Platform Release · Summer 2026

An entire agentic platform — in your hands.

V27 is the release where the platform becomes something you direct rather than operate.

Crew

A crew of AI specialists you lead by conversation.

Data

A home for every kind of data — to learn from, process, and query.

Build

A no-code way to build the apps your teams need.

Trust

A layer of trust over every action.

Four capabilities, one platform.

V27 · by the numbers

5copilot specialists behind Piper
200+platform actions at its command
36drag-and-drop experience cards
6agent types, one factory
Mesh throughput, re-architected
1new Studio pillar: Governance

Experience

Build the app by configuration

One idea quietly rewrites delivery economics this cycle: one rendering engine assembles many different applications from a specification. No bespoke front-end per customer. Cards are the building blocks; cards compose into Pages; Pages bundle into Packs — a whole app layered over the base product.

Build once, reuse everywhere

Each card is screen- and data-aware but free of any single business's specifics — so a card built for a lending workbench works unchanged on a claims desk.

Configuration makes it yours

Cards wire to your data model through configuration, with a visual designer and live preview. Adding a brand-new card is a registry entry, with zero edits to the renderer.

This is not a preview — whole solutions run on this layer today:

In production

Credit-spreading workbench

A source viewer beside a two-period spread grid, per-line verify and flag, ratios, covenants and a decision panel.

In production

Loan origination

Documents drop in, classify, split and extract with provenance, then self-fill a form dozens of fields deep.

In production

Operational dashboards

Assembled entirely by configuration — metric cards, charts and structured-data lists, with live refresh.

Why it matters

A tailored, customer-specific application now takes days, by configuration — and it travels between environments as a single portable package.

Piper

The copilot for Solution Developers

Piper is one interface that navigates the platform, configures it, runs it, and reports back — in plain language. It does not work alone: behind Piper stands a crew of five specialists it directs.

You lead Piper; Piper leads the crew — routing each ask to the right specialist, with over 200 platform actions at its command.

Scoped to its role

Every specialist is scoped to its role. Ask it to read your data, and reading is all it does — changes and deletions run through the specialists and approvals built for exactly that.

The agent factory

Six agent types, one set of building blocks

Underneath the copilot, Botminds is an agent factory. Six agent types assemble from the same building blocks.

Extraction

Pulls structured fields from any document.

Classification

Categorizes and routes — by section, page, or whole document.

Standardization

Normalizes messy values into canonical forms.

Quality-Control

A first-class reviewer that checks another agent's work before you trust it.

Conversational

Grounds chat in your own documents.

Team

Orchestrates a crew of member agents, shown live in the Crew view.

Loaded on demand

Tools and skills load into context on demand — an agent stays fast and lean no matter how large your library grows.

Mesh

Direct a crew, run it at scale

Directing a crew is half the story; running it at enterprise volume is the other. The Mesh execution engine was re-architected this cycle: dispatch is decoupled from execution, so the ceiling is your real worker pool.

Before
2.8 runs/s
V27
11.2 runs/s
≈4× the throughput per server

4× throughput

Dispatch decoupled from execution — 2.8 runs per second before, 11.2 after. Large batches clear dramatically faster on the same hardware.

Park & resume

Long, human-in-the-loop work now parks and resumes instead of holding a worker hostage.

Live status board

Every document, stage and agent moving in real time — you watch work flow through the platform as it happens.

Collections

Three types, one home for every kind of data

Data comes in three shapes — a library to learn from, files to process, and tables to query. You pick a Purpose when you create a collection, and the platform shapes itself around it.

Knowledge

A corpus your agents learn from. Automatically indexed, kept fresh, protected from accidental deletion, and attachable to any agent for retrieval grounded in exactly your material.

Processing

Documents flow through stages while agents act on them, on a lifecycle you control, with a built-in two-approval "four-eyes" workflow before anything is final.

Structured NEW

A rate card, a rulebook, a catalog kept as a real table: typed columns you query, sort and edit — and still a full collection with workflow, agents, review and dashboards behind it.

Structured, without the spreadsheet sprawl

Typed columns you query, sort and edit, a Reference field that links records into connected master data, all defined in the no-code Schema Designer — with agent-powered self-enrichment behind it.

Why it matters

Your data keeps its natural shape: a knowledge base behaves like a knowledge base, a pipeline like a pipeline, a table like a table — each with the right indexing, workflow and behavior by default.

