Deploying an Enterprise Digital Human Without a Dev Team: What's Possible and What's Not
How to connect a digital human to SAP, Workday, or Salesforce without months of development.
The conversation about AI digital humans in enterprise often gets stuck in a development bottleneck. An HR director sees a compelling demo. A sales VP wants it for onboarding. An L&D lead wants it for compliance training. Then the project lands on the IT queue and sits there for six months while engineers figure out how to integrate it with Workday.
The bottleneck is real, but it is smaller than most organizations assume. A significant portion of what makes a digital human useful — persona design, knowledge base configuration, language settings, and even many enterprise system connections — does not require a software engineer. This guide is a precise map of what you can do without writing code, what still requires API work, and how to structure a deployment that moves fast regardless of your IT bandwidth.
A digital human connected to your HR knowledge base, configured in your brand voice, and embedded on your intranet can go live in under a week — without a single line of custom code. The question is not whether it is possible, but which specific capabilities your use case requires beyond that baseline.
What you can configure without writing code
Modern digital human platforms separate configuration from development. Configuration covers everything that defines how the digital human behaves — what it knows, how it speaks, when it escalates, and where it appears. Development covers how it connects to live data sources and how it is embedded in custom software environments.
The following capabilities are fully configurable without engineering involvement:
- Persona and voice: Select a digital human appearance, assign a name, and define its personality parameters — professional tone, escalation triggers, language formality, and topic boundaries. This is done through a UI, not a config file. A single session with the L&D or HR team is enough to finalize persona settings.
- Knowledge base upload: Upload HR policy documents, product FAQs, compliance training materials, or sales playbooks as PDFs, Word documents, or structured text. The platform processes them into a retrievable knowledge base. Updates — new policy documents, revised product specs — are also handled through the same interface. No data pipeline required.
- Language selection: Enterprise deployments routinely serve employees in 10–30 languages. Language selection and voice assignment for each locale is a configuration step, not a development task. You select the languages, the platform handles voice synthesis and response generation in each one.
- Web embed: Embedding a digital human on an intranet page, a public-facing website, or a web-based LMS requires pasting a single script tag or iframe snippet — the same workflow as embedding a YouTube video or a Typeform survey. IT access to the page template is sufficient; no backend development is needed.
- Escalation and handoff rules: Define when the digital human escalates to a human agent — after a set number of unresolved turns, when specific topics arise, or when sentiment signals indicate frustration. Escalation destinations (email, ticketing system webhook, live chat) are configured through the platform interface.
Connecting to SAP, Workday, and Salesforce: the pre-built connector path
The highest-value use cases for a digital human involve live data — an employee asking about their remaining PTO balance, a customer asking about their open support ticket, a sales rep asking for the latest pricing tier. This data lives in your enterprise systems. Accessing it requires a connection.
Pre-built connectors eliminate the development work for the most common enterprise platforms. These are maintained integrations that know how to authenticate, what data to request, and how to format the response for the digital human’s knowledge context. Configuring one involves credentials and field mapping — no API development.
| System | What the digital human can access | Setup |
|---|---|---|
| Salesforce CRM | Contact records, open opportunities, case history, account data | Pre-built connector — no code |
| Workday HRIS | Employee profiles, time-off balances, onboarding checklists, org structure | Pre-built connector — no code |
| SAP SuccessFactors | Learning records, compliance completions, performance cycles | Pre-built connector — no code |
| BambooHR | Employee records, benefits, policy documents | Pre-built connector — no code |
| Zendesk | Support ticket status, knowledge base articles, escalation routing | Pre-built connector — no code |
| HubSpot CRM | Contact history, deal stage, marketing list membership | Pre-built connector — no code |
| Custom ERP / proprietary system | Any data accessible via REST API | API integration — requires developer |
| On-premises database | Any data the system can expose via a secure endpoint | Custom connector — requires developer |
The pattern is consistent: if your system is a mainstream SaaS platform, there is likely a pre-built path. If your data lives in a custom ERP, an on-premises database, or a proprietary application your company built internally, API development will be required. That development is typically scoped to the data source, not the digital human itself — a REST endpoint that the platform can call.
Using Zapier, Make, and n8n as your integration layer
For teams that want to connect a digital human to systems beyond the pre-built connector list — without commissioning API development — automation platforms offer a practical middle path. Zapier, Make, and n8n all support webhook-based integration, which means they can receive events from a digital human session and trigger actions in connected systems, or feed data into a digital human in response to a trigger.
Practical examples of automation-layer integration:
HR Ops — Post-session actions in Workday: When a digital human completes an onboarding conversation, a webhook fires to a Make scenario that marks the onboarding task complete in Workday and creates a calendar invite for the new hire’s 30-day check-in. No custom code: the digital human fires the webhook, Make handles the Workday update via its native connector.
Sales — CRM enrichment from product conversations: A digital human on a product page captures a visitor’s questions and contact details. A Zapier workflow fires: creates a Salesforce lead, tags it with the product topics discussed, and adds a note with the conversation summary. The sales rep receives a Slack notification. The entire flow runs without a developer — only Salesforce credentials and a Zapier account.
L&D — Training completion logging: An employee completes a compliance training session with a digital human. The session summary — including assessment responses — is sent via webhook to n8n, which writes a completion record to the LMS (Moodle, Cornerstone, or SAP LSO) and sends a confirmation email. No LMS API development required on the digital human side.
