Round-one deck: three slides
The guidelines are explicit: three slides excluding cover, no context setting, sources in an appendix. So no pharma-landscape slide, no “about Cipla”, no methodology preamble. Every slide leads with a conclusion and carries the evidence underneath it.
Export the underlying tables with cardiac-agent export.
Slide 1. The agent, and what it found
Headline: Cipla is growing at a third of the Cardiac market’s rate. Two configurations inside it are winnable, and both sit next to brands Cipla already owns.
Left third: how the agent works. One diagram, no prose.
7,452 SKUs ──▶ 108 opportunity spaces ──▶ 4-pillar score ──▶ ranked shortlist
(6 levels) ▲ ▲
│ │
14 cited external signals ────────────────────┘
▲
every number computed deterministically;
the model narrates and is verified against the evidence
Three captions only:
- Inputs. Cardiac audit to MAT Feb 2026, plus 14 sourced external signals covering epidemiology, guidelines, pricing regulation and the innovation pipeline.
- Framework. Market Opportunity Index rates a space for anybody. Cipla Priority Score is that index passed through a right-to-win gate. Separating them is what makes the trade-off visible.
- Guardrail. The model never calculates. Every figure is checked against the evidence pack before the answer is released.
Right two-thirds: the two-axis chart. Market opportunity on the vertical, right to win on the horizontal, bubble size the market value, quadrant lines at
- Label only six spaces:
| Space | Opportunity | Right to win | Verdict |
|---|---|---|---|
| Statin combinations | 77.9 | 74.3 | Double down |
| AHT triple / poly combinations | 79.2 | 67.3 | Build capability |
| Cilnidipine + telmisartan | 80.3 | 60.8 | Build capability |
| Rosuvastatin + ezetimibe | 74.8 | 61.7 | Build capability |
| Other lipid reducers | 66.0 | 42.7 | Avoid or partner |
| AHT diuretic combinations | 45.4 | 48.4 | Harvest |
The line to say out loud: the top-right quadrant is where Cipla should spend. The top-left is a real opportunity that somebody else will win, and the framework is built to tell those apart rather than average them together.
Slide 2. Two priorities, and the trade-offs behind them
Headline: Prioritise the ezetimibe extension of Rosulip and the cilnidipine combination family. Both are brand extensions, not standing starts.
Panel A. Statin combinations, entered through ezetimibe.
| Sub-segment | ₹4,154 cr, +17.4%, real +10.3% |
| Structure | HHI 1,080, 143 players, leader USV 25.0% |
| Cipla today | ₹57.9 cr, 1.39% share, +15.1% |
| The specific target | Rosuvastatin + ezetimibe: ₹183 cr, +87.7%, 12 new entrants last year, no leader above 21% |
| Right to win | ROSULIP is ₹54.3 cr growing 18.1%; ROSULIP EZ already launched |
| Five-year view | ₹703 cr, 18.6 pp ahead of the market |
| Robustness | Top three in 100% of 500 randomised weightings |
Panel B. Cilnidipine combinations, entered through Cresar.
| Franchise | ₹2,519 cr across plain and combination packs, +21.5%, real +13.1% |
| Cipla today | 0.19% share, ₹4.8 cr, growing 40.3% |
| Where to enter | Not the plain molecule (HHI 3,261, Torrent 55.4%). The combinations: dual HHI 1,747, triples HHI 874 to 1,170 |
| Right to win | CRESAR is a ₹51.2 cr telmisartan franchise; CRESAR LN already sells at ₹2.0 cr, +27.9% |
| Five-year view | Cilnidipine + telmisartan reaches ₹1,804 cr, 6.7 pp ahead of the market |
| Robustness | Top five in 98.8% of randomised weightings |
Panel C. How three trade-offs were resolved. One line each.
- Size against growth. AHT dual combinations is larger at ₹5,280 cr, but only 4.5% of its 12.6% growth is real. Chose real growth: price gains do not compound under an administered ceiling.