Governance

Trust, on by default

In V27, control comes built in. Governance is a first-class pillar of Studio, sitting beside Agents, Data and Experience — backed by mechanism you can point to.

AgentsDataExperienceGovernance NEW PILLAR

A copilot that asks permission

The copilot that acts on your platform also asks: destructive and sensitive operations route through explicit human approval.

An audit trail you can stand behind

Every action lands in the trail — who did what, on which object, under which policy.

Spend, in plain sight

Cost is visible where the work happens, so you ship under control and see every dollar.

The full list

Everything new in V27

The keynote above is the story — this is the changelog: 700+ enhancements across seven areas, item by item.

Built for scale — faster, leaner, enterprise-ready

  • Re-architected execution engine — document-processing dispatch decoupled from execution, delivering roughly 4× more throughput per worker so large batches clear dramatically faster.
  • Unified data core — a single high-performance data store now serves documents, dashboards, labels and audit data, replacing several parallel storage systems and removing an entire class of sync delays.
  • Faster extraction turnaround — document reads along the extraction path re-routed to the new data core, cutting end-to-end extraction time on production workloads.
  • Leaner document screens — the document review page rebuilt to load a fraction of its former payload (over 4× lighter), so reviewers open documents near-instantly.
  • Concurrency hardening at 200+ users — caching, batched rendering and lost-update protection proven under sustained 200-concurrent-user load tests.
  • Snappier building experience — the configuration workspace eliminated redundant database round-trips, with heavy queries cut to a fraction of their previous cost.
  • Real-time processing visibility — live per-document run boards and progress indicators replace wait-and-refresh; you watch work flow through the platform as it happens.
  • Bulk ingestion at scale — large structured uploads now process in bounded, automatically re-queued batches, so big drops can't stall the pipeline.
  • Resource-safety hardening — connection pooling, backpressure and self-healing recovery across the processing tier eliminate stampedes and stuck runs under burst load.

Piper — an AI copilot that runs the platform with you

  • Piper, everywhere — one copilot available as a dockable side panel, a floating window, or a full-screen workspace, on every page of the platform.
  • Acts, not just answers — Piper performs real work through a governed action layer: creating and configuring agents, collections, dashboards and workflows conversationally.
  • Full platform coverage — every configuration menu is now reachable by conversation, backed by a large catalog of governed platform actions.
  • Compose — co-author extraction fields, rules and configuration with the copilot, iterating in plain language instead of forms.
  • Cost awareness — ask Piper what your AI spend looks like and where it's going; it reads live cost telemetry.
  • Grounded & hardened answers — stress-tested against a 651-question scenario storm and hardened with grounding rules, so answers stay anchored in your actual data and configuration.
  • Human confirmation on sensitive actions — anything consequential requires an explicit confirm step before Piper proceeds.
  • Smart fast/deep thinking — Piper automatically chooses between fast responses and deeper reasoning per question, so simple asks stay instant.
  • Guided learning built in — Piper draws on the new in-app Academy and product guide to explain any concept as it works with you.

Build solutions without code

  • Visual experience builder — assemble tailored review screens, workbenches and portals from a library of reusable cards, with a visual designer and live preview — no development required.
  • Configurable dashboard cards — metric cards, charts and structured-data lists wire directly to your data model through configuration, with a visual inspector for fine-tuning.
  • Experience packs — bundle a complete tailored interface as a pack and surface it as an additional tab inside the standard product, side by side with existing screens.
  • Redesigned Studio — the builder workspace reorganized around four clear pillars (Agents · Data · Experience · Governance), with a cleaner header and streamlined navigation.
  • Plain-English language — consistent, human terminology across the product (Lifecycle, Taxonomy, Collections) replaces internal jargon.
  • New visual flow editor — a rebuilt drag-and-connect workflow canvas with smooth connections and reliable editing.
  • Live Crew view — one unified tree shows every member of a solution's automation crew with live per-member status while work is running.
  • Structured Collections — a first-class home for typed tabular data, with schema, records pages, and agent-powered self-enrichment.
  • Botminds Academy in-app — a full learning curriculum built into the product, so new builders self-serve their onboarding.

