enterprise42.co

Enterprise memory · AI agent security

Better context.
Safer AI agents.

MindTwin turns customer conversations into current account context. AAPT tests how agents use data, tools and permissions. Enterprise42 builds both and helps teams take AI from strategy to systems they can operate.

See the products Request materials

Remember what matters.
Test what can go wrong.

Two products for distinct needs: account context for Customer Success, and release evidence for engineering and security teams.

MindTwin

Governed memory for Customer Success

See changing customer intent earlier.

Bring account conversations from Slack, Gmail and Gong into a living memory. MindTwin links new signals to earlier commitments, flags potential renewal risk with dated sources, checks drafts and prepares account handoffs.

For
Customer Success and account teams
Governance
PII redaction and crisis-content suppression before storage, with per-team memory isolation
Access
Dashboard, Slack, REST API and MCP
Start
A six-week paid pilot with one account segment, an agreed fixed fee and a shared success metric
Explore a MindTwin pilot ↗ Request the product deck

AAPT

Security testing for AI agents

Make the next release decision with evidence.

Assess how an agent could expose data, abuse tool access or exceed its permissions. AAPT combines automated adversarial tests with manual red teaming to give engineers and security leaders a prioritised remediation plan.

Coverage
74 probes, 28 categories and eight attack families, including MCP, GraphQL and multi-turn attacks
Outputs
Reproduction steps, CVSS-A scored findings, fix guidance, executive and technical reports, and remediation re-tests
Deployment
On-premises and air-gapped assessments available; the judge runs locally in sovereign mode
Start
A free 30-minute scoping call, then a one-off assessment, recurring testing or an enterprise engagement
Scope an AAPT assessment ↗ Request the product deck

Explore the evidence. See MindTwin’s product dashboard and published research and limitations, or review AAPT’s assessment methodology and severity scoring.

An AI-native services practice, from strategy to the GPU

Put the products to work in your organisation, or build for a different AI workflow. Our consulting and engineering practice connects strategy, application development, infrastructure and security.

  1. 01 / STRATEGYDecideWhat changes when the organisation assumes intelligence is cheap.

    • AI-native strategy

      Where AI changes the operating model rather than the feature list: which decisions get automated, which data becomes an asset, how the org and its economics shift.

    • AI-first digitisation

      Digitise processes so agents can consume them from day one: structured records, APIs and event streams, not scanned forms behind a portal.

  2. 02 / BUILDMakeSystems that perceive, reason and act, evaluated before they meet a customer.

    • AI agent building

      Multi-agent systems with tool use, memory and guardrails, tested against real tasks with a proper eval harness, not a demo prompt.

    • Physical AI

      Perception and control for machines: vision models on Jetson Orin with TensorRT, sorting lines, robotic gantries and field equipment.

    • Training and inference

      Fine-tuning and distillation on your own data, and serving that hits a stated latency and cost budget.

  3. 03 / RUNOperateYour models, your data, your hardware, your jurisdiction.

    • Sovereign AI

      Design models, data and inference for infrastructure and jurisdiction you control. Agree deployment boundaries, model dependencies and data handling before the build.

    • AI infrastructure

      GPU cluster design and sizing, procurement guidance, scheduling, storage and observability, on-premise or in your cloud account.

    • AI workload optimisation

      Quantisation, batching, KV-cache and kernel-level tuning so the GPUs you already own do more of the work.

  4. 04 / SECUREDefendAcross every layer, because agents that act can be made to act badly.

    • AI agent security reviews

      Threat modelling and red-teaming of agent systems before launch, using AAPT and hands-on testing, with a hardening plan your engineers can execute.

Fixed-scope engagements, from a two-week assessment to a full build. Every engagement leaves your team able to run what we built.

Why 42

In Douglas Adams's story, a planet-sized computer spends seven and a half million years computing the answer to the ultimate question of life, the universe and everything, and returns 42. The joke lands because nobody had worked out what the question was.

That is most of our consulting work: getting the question right before the compute is spent. The number is the reminder. The full stop after it is the point. We finish things.

The people behind the products

Enterprise42 brings product development and AI engineering together. Meet the founder behind MindTwin and AAPT, and explore the teams taking shape around them.

FOUNDERSubrat Panda

  • Subrat Panda

    Founder and CEO

    PhD in Computer Science from IIT Kharagpur and faculty at Plaksha University. Previously CTPO at AgNext and VP at Capillary Technologies. Subrat builds MindTwin and AAPT and authored TWIST, an open benchmark for conversational-memory intervention quality.