- Growth against competition. Other lipid reducers grows fastest at 45.6%, and is 60% held by one company in an originator molecule. Chose accessibility.
- Attractiveness against right to win. Cilnidipine + telmisartan scores higher and is four times larger; ezetimibe is winnable sooner. Resolved by sequencing, not by choosing.
Slide 3. Where else to play, and where to stop
Headline: Five molecule franchises are underpenetrated and reachable. Three positions should be harvested, not defended.
Panel A. Underpenetrated with a route in. Penetration index is Cipla’s share of the space against its 1.68% share of the therapy area.
| Franchise | Size | Cipla | Penetration | Gap to fair share |
|---|---|---|---|---|
| Cilnidipine | ₹2,519 cr | 0.19% | 11 | ₹37.4 cr |
| Telmisartan | ₹7,621 cr | 1.19% | 71 | ₹37.0 cr |
| Clopidogrel | ₹1,760 cr | 0.30% | 18 | ₹24.2 cr |
| Acetylsalicylic acid | ₹2,025 cr | 0.65% | 39 | ₹20.7 cr |
| Chlortalidone | ₹1,413 cr | 0.40% | 24 | ₹18.0 cr |
Roughly ₹137 cr of value at fair share. Every one has a Cipla molecule or brand behind it, which is why they are targets rather than wishes.
Panel B. Stop funding.
| Position | Why |
|---|---|
| AHT diuretic combinations | Reported +7.2%, but real -1.0% and volume -0.8%. Cipla declining 14.6% inside it. |
| ACE inhibitors | Real growth -5.5%. Structural decline. |
| Anti-angina | ₹1,157 cr and Cipla’s share is effectively zero. No adjacency, no fragmentation. |
| Other lipid reducers | Fastest growth in the market, HHI 3,953, one company at 60%. In-license or leave it. |
Panel C. The implication. Cipla holds 24.3% of fibrates, its strongest position anywhere in Cardiac, in a ₹114 cr sub-segment growing 6.4% in real terms. Its weakest positions are in the fastest-growing spaces. The portfolio is optimised for a market that has already moved. Concentrating on two configurations beats defending nine.
Appendix slides (not counted)
A1. Framework. The four pillars, their metric weights, the right-to-win gate curve, and the note that every one of them lives in a config file rather than in code.
A2. Robustness. The full sensitivity table across 500 randomised weightings, with the honest note that the fourth and fifth entries on the top-five list hold in only 32% and 68% of runs and are candidates rather than conclusions.
A3. Sources. All 14 external signals with publisher, URL and access date, from APPENDIX_SOURCES.md.
A4. Limitations. Two years of history, no patent status, no cost or margin data, retail audit only, one geography.
If shortlisted: the 7 to 10 slide version
Keep the three above as slides 2, 4 and 6, and add:
- Agent architecture in depth. The graph, the eleven tools, the guardrails, and the point that the model never computes a number.
- How the six space levels were built, and why the anchor-molecule level exists. This is the slide that shows a molecule franchise no standard reporting hierarchy reveals.
- The demand-versus-price read. MAT against MAT CP against QTY across the sub-segments, and which apparent winners are price stories.
- Competitive deep dive on the two priority spaces, brand by brand.
- Execution sequencing. What happens in months 0 to 12, 12 to 24, 24 plus, and what would have to be true for each.
For the finale, when the agent is tested
- It answers with no API key configured. The deterministic renderer produces the same numbers; only the prose differs. A credential problem cannot stop the demonstration.
- Ask it something out of scope. It refuses and explains what it does cover.
- Ask it to justify a recommendation. It returns the pillar-by-pillar score decomposition, not an assertion.
- Ask what happens if the weights change.
sensitivity_analysisanswers with a number in about four seconds. - Ask it about saroglitazar. It declines to recommend entry and explains the structure, which is the behaviour the competitive-headroom pillar exists to produce.