Tools making dashboards fast

Making a dashboard is starting to look more like writing a prompt than laying out charts by hand. Lightdash, Anthropic’s Claude, and ChartCraft are all showing tools that turn data or plain-language requests into dashboard mockups and interactive views. (lightdash.com, anthropic.com, chartcraft.io) Lightdash says its product sits on top of dbt, the data build tool used by analytics teams to define metrics and business logic in code. On its homepage, the company says AI can “build, refactor, and ship analytics in minutes” and shows a prompt that creates a sales dashboard from an approved revenue model. (lightdash.com) That pitch targets a familiar bottleneck inside data teams: the metric definitions may already exist, but someone still has to assemble charts, filters, and layouts for each stakeholder request. Lightdash is framing that assembly step as something an agent can automate while still routing queries through a governed semantic layer. (lightdash.com) Anthropic is pushing the same idea from the chat side. Its product pages now list Claude for Excel, Claude for PowerPoint, and interactive apps inside Claude, and the company’s platform says developers can “create agents and applications” with managed infrastructure. (anthropic.com, platform.claude.com) Anthropic expanded that interface on January 26, 2026, when it launched interactive apps inside Claude for paid users including Pro, Max, Team, and Enterprise plans. TechCrunch reported the launch apps included workplace tools such as Slack, Canva, Figma, Box, and Clay, with Anthropic saying visual interfaces help users analyze data and design content faster. (techcrunch.com) ChartCraft is approaching the problem as a design studio for business intelligence teams rather than a general assistant. Its site says users can describe a dashboard in natural language, get suggested chart types and layouts, analyze an existing design for hierarchy and accessibility issues, and export the result to tools including Power BI and Tableau. (chartcraft.io) A BetaList profile published this week describes ChartCraft’s workflow even more directly: upload an existing dashboard, get an artificial-intelligence design score, receive recommendations on layout, contrast, typography, and color harmony, and generate an enhanced version ready to implement. That positions it less as a database query tool and more as a polishing layer between analysts and executives. (betalist.com) The common thread is that dashboard work is being split into smaller jobs that software can handle: one system understands the data model, another generates the interface, and another critiques the presentation. Vendors are now packaging those steps into tools that promise faster drafts from dbt projects, spreadsheets, and prompt-based requests. (lightdash.com, chartcraft.io, platform.claude.com) That does not remove the hard part of analytics, which is agreeing on the metric definitions and deciding what a leadership team actually needs to see. It does move more of the visual assembly work into software, which is why dashboard generation is increasingly being sold as a minutes-long task instead of a multi-meeting project. (lightdash.com, chartcraft.io)