Vancouver, BC · Slalom

AI Solutions Architect & Builder in Vancouver, BC

I solve real business problems with AI — not the other way around. Director, Global Salesforce at Slalom. Salesforce and Agentforce are the enterprise platform I work on.

AI for real business problems

Problems I solve

  • Teams looking for AI uses instead of starting from a business problem that needs a better answer.
  • AI experiments that never become working agents, automation, or production workflows — including Agentforce at work.
  • Practitioners who need practical upskilling, not another slide deck about generative AI.

How I approach it

I treat AI as a way to solve a real business problem, not as a solution looking for a home. That means agents and automation that show up in delivery work — including Agentforce on the Salesforce platform — with clear guardrails and time spent upskilling practitioners. I experiment in the open, then bring what holds up into enterprise architecture.

Relevant experience

  • Director, Global Salesforce at Slalom, sitting at the intersection of enterprise delivery and AI transformation.
  • Spearheaded Slalom’s internal AI Task Team to scale AI capabilities and adoption across the Western Canada practice.
  • Earned Agentforce credentials, including Salesforce Certified AI Associate (Agentforce), Agentblazer Innovator 2025, and the Agentforce Service and Prompt Builder Templates superbadges.
  • Currently architecting production Agentforce deployments for enterprise clients, with 2025 work expanding into AI enablement and agentic workflows.
  • Built AI-assisted delivery frameworks used on data migration work, including the Duravant migration, and reusable toolkits reused across engagements.

AI-assisted development

Problems I solve

  • Delivery teams that want AI in the software lifecycle — development, automation, and testing — without a toy demo.
  • Too many coding assistants, not enough clarity on which workflow actually fits how people work.
  • Manual scripting and repeatable delivery steps that still have not been turned into reusable AI-assisted frameworks.

How I approach it

I use AI the way I use any other tool: on real software. Hands-on with Cursor, Claude Code, GitHub Copilot, and custom agents for development, automation, and testing. The test is whether the work ships — a site, a calculator, a migration framework — not whether the prompt looked clever.

Relevant experience

  • Built this profile site with AI-assisted development (Nuxt 3, Nuxt Content, Tailwind, GitHub Pages), documented in the open.
  • Published a same-model comparison of Cline, GitHub Copilot, Claude Code, and Cursor for 2026 — interface, context, agentic behaviour, and cost.
  • Built AI-assisted delivery frameworks used on data migration work, including the Duravant migration, now reused across engagements.
  • 70+ GitHub repositories spanning Claude integrations, custom AI agents, Next.js, Vue/Nuxt, and experimental frameworks. Current tools include Cursor, Claude Code, Copilot, OpenRouter, and OpenAI APIs.

AI governance & security

Problems I solve

  • Teams shipping prompts before legal, security, and compliance have a repeatable way to say yes or no.
  • Shadow AI, unclear data residency, and vendor risk treated as an AI-specific panic instead of ordinary cloud due diligence.
  • Healthcare, financial services, and public-sector constraints that generic AI policies do not cover.

How I approach it

I help teams move fast and move safely: data classification, vendor assessment, access controls, and the regulations that actually apply. The goal is governance that prevents shadow AI and protects sensitive data without becoming a policy document nobody uses.

Relevant experience

  • Written a practical AI data and security governance checklist organized by organisation type, including healthcare, financial services, government, enterprise, and small business.
  • Published a detailed breakdown of Claude, Anthropic, and cloud security — including when Amazon Bedrock is the more appropriate control plane — and recorded a podcast episode on the same question.
  • Delivery work in regulated and high-trust environments, including Fraser Health, IG Wealth Management secure file exchange, and earlier HSBC Trade & Supply Chain platforms.
  • Served as Enterprise Architect on a long-running enterprise account, covering cross-cloud governance, roadmap alignment, and technical risk management.

Experiments with AI tools

Problems I solve

  • AI enthusiasm that never turns into a business use case someone can actually run.
  • New tools shipping every week, with no habit of trying them on real work before recommending them.
  • Side projects and production work treated as separate worlds, so experiments never inform delivery.

How I approach it

I am an AI enthusiast who tests tools by building with them. This site, the Canadian finance calculators, and live experiments — including an AI assistant that prepares weekday trading tickets — are how I learn what is useful, then bring the pattern back to business automation.

Relevant experience

  • Built and maintain this site as a living profile: Nuxt 3, markdown content, and GitHub Pages, with AI pair programming in the loop.
  • Shipped free Canada-focused tools — mortgage, tax, and HELOC calculators, plus a rental application and todo board — as practical software, not slides.
  • Ran a live micro-trading experiment with Malli, an AI assistant in the Muse app: weekday tickets, two Canadian stocks, real brokerage orders.
  • 70+ GitHub repositories and ongoing experiments with Claude, Cursor, custom agents, Next.js, Vue/Nuxt, Supabase, and related tools.

Salesforce & Agentforce at work

Problems I solve

  • Multi-cloud programs (Sales, Service, Experience, CPQ) that need one architecture, not four disconnected implementations.
  • Org mergers, large data volumes, and Experience Cloud portals that have to stay secure under real traffic.
  • Executives who need demos and options they can decide on, not a solution already locked in the architect’s head.

How I approach it

Salesforce and Agentforce are the enterprise platform I work on — experience and proof, not the headline. I stay hands-on across Sales Cloud, Service Cloud, Experience Cloud, CPQ, and Agentforce, and translate that into language executives can use.

Relevant experience

  • 12+ years specializing in Salesforce architecture, delivery, and practice leadership, with 12+ Salesforce certifications including Experience Cloud Consultant.
  • Director, Global Salesforce at Slalom. Currently architecting production Agentforce deployments for enterprise clients.
  • Program Architect for Avigilon (Motorola Solutions) Fusion X, a multi-brand CRM consolidation spanning cross-cloud governance and data migration.
  • Led technical architecture streams for Rockwell Automation and Fraser Health discoveries, including PRM work and demos for non-technical stakeholders.
  • 2025 platform work included FCL Sales Cloud and Distributed Marketing, 407 Express Phase 1 platform stabilization and data architecture, and reusable Map Point Management for desktop and mobile.
  • Earlier: Director of Development Services at Groundswell Cloud Solutions in Vancouver, and Salesforce / Force.com delivery at TELUS.

How I work

Great technology serves people, not the other way around. Whether the work is introducing AI into a delivery team or architecting on an enterprise platform, the same sequence applies: start with the human need, design backward from there, and measure outcomes that actually matter.

AI adoption fails when it is treated as a technology project. It succeeds when it is treated as a people and process transformation — with governance, guardrails, and leaders who can bring skeptics along.

That is also how I lead. At Groundswell I learned people leadership by doing, starting with co-op students. At Slalom I have grown the Western Canada Salesforce practice, mentored developers into architects, and stayed hands-on as Program Architect, Integration Architect, and Enterprise Architect on strategic accounts.

Get in touch

Have a real business problem AI could solve? Let's talk.

The best way to reach Prishan is on LinkedIn.

Connect on LinkedIn