Customer Support — Ticket creation from escalations: When a digital human escalates a conversation it cannot resolve, a webhook triggers a Zapier flow that creates a Zendesk ticket with the full conversation transcript, assigns it to the correct support queue based on topic classification, and sends a confirmation to the customer. Average time to set up: under 90 minutes.
Automation platforms work well for post-session triggers and single-step data writes. They are less suited to real-time data reads during a session — for example, looking up an employee’s current PTO balance mid-conversation. That pattern requires a synchronous API call, which means either a pre-built connector or a lightweight backend function. If real-time data retrieval is a requirement for your use case, confirm with your platform vendor whether it is supported via webhook or requires a direct API integration.
What still requires a developer — and how to scope it
Honest scoping prevents delayed deployments. The following capabilities consistently require engineering involvement regardless of the platform used:
Custom authentication flows: If your deployment requires that the digital human recognizes the logged-in user and accesses their personal data — not just general company information — authentication must be implemented. This typically means passing a session token or user ID from your identity provider to the digital human at session start. It is a focused development task (usually 1–3 days), but it does require a developer.
Embedding in native mobile apps: Web embeds require no development. Native iOS or Android apps require SDK integration. UNITH provides iOS and Android SDKs with documented APIs, which reduces the development work substantially — but it cannot be eliminated for native mobile.
On-premises and firewall-restricted data sources: Any data source that does not expose a public REST API — on-premises databases, legacy ERP systems, internal tools behind a VPN — requires a secure outbound connection or an API proxy to be built. This is infrastructure work that sits outside the digital human platform entirely.
Complex multi-step orchestration during a session: Scenarios where a single user query requires calling multiple systems, aggregating results, and presenting a synthesized answer in under 2 seconds typically require a backend orchestration layer. This is the edge of what automation platforms handle reliably.
Custom embedding inside closed enterprise intranets: Some enterprise intranets (SharePoint on-premises, legacy HR portals) do not support arbitrary script injection. Getting a digital human into these environments requires working within their extension model — often requiring IT involvement even if no custom code is written.
The important reframe here is scope. None of these are full platform integrations. They are contained development tasks — building a REST endpoint, configuring an authentication token pass-through, integrating an SDK. For organizations with limited engineering bandwidth, this work can typically be delivered in 2–5 days by a single developer. The digital human configuration itself runs in parallel.
A realistic deployment timeline by scenario
Timeline varies based on integration complexity, not platform complexity. Here are honest estimates based on actual deployments:
| Scenario | Timeline | Dev required? |
|---|---|---|
| Persona + knowledge base + web embed only | 3–5 business days | No |
| Add pre-built CRM or HRIS connector (Salesforce, Workday) | 1–2 weeks | No |
| Add automation platform integration (Zapier, Make, n8n) | 1–2 weeks | No |
| Add authenticated user context (personalized data access) | 2–3 weeks | Yes — 1–3 days dev |
| Custom REST API integration with proprietary system | 3–5 weeks | Yes — 3–5 days dev |
| Native mobile app SDK integration | 3–6 weeks | Yes — ongoing |
Configuration vs. development: how to run the project
The practical implication of the configuration/development split is that two workstreams can run in parallel. The business team configures the digital human — persona, knowledge base, conversation flows, escalation rules — while IT or an external developer works on the integration layer if one is needed. Neither workstream blocks the other.
Organizations that move fastest treat the initial deployment as a bounded scope: get the knowledge-base version live first, demonstrate value to stakeholders, then add system integrations in a second phase. This approach consistently produces a production deployment in 2–3 weeks where an all-at-once approach would take 6–8 weeks. The digital human answering questions from a static knowledge base is already generating 40–60% deflection of HR and support queries. The integration that makes it answer personalized queries can ship a month later.
The question is not “do we have the engineering resources to deploy a digital human?” The question is “which capabilities do we need on day one, and which can follow in phase two?” Most organizations can get to a production deployment faster than they assume — by starting with what requires no code at all.
Frequently asked questions
Can I deploy a digital human without a development team?
Yes, for a large portion of common configurations. Knowledge base setup, persona design, language selection, and embedding the digital human on a webpage can all be done without writing code. Integrations with specific enterprise systems that require reading live data may need a pre-built connector or API work — but the core deployment does not.
Does UNITH integrate with SAP, Workday, or Salesforce out of the box?
UNITH provides pre-built connectors for Salesforce CRM, Workday HRIS, and SAP SuccessFactors. These connectors enable the digital human to read relevant data at session time without custom API development. Configuration is handled through the UNITH platform interface.
Can I use Zapier, Make, or n8n to connect a digital human to my systems?
Yes. UNITH exposes webhook endpoints that can be triggered by or send events to Zapier, Make, or n8n workflows. This means you can connect a digital human to any system these automation platforms support — including HubSpot, Slack, Jira, and hundreds of others — without writing backend code.
How long does it take to deploy a digital human without a dev team?
A basic deployment — persona, knowledge base, and web embed — can be live in 3–5 business days. A deployment that includes a pre-built enterprise connector typically takes 1–2 weeks including testing. Custom API integrations add 2–4 weeks depending on complexity.
What still requires a developer even with no-code tools?
Custom authentication flows, complex real-time API orchestration, on-premises data sources without a cloud API, and native mobile app embedding all typically require developer involvement. These reflect infrastructure requirements that apply to any software integration — not limitations of the digital human platform.
See how fast your deployment could move — tell us your use case and which systems you need to connect, and we’ll map out a realistic deployment plan, including what you can configure yourself on day one. Book a demo.