    LinkedIn · Full profile

FOUNDING TEAMFounding team

Help shape the product teams across sales, solutions, customer support, integrations and global expansion. MindTwin and AAPT welcome introductions for future founding roles; both are currently exploring relationships without active hiring.

MindTwin careers ↗AAPT careers ↗

Customers, advisors and investors

Start with a scope

Customer engagements

Evaluate MindTwin on a focused account segment, or test an agent with AAPT. Agree the scope and success criteria before the work begins.

Explore a memory pilot · Discuss a security assessment

AAPT advisory

Product advisors

Prof. Anupam Chattopadhyay of NTU Singapore advises AAPT on secure AI and AI-generated hardware. MindTwin’s advisory group is being formalised.

Meet AAPT’s advisor · MindTwin team and advisors

Materials on request

Investors

Explore the company thesis, the two product opportunities and the engineering practice. Request an investor deck and tell us your investment focus.

Request the investor deck

Questions people ask us

What is Enterprise42?

An India-based AI company behind MindTwin, governed account memory for Customer Success, and AAPT, security testing for AI agents. We also offer AI-native consulting and engineering from strategy through infrastructure and security.

What is MindTwin?

MindTwin connects account conversations from Slack, Gmail and Gong into governed enterprise memory. It tracks changes in customer intent, surfaces potential renewal risk with dated evidence, checks outgoing drafts and helps teams hand off accounts. Its dashboard, Slack integration, REST API and MCP make that context available to people and agents. Explore it at mindtwin.me.

How does a MindTwin pilot work?

A six-week paid pilot focuses on one account segment. The fixed fee, scope and success metric are agreed together before starting, and your team keeps the findings. Discuss a pilot with MindTwin.

What is AAPT?

AAPT combines automated adversarial testing and manual red teaming for AI agents. Its current library has 74 probes across 28 categories and eight attack families. Assessments produce reproducible findings, remediation guidance and executive and technical reports, with fixes re-tested. AAPT’s proprietary CVSS-A score adds AI-specific severity context. Explore it at aapt.online.

Can AAPT test agents inside our infrastructure?

Yes. AAPT offers on-premises and air-gapped assessments. In sovereign mode, the judge runs locally without third-party model calls or data egress. Relevant interfaces and attack surfaces are agreed during scoping. Explore sovereign testing.

Do security assessments certify compliance?

AAPT provides testing evidence mapped to frameworks such as OWASP LLM Top 10, MITRE ATLAS, NIST AI RMF and ISO/IEC 42001. These mappings support risk reviews and audit discussions; they do not certify compliance.

What services does Enterprise42 offer?

AI-native strategy, AI-first digitisation, AI agent building, physical AI (vision and control on edge hardware such as NVIDIA Jetson Orin), model training and inference, sovereign AI, AI infrastructure design, AI workload optimisation and AI agent security reviews. Engagements are fixed-scope, from a two-week assessment to a full build.

What does sovereign AI mean here?

Designing models, data and inference to operate within infrastructure and jurisdiction you control. For our engineering engagements, the deployment boundary, model dependencies and data handling are agreed in the scope. AAPT separately offers a local-judge mode for sovereign security assessments.

Who founded Enterprise42?

Subrat Panda, who holds a PhD in Computer Science from IIT Kharagpur and serves as faculty at Plaksha University. He previously served as CTPO at AgNext and VP at Capillary Technologies, and is the founder behind MindTwin and AAPT.

How do investors get the pitch deck?

Request the investor deck using the form below, or email investors@enterprise42.co. We review requests and share materials directly.

Why the name?

The 42 is a nod to Douglas Adams: the famous answer to the ultimate question, useless because nobody had worked out the question. Our consulting starts by getting the question right before compute is spent.

Get a clearer picture.

Request a product deck, a one-pager, or the investor deck. Tell us what interests you, and we’ll follow up with the relevant materials.

  • Product deckExplore the product, its workflows, and how an engagement works.
  • One-pagerA concise overview to share with your team.
  • Investor deckThe company thesis, products, and business opportunity.

Prefer email? hello@enterprise42.co

Which materials would you like?

Choose one or more.

Your details are used to review and respond to this request. To ask for correction or deletion, contact hello@enterprise42.co.

Preparing the request form…

Talk to us

Tell us what you are building and what is in the way. We reply within two working days. Based in India, working with clients worldwide.

Request a product deck, one-pager, or investor deck

hello@enterprise42.co