to test impacts before rollout; they need evidence‑backed simulations to reduce risk when changing policies YC launch post.
• Compliance and audit teams: They must justify lending decisions to regulators and auditors and need clear, explainable documentation tying any policy change to expected business and consumer impacts YC launch post.
• Core‑banking/integration engineers and rule‑engine owners: They translate policy text into executable rules by hand, burning engineering time on coding and testing and risking production errors; automation of rule generation would cut cycle time YC profile.
• Product and growth teams at fintech lenders: They want to increase originations without raising defaults and need fast, data‑backed experiments to tune approval criteria by segment or region YC profile.
• Regional or portfolio managers (branch/city‑level): They lack granular visibility under national policies and need tools to spot local pockets of higher risk and adjust rules quickly to limit losses LinkedIn example.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run 3–6 week paid proof‑of‑value pilots with nearby banks/NBFCs via YC intros and founder networks, converting one live policy into executable rule code to show immediate engineering/time savings YC launch post.
• First 50: Productize a fixed‑scope “policy → simulation → rule” package and sell short, paid pilots to similar regional banks/fintechs using early case studies; add 1–2 integration partners (core‑banking/BRE vendors) to handle implementations and shorten procurement YC launch post.
• First 100: Hire a few enterprise sellers with banking risk backgrounds; ship templated connectors and audit‑ready artifacts to cut onboarding friction, and scale via channel partners, regulator/industry events, and a reference program that trades discounts for public case studies.

What is the rough total addressable market

Top-down context:

Combining bank risk‑management software (~$11.4B) and credit‑decisioning software (~$8.1B) implies a current global TAM around $18–20B, with growth in the low double digits annually Decision Advisors / DataIntelo.

Bottom-up calculation:

Illustratively, if ~1,200 large/upper‑mid lenders buy platforms at $0.5–1.5M ACV and ~10,000 midsize lenders adopt lighter packages at $50–150k ACV, the blended total lands in the high‑teens billions; India alone contributes a large pool with 135 scheduled banks and ~9,300 NBFCs MoF/RBI.

Assumptions:

• Average ACVs: large $0.5–1.5M; midsize $50–150k (software + support).
• Institution counts are approximate; adoption is staggered, not 100%.
• Categories overlap; $18–20B is an order‑of‑magnitude TAM, not a precise total.

Who are some of their notable competitors

• FICO: Long‑established decisioning/scorecard vendor; its platform and rule‑engine products cover model governance, policy automation, and monitoring used by large banks—overlapping Tejas’s test/explain/deploy workflow.
• Experian PowerCurve: A commercial decision engine bundling data, analytics, and rule execution for origination, collections, and fraud; banks use it for automated policies, monitoring, and regulator‑facing reports.
• Zest AI: AI underwriting and automated credit‑decisioning with explainability and connectors into loan‑origination systems; positioned for faster ML‑based rule updates and auto‑decisioning.
• H2O.ai (Driverless AI): AutoML and model‑deployment tooling with built‑in explainability and reason codes; banks use it to build/validate/serve credit and fraud models with compliance‑friendly outputs.
• Red Hat Decision Manager / Drools: Widely used BRMS/rule engine (DMN/DRL) for authoring, testing, and running executable rules; a common target integration for any “policy → executable rule” automation.