Kritmatta has launched an AI agent platform for small and medium-sized businesses, built to run workflows across the software teams already use. The goal is straightforward: move AI out of the experimentation phase and into the systems people rely on every day, without asking anyone to learn a new tool from scratch.

The platform works across email, CRM systems, LinkedIn, messaging platforms, productivity tools and internal data sources, and it is aimed squarely at non-technical teams managing work across multiple systems at once.

The problem was never a data problem

"Most teams do not have a data problem. They have a workflow problem," says Kalpesh Baxi, Kritmatta co-founder. "The signal is already there, but it is spread across inboxes, CRMs, LinkedIn, WhatsApp and internal notes. Kritmatta is designed to connect that signal and turn it into action."

That is the starting point for most smaller businesses. Information is not missing, it is scattered: an email here, a WhatsApp message there, a CRM note that never quite gets updated, a LinkedIn reply that lands and gets forgotten. Someone still has to connect the context by hand, decide what matters, and remember to follow up. Kritmatta's agents are built to do that connecting work, monitoring signals, structuring information, identifying relationships, and queuing the next action for a human to review.

Built around entities, not prompts

The core of the platform is an intelligence layer based on a knowledge graph and entity resolution. In practice, that means the system recognises when the same person, company, customer, project or opportunity shows up in different places, and treats it as one connected piece of work rather than several disconnected mentions.

"A lot of AI tools still behave like search boxes with better language," says Manjit Johal, Kritmatta co-founder. "We have built Kritmatta around entities, relationships and context, so agents can understand how people, companies, projects and opportunities connect across a business."

That distinction matters in practice. A prospect who messages on LinkedIn, emails a week later, and shows up in a CRM note about the same account is not three separate signals. It is one relationship with three touchpoints, and treating it that way is what lets an agent queue the right next action instead of three unrelated ones.

Three layers, one workflow model

The platform operates across three layers. The productivity layer handles the everyday load: inbox summaries, meeting summaries, communication aggregation. The intelligence layer sits above that, connecting signals across channels and prioritising what actually needs attention.

The third layer, Vega, is built for lead generation and growth: opportunity identification, relationship expansion, account penetration and signal-based outreach. The same underlying workflow model carries across teams handling very different kinds of relationships, whether that is clients, candidates, campaigns, projects or commercial pipeline.

Designed for teams without specialist support

"AI only becomes useful when people can apply it inside the way they already work," says Duncan Stuart, Kritmatta co-founder. "Our focus has been on designing workflows that operational teams can configure directly, without needing technical expertise."

More than 60 per cent of registered users have an active agent running within 24 hours of signing up, which is the number Kritmatta points to as evidence that the configuration bar is low enough for non-technical teams to clear on their own.

Backed with that focus in mind

Kritmatta is backed by Velocity Capital as part of the firm's technology portfolio. Thomas Lindup, Chief Operating Officer at Velocity Capital, framed the investment around the same non-technical audience: "Kritmatta is focused on how AI and AI agents can be effectively deployed within the workflows of non-technical teams in small and medium-sized businesses, in a predictable, auditable and safe manner."

Where it is already working

The platform is already in use across recruitment, talent, professional services, SaaS businesses and other commercial settings where teams manage fragmented communication and a high volume of day-to-day workflow. A typical case: a team receives a LinkedIn message from a prospect, an email from a client, a CRM note tied to an opportunity, and an internal update, all about the same account. Kritmatta connects those signals and queues the next action for review, rather than leaving someone to piece the account's story back together from four different places.

Simple AI. Real results.

This post draws on coverage of the launch published by a UK business news outlet on 22 July 2026, including on-record comments from Kalpesh Baxi, Manjit Johal, Duncan Stuart and Velocity Capital COO Thomas Lindup.