Solution Hub — build once, deliver to many

  • Full-fidelity publishing — solutions travel complete: agents, data model, screens, workflows and experience packs arrive intact, verified by content-addressed assets.
  • Versioning & promotion — controlled roll-out across environments with full provenance, next-version publishing, and promote-to-next-ring actions.
  • One-click demo — install a solution and watch it run on sample data immediately, with live progress, so stakeholders see value in minutes.
  • Governed distribution — production releases go through explicit approval workflows, with a complete audit ledger on everything published or installed.
  • Cryptographically signed packages — every published solution is signed and verified at install, so what you install is exactly what was published.
  • Credential safety by design — publishing automatically strips every credential and secret, and installation prompts you to re-enter your own — keys never travel.
  • Install where you want — bring a Hub solution into a new project or an existing one, directly from the workspace's project menu.
  • Smart model binding — installed solutions automatically bind to your workspace's approved AI models, no manual re-wiring.
  • Honest progress everywhere — export, publish and install all report live step-by-step progress instead of a spinner.

Security, isolation & compliance

  • Strengthened tenant isolation — a dedicated adversarial security program reviewed and reinforced every boundary between customers, with automated tests that continuously prove one tenant can never read another's data.
  • Comprehensive threat hardening — systematic remediation across injection, request-forgery and access-control risk classes, covering every ingestion and integration pathway.
  • Access Overview — a single, clear view of who can access what across the platform, with integrated user and role management.
  • Stronger authentication — hardened admin-role gating, lockout protection against password-guessing, forced password reset on newly created accounts, and hardened two-factor flows.
  • Graceful, secure sessions — resilient sign-in with silent token refresh, idle-timeout countdowns, and context-aware logout — fewer surprise logouts, tighter session control.
  • Governed data lifecycle — controlled soft-delete and hard-delete for documents with an immutable audit trail of every removal.
  • End-to-end audit — builder actions, human review decisions and API-driven changes all land in one aligned governance audit stream.
  • Secret hygiene everywhere — solution exports scrub every credential shape, and platform pipelines run continuous secret-leak and vulnerability scanning as merge gates.
  • Approval-gated releases — production deployments and marketplace publications both require explicit human approval with full traceability.

See and control your AI spend

  • Unified AI gateway — all AI model traffic now flows through one governed gateway on every environment, giving a single accurate point of measurement and control.
  • Spend dashboards — real-time AI usage and cost analytics in the admin console, from daily trends down to individual workloads.
  • Per-workspace & per-project attribution — every unit of AI spend is tagged to the workspace and project that incurred it, ready for internal chargeback.
  • Durable spend history — daily cost rollups retained over a six-month range, so trends and month-over-month comparisons are always available.
  • Budgets & alerts — set spend expectations and get flagged before consumption runs ahead of plan.
  • Honest, accurate telemetry — cost accounting reconciled against provider billing (including cached-token pricing), so the numbers you see are the numbers you pay.
  • Cost in conversation — ask Piper about spend and get a live, accurate answer with breakdowns.
  • Efficiency behind the scenes — internal token-usage optimization cut copilot conversation overhead by an order of magnitude, lowering the platform's own AI footprint.

More sources, smarter answers, new reach

  • Modernized SharePoint connector — rebuilt on Microsoft Graph incremental sync, with automatic metadata mapping and system fields.
  • Azure Blob Storage connector — ingest directly from your cloud storage containers out of the box.
  • More file types, natively — JSON, PowerPoint, plain text, Word and Excel now ingest, index and preview in the built-in viewer alongside PDFs.
  • Knowledge Collections — auto-indexed, scoped knowledge bases that keep agent answers grounded, with self-healing indexing proven at thousand-document scale.
  • Structured-data ingestion — spreadsheets and CSV/JSON sources explode into clean, row-level records automatically.
  • Multi-source analysis — query across multiple datasheets with familiar SQL-style analysis and get inline charts in the answer.
  • Duplicate protection — strict duplicate detection on ingestion with proactive alerts, so the same file never silently processes twice.
  • Source-anchored answers — extracted values deep-link to the exact page in the source document, one click from value to evidence.
  • Deeper document intelligence — enhanced OCR options and entity-scoped retrieval sharpen answer accuracy on complex documents.
  • French-language foundation — the first step toward serving French-speaking users and customers.

The shape of it

A crew to do the work. A home for every kind of data. A no-code canvas to build on. One layer of trust over it all.

That's V27.