Move fast, with guardrails
Give departments a sanctioned way to build, while IT sets the models, tools, roles and approval policies they build with.
Trust & security
New Prompt version history, reranking & audit logs →
AI that makes it past the pilot.
The governed AI Operating System. Build agents and workflows on the models you choose, ground them in your knowledge and systems, and deploy them to the web, Teams, Slack and the phone — with every run visible to central IT.
Real platform UI · representative demo data
Trusted by 100+ institutions
eCampusUniversity System of Georgia
Lamar University
≈200K
conversations handled by Gecko AI agents each month
200+
AI agents launched across education
<1 wk
to a production agent on average, with managed implementation
3
independent regional deployments — Ireland, US and Canada
The CIO imperative
AI demand is everywhere. Institutional control has to catch up.
Your teams are already experimenting. The job now is turning scattered enthusiasm into a secure, connected and measurable institutional capability.
Connect the whole institution
Ground agents in your websites, SharePoint and drives, and let them act in the SIS, CRM and proprietary systems through open MCP.
Models & integrationsMake value visible
See what ran, which model and tools it used, how long it took, how many tokens it cost — and whether a human signed it off.
Observability
Explore the platform
One operating system for every AI use case.
Six capability groups, one set of controls. Pick an area to see what it does — then dive into the detail.
Build agents from governed parts.
An agent is a versioned prompt, an approved model, scoped knowledge and exactly the tools IT grants it. Or describe the outcome in plain English and let the AI Agent Builder draft it for review.
- Central prompt library with fragments, variables and version history
- Swap the model behind a live agent from a dropdown
- Nothing is created by the AI builders until you approve it
Design the process. Let agents do the work.
A visual builder with 20 step types — AI, retrieval, branching, parallel paths, MCP tools, messages, outbound calls, sandboxed code and human checkpoints — triggered by webhooks, schedules or agent events.
- Approvals as interactive cards in Slack or Teams
- Published versions are immutable; every run binds to one
- Export as JSON — your process logic stays portable
Answers grounded in sources you can inspect.
Crawl your websites and SharePoint, sync folders from Google Drive, OneDrive, Dropbox and S3, or upload PDF, Word and CSV files. Hybrid search blends meaning with exact matches — and every answer records its sources.
- Hybrid retrieval — vector plus full-text, fused by rank
- Relevance scores on every cited source
- Page-level usage shows which content earns its keep
Vendor-neutral models. Agent-ready systems.
Use OpenAI, Anthropic, Google, Amazon Bedrock, Microsoft Foundry, OpenRouter or any OpenAI-compatible endpoint — with your own keys or ours. Connect Ellucian Ethos, Salesforce and Gecko Engage, and any other system through MCP.
- IT curates which models each tenant can use
- 22+ built-in integrations — and any system via MCP
- Tools granted one checkbox at a time — nothing by default
One agent, wherever people ask for help.
Publish the same governed agent to a website chat bubble, inline chat or AI search, to Microsoft Teams and Slack, to a phone line answered by an AI receptionist, and to SMS, email and the API.
- One script tag for web embeds — WCAG 2.2 AA enforced
- AI receptionist routes calls or warm-transfers to people
- Same knowledge, tools and guardrails on every channel
Governance as a query, not a policy PDF.
Guardrails screen input before the model sees it. Approvals and identity checks pause work for people. Every model call, tool call and decision lands in one explorer — and admin changes land in the audit log.
- 32 built-in detectors for prompt injection and sensitive data
- Roles scoped down to individual agents and workflows
- Filter every run by guardrail hit or approval state
Proven in the real world
AI outcomes from institutions like yours.
Published results from Gecko AI agents in production. All success stories
96%
of student questions answered after deploying its Gecko AI agent.
Read the story
1,081 hours
saved in seven months, while 63% of agent conversations happened out of hours.
Read the story eCampusUniversity System of Georgia 30,000students supported around the clock by Gecko AI agents.
Read the story
95%+
success rate for its Gecko AI agent after connecting it to live university website content.
Read the storyResults are from published Gecko agent case studies. Every institution and use case is different.
Start where value is clear
One platform. Every team.
Build a reusable AI capability once, then let each department apply it to the work that matters most.
Student services
Always-on, source-grounded answers — with sensitive cases routed to the right human team with full context.
See the workflowAdmissions & recruitment
Qualify enquiries and personalise follow-up at scale, with recruiter approval before anything is sent.
See the workflowIT & operations
Verify the user, check live systems through MCP, guide the fix — and raise an enriched ticket when people are needed.
See the workflow
Institutional assurance
Built for the questions your security team will ask.
The controls behind the platform, stated plainly — because “trust us” is not an architecture.
- SSO including SAML; the console is SSO-only in production
- AES-256-GCM encrypted, write-only secrets and credentials
- Audit log of every create, update and delete — by user or API key
- AWS hosted with SOC 2 and ISO 27001 assurance
eu-west-1IrelandUK & EU institutions
us-east-1United StatesUS institutions
ca-central-1CanadaCanadian institutions
Independent regional stacks — each with its own database, secrets and domain.
32 guardrail detectors
Prompt injection and sensitive data — flag, redact or block before the model sees it.
Humans where it matters
Single-use approval cards in Slack and Teams, and emailed identity checks mid-run.
A flexible operating model
Start fully managed. Take the wheel when you’re ready.
Gecko can launch and run your first agents, then help your teams build the confidence and patterns to deploy their own — on the same platform.
Fully managed launch
We define the outcome, connect trusted systems, configure guardrails, test, deploy and monitor your agent.
Gecko leadsBuild together
Your team works alongside ours, reusing proven implementation patterns while capability grows.
Shared deliveryBuild your own
Trained teams create and deploy their own agents with central controls, observability and support in place.
Your team leads
Make your next pilot a path to production
Bring us one process. Leave with a plan for a governed agent.
We’ll help you define the outcome, the controls and the evidence needed to put it into production.
- A 30-minute conversation, not a sales deck
- A scoped first use case and the controls it needs
- Assurance documents in our Trust Center