The context
By 2023, AI had rapidly become table stakes in enterprise SaaS. Customers expected semantic understanding, instant answers, and intelligent assistance — but only if the results were accurate, explainable, and trustworthy.
At Bigtincan, AI sat at a critical intersection:
Content discovery
Sales productivity
Platform credibility
When it worked, it accelerated adoption.
When it didn’t, it eroded trust faster than almost any other feature.
How it started
The first AI capability shipped was SearchAI — a semantic search experience layered alongside the existing keyword search as a proof of concept.
Technically, it worked.
Experientially, it exposed something deeper.
AI was being introduced faster than the platform could support it coherently.
caption
The future pressure
As GenieAI entered the roadmap and acquisition planning accelerated, AI could no longer exist as isolated features.
It needed to operate as a shared platform capability — consistently, safely, and across products.
The tension
AI raised expectations — but the platform wasn’t designed for intelligence as a system.
Multiple search paradigms across products
Fragmented retrieval logic
Product-specific AI implementations
Limited analytics and experimentation tooling
AI logic often defined before UX involvement
The challenge wasn’t adding AI features.
It was preventing intelligence from fragmenting the experience further.
The problem
SearchAI and traditional search could surface different answers on the same page.
GenieAI relied on SearchAI for retrieval, but its UI patterns weren’t consistently exposed across products. Other areas lacked true search entirely, relying on inline filtering or manual navigation.
Users couldn’t tell:
Which answer to trust
Where AI applied
Why results differed
Whether content was complete or scoped
A concrete moment where trust broke
A sales rep searches for a pitch deck before a customer call.
On the same screen:
SearchAI surfaces a summarized document as the “best result”
Traditional search ranks a different file first
Both appear valid.
Neither explains why.
The rep hesitates — not because AI failed, but because the system offered two conflicting truths.
Fundamental issue
Understanding the system
Through close collaboration with Product, Customer Success, and Engineering, it became clear these weren’t isolated UX issues.
They were symptoms of:
Fragmented retrieval models
Inconsistent entry points into discovery
No unified search surface across products
AI logic divorced from user mental models
A pattern emerged:
AI needed to help users find the right content, understand it in context, and act — not exist as separate tools scattered across the platform.
The constraints
AI logic often defined before design involvement
Limited access to usage analytics and experimentation tools
Acquisition and merger priorities shifting roadmap focus
Multiple products with inconsistent search foundations
Impact for this work was therefore qualitative and directional, validated through Customer Success feedback, leadership alignment, and usability patterns rather than clean quantitative attribution.
What I owned
My role focused on experience definition, interaction patterns, and trust models, working within engineering-defined AI constraints.
SearchAI
UX and interaction design within dev-defined logic
Visual and UI refinement to reduce confusion and duplication
Desktop and mobile experiences
GenieAI
Research and early concepting
Mental models and trust patterns
UI foundations for Knowledge Scopes
Handoff prior to full execution
Global AI Search POC
Design-led research and framing
Platform-level recommendations
Exploratory wireframes to align teams
How it came together
Rather than arriving as a single vision, the intelligence layer emerged over time.
Each phase surfaced new constraints — and new clarity.
From this, a set of consistent design principles began to take shape.
Phase 1 — SearchAI Stabilization
2023
Focused on clarifying AI vs keyword results, reducing duplication, and improving comprehension within an engineering-defined system.
Outcome
Improved usability and reduced confusion — but structural issues remained.
Phase 2 — GenieAI & Trust Patterns
2023-2024
Expanded AI beyond search into an assistant experience.
Knowledge Scopes became critical for enterprise trust and safe adoption.
Outcome
Customers could adopt AI while maintaining control over scope and accuracy.
Phase 3 — Global Search POC
2025
Reframed AI as a shared discovery foundation across products.
Defined principles for:
Unified search entry points
Assistant behaviors
Platform-wide intelligence
Status
Exploratory and direction-setting. Not shipped due to acquisition priorities.
AI as an Overlay, Not a Destination
Rather than pulling users into a separate AI experience, AI was designed to live alongside existing UI:
Inline summaries
Contextual suggestions
Side-panel reasoning
Action prompts that respected existing workflows
Why it mattered:
Users stayed oriented, cognitive load dropped, and trust increased — without forcing behavior change.
Separate Retrieval from Interpretation
A deliberate distinction emerged:
Search (retrieval): finding the right things
AI (interpretation): summarizing, explaining, suggesting
Combined moments: accelerating decisions
This prevented AI from feeling random or magical.
Why it mattered:
Users understood where answers came from — and why they could trust them.
caption
Optimize for “Next Steps,” Not Clever Answers
The most valuable AI moments weren’t verbose responses — they were clear next actions:
Summaries without opening files
Suggested actions across one or multiple items
Creation flows triggered directly from context
Why it mattered:
AI supported momentum and decision-making instead of becoming another content surface.
caption
Scope as a Trust Boundary
Enterprise adoption required control.
Knowledge Scopes became a trust mechanism, not just a configuration:
Users understood what AI could and couldn’t see
Admins reduced risk without disabling value
caption
Behavioral & UX impact
Reduced confusion around AI vs non-AI results
Clearer mental models for AI outputs
Safer adoption through scoped knowledge
Strategic impact
Positioned AI as a platform capability, not a feature
Informed future consolidation discussions
Reduced long-term UX debt risk
Validated impact
Customer Success feedback confirmed improved trust and comprehension
Leadership alignment on the need for unified search
POC used as reference in platform-level discussions
Where it stalled
Platform and merger priorities paused execution
Limited analytics constrained validation
AI systems evolved faster than UX maturity
The real lesson
What I'd do differently
I would push harder for earlier design involvement in AI architecture decisions — particularly around retrieval and ranking logic. Several UX issues we later addressed were already “locked in” by technical choices made before design input.
Even so, establishing platform-level principles — separating retrieval from interpretation, treating AI as an overlay, and optimizing for next steps — created a durable foundation that extended beyond any single feature.
