Agents / Portfolio Agent

Which product, pack, price and promotion will win this opportunity – profitably?

The Portfolio Agent works on the whole mix: the product, its pack, its price and its promotion. It writes the product brief from consumer research, product tests and where demand is growing, and checks the margin of every move before it reaches the market.

Works on its own or as part of the chain – guided step by step, or autonomously.

See it work, step by step Request a demo
Step 1
Product brief
Step 2
Pack & price
Step 3
Promotion & margin
Demand Agent
Consumer Agent
Sales Agent
Portfolio Agent
The Sales agent hands over to the Portfolio agent
Picked up from the Sales Agent

The case so far

Demand sized the opportunity. Consumer defined who to win. Trade made the shelf ready: a single-bar listing for convenience, 20 priority stores and order quantities for 46 stores. It handed over three questions for the Portfolio Agent.

See how the Sales Agent got here
Most of the time
The brief comes from research, the price from last year’s ladder, the promotion from the trade calendar. Each team optimises its own lever, and the misalignment shows up in the P&L.
With the Portfolio Agent
One view of who the product is for, what they will pay and which promotions pay back, so product, price and promotion pull in the same direction.
What the Portfolio Agent reads
  1. Consumer research & U&A
  2. Product & concept tests
  3. Market data & competitor prices
  4. Sell-out by SKU & price point
  5. Promotion history
  6. Margin & cost data
  7. CRM discount response
  • Observed what people did
  • Stated what they told research
  • Predictive model scores
  • Kept apart and labelled on every number.
Three steps · one example

You ask in plain words. The agent plans, you approve, it answers.

Each step is one question to the Portfolio Agent. The answer comes back validated, with weak evidence flagged and outputs ready for your systems.

Shown step by step – it can also run on its own. The case below is guided: one question per step, with suggested prompts leading to the next. Give the Portfolio Agent a broader brief and it works autonomously – it plans the analysis itself, goes as deep as the question needs, skips steps that add nothing and comes back with the answer and its full audit trail.

Portfolio agent - Product brief
Step 1

Product brief

A brief for a new SKU, or for adjusting an existing one, grounded in product tests, consumer research and where demand is actually growing.

Other questions it answers
  • “Should we reformulate the current bar or add a new variant?”
  • “Which claim should lead the pack: ‘less sugar’ or ‘high protein’?”
  • “Which SKUs could we cut without losing shoppers?”
Next: step 2 ↓
Portfolio agent - Pack & price
Step 2

Pack & price

Price tiers, pack sizes and occasions read against competitor positioning and against how each group responds to price. Where price history is too thin for a reliable elasticity estimate, the agent says so.

Other questions it answers
  • “Where does our 2 L pack sit against private label?”
  • “Is there room for a smaller pack at a higher price per kilo?”
  • “Which price points does each consumer group actually buy at?”
Next: step 3 ↓
Portfolio agent - Promotion & margin
Step 3

Promotion & margin

Promotion response and calendar, range fit by store cluster, cannibalisation, and margin guard-rails on every move.

Other questions it answers
  • “Which promotions buy volume at the expense of margin?”
  • “Which SKUs could we delist with the least volume loss?”
  • “What happens to margin if we move to everyday low price in discounters?”
Hand-off ↓
Activate directly

Every move goes to the system that executes it.

The brief goes to R&D and the agency, the price corridor to your pricing tool, the launch plan to trade-promotion management and the sampling list to the field – each with its expected margin impact.

Every move goes to the system that executes it

Shown in this case
Product briefs
Pack & price
Promotion & margin
Examples we tune or build per client
+ Range rationalisation
+ Claim testing
+ Everyday-low-price simulation
Inside the Portfolio Agent

A family of sub-agents, tuned to you

The three steps above are run by specialised sub-agents. We tune them to your categories, data and ways of working – and build new ones when your teams need them.

The plan goes live – and comes back as the next question
Close the loop

The plan goes live – and comes back as the next question.

Audiences, store lists, prices and promotions go to your CRM, field force, pricing and trade-promotion systems. In 12 weeks the Demand Agent measures the result against a control.

What you get

Every recommendation is computed by statistics and machine learning, validated, and traceable to the data cell it came from. The language model picks the variables and writes the explanation; it does not calculate the numbers.

A product brief for a new or adjusted SKU, backed by research, product-test and demand evidence
Pack, price, promotion and range moves, each with its expected margin impact
Moves that go into your pricing and trade-promotion systems

Numbers from live deployments

Example
Lifting beverage margin without losing volume
Repricing 17 SKUs, delisting 9 under-performers, rebuilding the promotion calendar and redistributing 8 SKUs across 4 regions: +5% gross margin, with volume share held within one point.
+7%
gross margin, year on year
−47%
dead stock
−63%
revenue lost to out-of-stocks

Bring your toughest question. Watch the Portfolio Agent answer it.

A 30-minute live session on prepared data: the agent interprets your question,
runs the analysis and returns a validated answer you can drill down to the data cell.

Request a demo
demo@sightpulse.ai