cd /news/ai-agents/introducing-taktile-ai-mcp-connectiv… · home › topics › ai-agents › article
[ARTICLE · art-171678] src=taktile.com ↗ pub= topic=ai-agents verified=true sentiment=↑ positive

Introducing Taktile AI & MCP Connectivity

Taktile released Taktile AI and Taktile MCP connectivity on November 18, 2026, letting banks' and insurers' risk, policy and product teams describe changes to decision flows and agents in natural language, test them, and route them for human approval. Taktile AI is available natively in the Taktile platform through a built-in chat box, while Taktile MCP connects compatible AI assistants to Taktile capabilities, including via a dedicated Taktile connector in the Anthropic directory. Senior Credit Underwriter Kyle Jacobs of Nivoda said the tool lets him "investigate the issue myself, understand what happened, test a change, and send it for review" without waiting for specialist support.

key takeaway

Taktile MCP lets technical teams connect a compatible AI assistant to Taktile's decision-flow capabilities, including through a dedicated Taktile connector in the Anthropic directory, while native Taktile AI keeps natural-language edits and testing inside the platform with human approval required fo

by read3 min views9 publishedNov 18, 2026
Introducing Taktile AI & MCP Connectivity
Image: Taktile (auto-discovered)

Back to blog Lucas Grabmaier

Senior Product Manager

ProductNov 18, 2026

Move forward faster by testing and optimizing agents and decision logic in natural language.

Banks and insurers combine data, policies, and judgment to make complex decisions like how much to lend, or whether a claim is fraudulent. But when the right decision is always changing, teams need a way to quickly adapt their decision strategy.

Taktile AI empowers risk, policy, and product teams to put their expertise into action through natural language.

That means fewer handoffs between business and technical teams to optimize and test policies, decision flows, and agents. The people who own KPIs like loan approvals or claims leakage can directly prepare changes, test how they behave, and improve outcomes.

AI accelerates routine work, while final changes still require approval from a human expert.

Here’s how your team get their hands on the dials of Taktile AI, and start taking even more ownership of their performance, with safeguards in place.

A more approachable way to work with agents and decision logic #

We built Taktile AI so our customers never have to wait to make a change that enhances their performance.

It enables both risk and product owners to quickly understand and adjust the decision processes that impact their KPIs. Teams can gather insights and act on them in one place, making continuous optimization part of how the business works.

Taktile AI significantly reduces the time it takes for users to get onboarded and extract value from the platform, and serve their own needs – whether it’s making structural changes to workflows, identifying optimization opportunities, or simply amending variables.

Taktile AI makes troubleshooting and problem-solving really easy, so I can resolve issues quickly without much effort. I no longer have to wait for specialist support to dig into the logic. I can investigate the issue myself, understand what happened, test a change, and send it for review.

— Kyle Jacobs, Senior Credit Underwriter at Nivoda

A policy or risk expert can describe a change in business terms and get help preparing an edit to the relevant agent or decision flow. They can test its behavior and review the result before deciding what should happen next.

A product team member can ask how a decision was made, follow the logic behind an outcome, or investigate a failed run.

Taktile AI connects these conversations to the work in Taktile. You can start with a question or a goal, then move into the relevant agents, flows, and testing surfaces to inspect and validate the change. The aim is not to replace your team's expertise, but to make it easier to apply—and easier for business and technical colleagues to work together.

Collaborate with Taktile AI natively or via MCP #

Teams can decide how they want to collaborate with Taktile AI, whether natively within the Taktile platform, or via MCP.

With Taktile AI in the platform, policy experts can type natural-language requests into a built-in chat box, and continue in Taktile's native interface to review flows and test changes.

Via Taktile MCP (Model Context Protocol), technical teams can connect a compatible AI assistant to Taktile capabilities and work with Taktile as part of a broader, multi-tool workflow. You can also connect via MCP using the dedicated Taktile connector (https://claude.ai/directory/taktile) in the Anthropic directory.

A business user might start in the platform to prepare and inspect a policy change, while Taktile MCP is helpful for pulling in context from Taktile while working on other tasks. Teams can choose the approach that fits the person and the work, and use both where it makes sense.

Keep people in control where it counts #

Taktile AI is designed to make routine work more accessible and collaboration more fluid—not to remove the judgment, oversight, or accountability that important decisions require.

Whether you work directly in Taktile or through a connected AI assistant, Taktile AI helps your team turn business intent into changes you can test and understand—while you stay in control of what goes live.

── more in #ai-agents 4 stories · sorted by recency
── more on @taktile 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/introducing-taktile-…] indexed:0 read:3min 2026-11-18 · —