Litmus-Tested: How Framework Friday Prioritizes What to Automate First in FP&A - Smarter automation starts with better questions

By Dasun Bhagya Sapuarachchi – Doctoral Researcher, Law Student, MBA, BBA (Sp.) Senior Manager – Financial Planning and Analysis, WebLife
Published on
July 22, 2025

Why Finance Can’t Afford Blind Automation

Finance teams today face a different kind of pressure. Customer acquisition costs shift by the hour. Supply chains tighten and loosen without notice. The spreadsheet-based status quo doesn’t cut it anymore.

At Framework Friday (formerly WebLife), we decided to rethink Financial Planning & Analysis (FP&A) from the ground up. Not just automating for speed, but turning workflows into intelligent systems that adapt, learn, and make decisions in real time.

But here’s the thing - automation only works when it’s deliberate. Not everything should be automated. That’s where our Litmus Framework comes in.

The Real Problem With Most Automation Efforts

Let’s call it out: many automation projects start from fatigue. A dashboard isn’t updating fast enough, or a month-end close is painful, so the default solution is “let’s automate it.”

This often backfires. You end up with disconnected tools, expensive subscriptions, and brittle logic that only one analyst understands. Efficiency doesn’t scale if your automation isn't strategic.

So, at FF, we flipped the question: not what can we automate? but what should we agentize - what deserves to become an intelligent, evolving system?

The Litmus Framework: Our Strategic Filter

We don’t automate based on trendiness. We use three specific criteria to evaluate any FP&A workflow before it moves into design:

1. Strategic Fit

Does this process influence a key performance area - like CAC, gross margin, inventory turns, or strategic planning? If not, it’s out.

2. Agentic Potential

Can the system adapt on its own? That means real-time data, dynamic logic, feedback loops, and a clear path from input to insight.

3. Reusable Learning Value

Does the output create assets we can reuse across pods - like prompt libraries, forecasting templates, or logic modules? If it’s a one-off, it’s not worth the build time.

A workflow must meet all three to move forward. That keeps our automation focused, adaptive, and scalable.

What Passed and What Didn’t

✅ CAC Forecasting Agents

We built this to forecast customer acquisition cost using GPT models and live marketing inputs. It pushes Slack alerts to align finance with campaign performance.

  • Strategic Fit: Yes, CAC is a top-line metric
  • Agentic Potential: High predictive modeling, auto-updates
  • Reusable: Prompt structures now used in LTV and cohort tools

✅ ClickUp-to-Slack Budget Deviation Alerts

This agent flags budget mismatches in ClickUp and sends timely updates via Slack.

  • Strategic Fit: Keeps teams on budget, supports discipline
  • Agentic Potential: Expands with complexity
  • Reusable: Logic now runs in ops and product pods

❌ Rejected: Automated Slide Deck Builders

Sounded appealing, didn’t pass the filter.

  • Strategic Fit: Weak KPI alignment
  • Agentic Potential: Low - it’s static formatting
  • Reusable: Too custom, hard to generalize

Why This Matters: From Reporting to Intelligence

When done right, agentic FP&A doesn’t just reduce manual work. It lifts the entire finance function - from being a reporting layer to being a strategic nerve center.

The Litmus Framework helps us:

  • Focus developers on real impact, not one-off tools
  • Build systems that scale as our business grows
  • Align finance automation with real decision-making needs

Most importantly, it ensures we’re not automating for automation’s sake - we’re building adaptable, intelligent infrastructure.

Advice for Other Finance Teams

If you're just starting your automation journey, try asking:

  • Does this process affect core business performance?
  • Can it adapt to real-time data and surface insights?
  • Will this build something reusable for other teams?

If the answer to all three isn’t a strong yes - don’t build it yet. In the world of intelligent systems, selective automation is your leverage.

Final Thought

We believe finance should lead - not react - in the age of intelligent business. That leadership starts by asking better questions before writing a single line of automation code.

And we’ll keep sharing what works, so others can build alongside us.


“Smart finance starts with better questions.”